Commit graph

741 commits

Author SHA1 Message Date
Eric Wong b6c5241606 hashmap_get takes "const struct hashmap_entry *"
This is less error-prone than "const void *" as the compiler
now detects invalid types being passed.

Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-10-07 10:20:10 +09:00
Eric Wong b94e5c1df6 hashmap_add takes "struct hashmap_entry *"
This is less error-prone than "void *" as the compiler now
detects invalid types being passed.

Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-10-07 10:20:10 +09:00
Eric Wong d22245a2e3 hashmap_entry_init takes "struct hashmap_entry *"
C compilers do type checking to make life easier for us.  So
rely on that and update all hashmap_entry_init callers to take
"struct hashmap_entry *" to avoid future bugs while improving
safety and readability.

Signed-off-by: Eric Wong <e@80x24.org>
Reviewed-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-10-07 10:20:09 +09:00
Elijah Newren 8e4ec3376e merge-recursive: fix the diff3 common ancestor label for virtual commits
In commit 743474cbfa ("merge-recursive: provide a better label for
diff3 common ancestor", 2019-08-17), the label for the common ancestor
was changed from always being

         "merged common ancestors"

to instead be based on the number of merge bases:

    >=2: "merged common ancestors"
      1: <abbreviated commit hash>
      0: "<empty tree>"

Unfortunately, this did not take into account that when we have a single
merge base, that merge base could be fake or constructed.  In such
cases, this resulted in a label of "00000000".  Of course, the previous
label of "merged common ancestors" was also misleading for this case.
Since we have an API that is explicitly about creating fake merge base
commits in merge_recursive_generic(), we should provide a better label
when using that API with one merge base.  So, when
merge_recursive_generic() is called with one merge base, set the label
to:

         "constructed merge base"

Note that callers of merge_recursive_generic() include the builtin
commands git-am (in combination with git apply --build-fake-ancestor),
git-merge-recursive, and git-stash.

Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-10-02 14:59:29 +09:00
Jonathan Tan 83e3ad3b12 merge-recursive: symlink's descendants not in way
When the working tree has:
 - bar (directory)
 - bar/file (file)
 - foo (symlink to .)

(note that lstat() for "foo/bar" would tell us that it is a directory)

and the user merges a commit that deletes the foo symlink and instead
contains:
 - bar (directory, as above)
 - bar/file (file, as above)
 - foo (directory)
 - foo/bar (file)

the merge should happen without requiring user intervention. However,
this does not happen.

This is because dir_in_way(), when checking the working tree, thinks
that "foo/bar" is a directory. But a symlink should be treated much the
same as a file: since dir_in_way() is only checking to see if there is a
directory in the way, we don't want symlinks in leading paths to
sometimes cause dir_in_way() to return true.

Teach dir_in_way() to also check for symlinks in leading paths before
reporting whether a directory is in the way.

Helped-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-09-20 10:15:57 -07:00
Elijah Newren 4615a8cb5b merge-recursive: alphabetize include list
Other than cache.h which needs to appear first, and merge-recursive.h
which I want to be second so that we are more likely to notice if
merge-recursive.h has any missing includes, the rest of the list is
long and easier to look through if it's alphabetical.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren 45ef16f77a merge-recursive: add sanity checks for relevant merge_options
There are lots of options that callers can set, yet most have a limited
range of valid values, some options are meant for output (e.g.
opt->obuf, which is expected to start empty), and callers are expected
to not set opt->priv.  Add several sanity checks to ensure callers
provide sane values.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren f3081dae01 merge-recursive: rename MERGE_RECURSIVE_* to MERGE_VARIANT_*
I want to implement the same outward facing API as found within
merge-recursive.h in a different merge strategy.  However, that makes
names like MERGE_RECURSIVE_{NORMAL,OURS,THEIRS} look a little funny;
rename to MERGE_VARIANT_{NORMAL,OURS,THEIRS}.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren 5bf7e5779e merge-recursive: split internal fields into a separate struct
merge_options has several internal fields that should not be set or read
by external callers.  This just complicates the API.  Move them into an
opaque merge_options_internal struct that is defined only in
merge-recursive.c and keep these out of merge-recursive.h.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren e95e481f9e merge-recursive: avoid losing output and leaking memory holding that output
If opt->buffer_output is less than 2, then merge_trees(),
merge_recursive(), and merge_recursive_generic() are all supposed to
flush the opt->obuf output buffer to stdout and release any memory it
holds.  merge_trees() did not do this.  Move the logic that handles this
for merge_recursive_internal() to merge_finalize() so that all three
methods handle this requirement.

Note that this bug didn't cause any problems currently, because there
are only two callers of merge_trees() right now (a git grep for
'merge_trees(' is misleading because builtin/merge-tree.c also defines a
'merge_tree' function that is unrelated), and only one of those is
called with buffer_output less than 2 (builtin/checkout.c), but it set
opt->verbosity to 0, for which there is only currently one non-error
message that would be shown: "Already up to date!".  However, that one
message can only occur when the merge is utterly trivial (the merge base
tree exactly matches the merge tree), and builtin/checkout.c already
attempts a trivial merge via unpack_trees() before falling back to
merge_trees().

Also, if opt->buffer_output is 2, then the caller is responsible to
handle showing any output in opt->obuf and for free'ing it.  This
requirement might be easy to overlook, so add a comment to
merge-recursive.h pointing it out.  (There are currently two callers
that set buffer_output to 2, both in sequencer.c, and both of which
handle this correctly.)

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren a779fb829b merge-recursive: comment and reorder the merge_options fields
The merge_options struct had lots of fields, making it a little
imposing, but the options naturally fall into multiple different groups.
Grouping similar options and adding a comment or two makes it easier to
read, easier for new folks to figure out which options are related, and
thus easier for them to find the options they need.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren 8599ab4574 merge-recursive: consolidate unnecessary fields in merge_options
We provided users with the ability to state whether they wanted rename
detection, and to put a limit on how much CPU would be spent.  Both of
these fields had multiple configuration parameters for setting them,
with one being a fallback and the other being an override.  However,
instead of implementing the logic for how to combine the multiple
source locations into the appropriate setting at config loading time,
we loaded and tracked both values and then made the code combine them
every time it wanted to check the overall value.  This had a few
minor drawbacks:
  * it seems more complicated than necessary
  * it runs the risk of people using the independent settings in the
    future and breaking the intent of how the options are used
    together
  * it makes merge_options more complicated than necessary for other
    potential users of the API

Fix these problems by moving the logic for combining the pairs of
options into a single value; make it apply at time-of-config-loading
instead of each-time-of-use.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren 7c0a6c8e47 merge-recursive: move some definitions around to clean up the header
No substantive code changes (view this with diff --color-moved), but
a few small code cleanups:
  * Move structs and an inline function only used by merge-recursive.c
    into merge-recursive.c
  * Re-order function declarations to be more logical
  * Add or fix some explanatory comments

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren bab56877e0 merge-recursive: rename 'mrtree' to 'result_tree', for clarity
It is not at all clear what 'mr' was supposed to stand for, at least not
to me.  Pick a clearer name for this variable.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:04 -07:00
Elijah Newren ff1bfa2cd5 merge-recursive: use common name for ancestors/common/base_list
merge_trees(), merge_recursive(), and merge_recursive_generic() in
their function headers used four different names for the merge base or
list of merge bases they were passed:
  * 'common'
  * 'ancestors'
  * 'ca'
  * 'base_list'
They were able to refer to it four different ways instead of only three
by using a different name in the signature for the .c file than the .h
file.  Change all of these to 'merge_base' or 'merge_bases'.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 4d7101e25c merge-recursive: fix some overly long lines
No substantive code change, just add some line breaks to fix lines that
have grown in length due to various refactorings.  Most remaining lines
of excessive length in merge-recursive include error messages and it's
not clear that splitting those improves things.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 724dd767b2 cache-tree: share code between functions writing an index as a tree
write_tree_from_memory() appeared to be a merge-recursive special that
basically duplicated write_index_as_tree().  The two have a different
signature, but the bigger difference was just that write_index_as_tree()
would always unconditionally read the index off of disk instead of
working on the current in-memory index.  So:

  * split out common code into write_index_as_tree_internal()

  * rename write_tree_from_memory() to write_inmemory_index_as_tree(),
    make it call write_index_as_tree_internal(), and move it to
    cache-tree.c

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 345480d1ed merge-recursive: don't force external callers to do our logging
Alternatively, you can view this as "make the merge functions behave
more similarly."  merge-recursive has three different entry points:
merge_trees(), merge_recursive(), and merge_recursive_generic().  Two of
these would call diff_warn_rename_limit(), but merge_trees() didn't.
This lead to callers of merge_trees() needing to manually call
diff_warn_rename_limit() themselves.  Move this to the new
merge_finalize() function to make sure that all three entry points run
this function.

Note that there are two external callers of merge_trees(), one in
sequencer.c and one in builtin/checkout.c.  The one in sequencer.c is
cleaned up by this patch and just transfers where the call to
diff_warn_rename_limit() is made; the one in builtin/checkout.c is for
switching to a different commit and in the very rare case where the
warning might be triggered, it would probably be helpful to include
(e.g. if someone is modifying a file that has been renamed in moving to
the other commit, but there are so many renames between the commits that
the limit kicks in and none are detected, it may help to have an
explanation about why they got a delete/modify conflict instead of a
proper content merge in a renamed file).

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren b4db8a2b76 merge-recursive: remove useless parameter in merge_trees()
merge_trees() took a results parameter that would only be written when
opt->call_depth was positive, which is never the case now that
merge_trees_internal() has been split from merge_trees().  Remove the
misleading and unused parameter from merge_trees().

While at it, add some comments explaining how the output of
merge_trees() and merge_recursive() differ.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 98a1d3d888 merge-recursive: exit early if index != head
We had a rule to enforce that the index matches head, but it was found
at the beginning of merge_trees() and would only trigger when
opt->call_depth was 0.  Since merge_recursive() doesn't call
merge_trees() until after returning from recursing, this meant that the
check wasn't triggered by merge_recursive() until it had first finished
all the intermediate merges to create virtual merge bases.  That is a
potentially huge amount of computation (and writing of intermediate
merge results into the .git/objects directory) before it errors out and
says, in effect, "Sorry, I can't do any merging because you have some
local changes that would be overwritten."

Further, not enforcing this requirement earlier allowed other bugs (such
as an unintentional unconditional dropping and reloading of the index in
merge_recursive() even when no recursion was necessary), to mask bugs in
other callers (which were fixed in the commit prior to this one).

Make sure we do the index == head check at the beginning of the merge,
and error out immediately if it fails.  While we're at it, fix a small
leak in the show-the-error codepath.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 10f751c06b merge-recursive: remove another implicit dependency on the_repository
Commit d7cf3a96e9 ("merge-recursive.c: remove implicit dependency on
the_repository", 2019-01-12) and follow-ups like commit 34e7771bc6
("Use the right 'struct repository' instead of the_repository",
2019-06-27), removed most implicit uses of the_repository.  Convert
calls to get_commit_tree() to instead use repo_get_commit_tree() to get
rid of another.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren f836bf3937 merge-recursive: future-proof update_file_flags() against memory leaks
There is a 'free_buf' label to which all but one of the error paths in
update_file_flags() jump; that error case involves a NULL buf and is
thus not a memory leak.  However, make that error case execute the same
deallocation code anyway so that if anyone adds any additional memory
allocations or deallocations, then all error paths correctly deallocate
resources.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Derrick Stolee 8e01251694 merge-recursive: introduce an enum for detect_directory_renames values
Improve code readability by introducing an enum to replace the
not-quite-boolean values taken on by detect_directory_renames.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 743474cbfa merge-recursive: provide a better label for diff3 common ancestor
In commit 7ca56aa076 ("merge-recursive: add a label for ancestor",
2010-03-20), a label was added for the '||||||' line to make it have
the more informative heading '|||||| merged common ancestors', with
the statement:

    It would be nicer to use a more informative label.  Perhaps someone
    will provide one some day.

This chosen label was perfectly reasonable when recursiveness kicks in,
i.e. when there are multiple merge bases.  (I can't think of a better
label in such cases.)  But it is actually somewhat misleading when there
is a unique merge base or no merge base.  Change this based on the
number of merge bases:
    >=2: "merged common ancestors"
    1:   <abbreviated commit hash>
    0:   "<empty tree>"

Tests have also been added to check that we get the right ancestor name
for each of the three cases.

Also, since merge_recursive() and merge_trees() have polar opposite
pre-conditions for opt->ancestor, document merge_recursive()'s
pre-condition with an assertion.  (An assertion was added to
merge_trees() already a few commits ago.)  The differences in
pre-conditions stem from two factors: (1) merge_trees() does not recurse
and thus does not have multiple sub-merges to worry about -- each of
which would require a different value for opt->ancestor, (2)
merge_trees() is only passed trees rather than commits and thus cannot
internally guess as good of a label.  Thus, while external callers of
merge_trees() are required to provide a non-NULL opt->ancestor,
merge_recursive() expects to set this value itself.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-19 10:08:03 -07:00
Elijah Newren 139ef37a2f merge-recursive: enforce opt->ancestor != NULL when calling merge_trees()
We always want our conflict hunks to be labelled so that users can know
where each came from.  The previous commit fixed the one caller in the
codebase which was not setting opt->ancestor (and thus not providing a
label for the "merge base" conflict hunk in diff3-style conflict
markers); add an assertion to prevent future codepaths from also
overlooking this requirement.

Enforcing this requirement also allows us to simplify the code for
labelling the conflict hunks by no longer checking if the ancestor label
is NULL.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-16 12:47:20 -07:00
Elijah Newren d8523ca1b9 merge-recursive: be consistent with assert
In commit 8daec1df03 ("merge-recursive: switch from (oid,mode) pairs
to a diff_filespec", 2019-04-05), an assertion on a->path && b->path
was added for code readability to document that these both needed to be
non-NULL at this point in the code.  However, the subsequent lines also
read o->path, so it should be included in the assert.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-16 12:47:20 -07:00
Junio C Hamano 0bdaec1d3d Merge branch 'en/disable-dir-rename-in-recursive-merge'
"merge-recursive" hit a BUG() when building a virtual merge base
detected a directory rename.

* en/disable-dir-rename-in-recursive-merge:
  merge-recursive: avoid directory rename detection in recursive case
2019-08-08 14:26:10 -07:00
Elijah Newren ff6d54771a merge-recursive: avoid directory rename detection in recursive case
Ever since commit 8c8e5bd6eb ("merge-recursive: switch directory
rename detection default", 2019-04-05), the default handling with
directory rename detection was to report a conflict and leave unstaged
entries in the index.  However, when creating a virtual merge base in
the recursive case, we absolutely need a tree, and the only way a tree
can be written is if we have no unstaged entries -- otherwise we hit a
BUG().

There are a few fixes possible here which at least fix the BUG(), but
none of them seem optimal for other reasons; see the comments with the
new testcase 13e in t6043 for details (which testcase triggered a BUG()
prior to this patch).  As such, just opt for a very conservative and
simple choice that is still relatively reasonable: have the recursive
case treat 'conflict' as 'false' for opt->detect_directory_renames.

Reported-by: Emily Shaffer <emilyshaffer@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-08-06 10:42:36 -07:00
Nguyễn Thái Ngọc Duy 34e7771bc6 Use the right 'struct repository' instead of the_repository
There are a couple of places where 'struct repository' is already passed
around, but the_repository is still used. Use the right repo.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-06-27 12:45:17 -07:00
Nguyễn Thái Ngọc Duy 90d3405196 match-trees.c: remove the_repo from shift_tree*()
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-06-27 12:45:17 -07:00
Nguyễn Thái Ngọc Duy 50ddb089ff tree-walk.c: remove the_repo from get_tree_entry()
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-06-27 12:45:17 -07:00
Junio C Hamano 20fbf7dd42 Merge branch 'en/merge-directory-renames-fix'
Recent code restructuring of merge-recursive engine introduced a
regression dealing with rename/add conflict.

* en/merge-directory-renames-fix:
  merge-recursive: restore accidentally dropped setting of path
2019-06-06 14:03:36 -07:00
Elijah Newren 481de8a293 merge-recursive: restore accidentally dropped setting of path
In commit 8daec1df03 ("merge-recursive: switch from (oid,mode) pairs
to a diff_filespec", 2019-04-05), we actually switched from
(oid,mode,path) triplets to a diff_filespec -- but most callsites in the
patch only needed to worry about oid and mode so the commit message
focused on that.  The oversight in the commit message apparently spilled
over to the code as well; one of the dozen or so callsites accidentally
dropped the setting of the path in the conversion.  Restore the path
setting in that location.

Also, this pointed out that our testsuite was lacking a good rename/add
test, at least one that involved the need for merge content with the
rename.  Add such a test, and since rename/add vs. add/rename could
possibly be important, redo the merge the opposite direction to make
sure we don't have issues with the direction of the merge.  These
testcases failed before restoring the setting of path, but with the
paths appropriately set the testcases both pass.

Reported-by: Ben Humphreys <behumphreys@atlassian.com>
Based-on-patch-by: SZEDER Gábor <szeder.dev@gmail.com>
Tested-by: Ben Humphreys <behumphreys@atlassian.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-06-05 09:30:40 -07:00
Junio C Hamano 0b179f3175 Merge branch 'nd/sha1-name-c-wo-the-repository'
Further code clean-up to allow the lowest level of name-to-object
mapping layer to work with a passed-in repository other than the
default one.

* nd/sha1-name-c-wo-the-repository: (34 commits)
  sha1-name.c: remove the_repo from get_oid_mb()
  sha1-name.c: remove the_repo from other get_oid_*
  sha1-name.c: remove the_repo from maybe_die_on_misspelt_object_name
  submodule-config.c: use repo_get_oid for reading .gitmodules
  sha1-name.c: add repo_get_oid()
  sha1-name.c: remove the_repo from get_oid_with_context_1()
  sha1-name.c: remove the_repo from resolve_relative_path()
  sha1-name.c: remove the_repo from diagnose_invalid_index_path()
  sha1-name.c: remove the_repo from handle_one_ref()
  sha1-name.c: remove the_repo from get_oid_1()
  sha1-name.c: remove the_repo from get_oid_basic()
  sha1-name.c: remove the_repo from get_describe_name()
  sha1-name.c: remove the_repo from get_oid_oneline()
  sha1-name.c: add repo_interpret_branch_name()
  sha1-name.c: remove the_repo from interpret_branch_mark()
  sha1-name.c: remove the_repo from interpret_nth_prior_checkout()
  sha1-name.c: remove the_repo from get_short_oid()
  sha1-name.c: add repo_for_each_abbrev()
  sha1-name.c: store and use repo in struct disambiguate_state
  sha1-name.c: add repo_find_unique_abbrev_r()
  ...
2019-05-09 00:37:25 +09:00
Junio C Hamano 96379f043f Merge branch 'en/merge-directory-renames'
"git merge-recursive" backend recently learned a new heuristics to
infer file movement based on how other files in the same directory
moved.  As this is inherently less robust heuristics than the one
based on the content similarity of the file itself (rather than
based on what its neighbours are doing), it sometimes gives an
outcome unexpected by the end users.  This has been toned down to
leave the renamed paths in higher/conflicted stages in the index so
that the user can examine and confirm the result.

* en/merge-directory-renames:
  merge-recursive: switch directory rename detection default
  merge-recursive: give callers of handle_content_merge() access to contents
  merge-recursive: track information associated with directory renames
  t6043: fix copied test description to match its purpose
  merge-recursive: switch from (oid,mode) pairs to a diff_filespec
  merge-recursive: cleanup handle_rename_* function signatures
  merge-recursive: track branch where rename occurred in rename struct
  merge-recursive: remove ren[12]_other fields from rename_conflict_info
  merge-recursive: shrink rename_conflict_info
  merge-recursive: move some struct declarations together
  merge-recursive: use 'ci' for rename_conflict_info variable name
  merge-recursive: rename locals 'o' and 'a' to 'obuf' and 'abuf'
  merge-recursive: rename diff_filespec 'one' to 'o'
  merge-recursive: rename merge_options argument from 'o' to 'opt'
  Use 'unsigned short' for mode, like diff_filespec does
2019-05-09 00:37:22 +09:00
Nguyễn Thái Ngọc Duy a133c40b23 commit.cocci: refactor code, avoid double rewrite
"maybe" pointer in 'struct commit' is tricky because it can be lazily
initialized to take advantage of commit-graph if available. This makes
it not safe to access directly.

This leads to a rule in commit.cocci to rewrite 'x->maybe_tree' to
'get_commit_tree(x)'. But that rule alone could lead to incorrectly
rewrite assignments, e.g. from

    x->maybe_tree = yes

to

    get_commit_tree(x) = yes

Because of this we have a second rule to revert this effect. Szeder
found out that we could do better by performing the assignment rewrite
rule first, then the remaining is read-only access and handled by the
current first rule.

For this to work, we need to transform "x->maybe_tree = y" to something
that does NOT contain "x->maybe_tree" to avoid the original first
rule. This is where set_commit_tree() comes in.

Helped-by: SZEDER Gábor <szeder.dev@gmail.com>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-16 18:56:51 +09:00
Elijah Newren 8c8e5bd6eb merge-recursive: switch directory rename detection default
When all of x/a, x/b, and x/c have moved to z/a, z/b, and z/c on one
branch, there is a question about whether x/d added on a different
branch should remain at x/d or appear at z/d when the two branches are
merged.  There are different possible viewpoints here:

  A) The file was placed at x/d; it's unrelated to the other files in
     x/ so it doesn't matter that all the files from x/ moved to z/ on
     one branch; x/d should still remain at x/d.

  B) x/d is related to the other files in x/, and x/ was renamed to z/;
     therefore x/d should be moved to z/d.

Since there was no ability to detect directory renames prior to
git-2.18, users experienced (A) regardless of context.  Choice (B) was
implemented in git-2.18, with no option to go back to (A), and has been
in use since.  However, one user reported that the merge results did not
match their expectations, making the change of default problematic,
especially since there was no notice printed when directory rename
detection moved files.

Note that there is also a third possibility here:

  C) There are different answers depending on the context and content
     that cannot be determined by git, so this is a conflict.  Use a
     higher stage in the index to record the conflict and notify the
     user of the potential issue instead of silently selecting a
     resolution for them.

Add an option for users to specify their preference for whether to use
directory rename detection, and default to (C).  Even when directory
rename detection is on, add notice messages about files moved into new
directories.

As a sidenote, x/d did not have to be a new file here; it could have
already existed at some other path and been renamed to x/d, with
directory rename detection just renaming it again to z/d.  Thus, it's
not just new files, but also a modification to all rename types (normal
renames, rename/add, rename/delete, rename/rename(1to1),
rename/rename(1to2), and rename/rename(2to1)).

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:08 +09:00
Elijah Newren e62d11239c merge-recursive: give callers of handle_content_merge() access to contents
Pass a merge_file_info struct to handle_content_merge() so that the
callers can access the oid and mode of the result afterward.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:08 +09:00
Elijah Newren 6d169fd321 merge-recursive: track information associated with directory renames
Directory rename detection previously silently applied.  In order to
allow printing information about paths that changed or printing a
conflict notification (and only doing so near other potential conflict
messages associated with the paths), save this information inside the
rename struct for later use.  A subsequent patch will make use of the
additional information.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:08 +09:00
Elijah Newren 8daec1df03 merge-recursive: switch from (oid,mode) pairs to a diff_filespec
There was a significant inconsistency in the various parts of the API
used in merge-recursive; many places used a pair of (oid, mode) to track
file version/contents, while other parts used a diff_filespec (which
have an oid and mode embedded in it).  This inconsistency caused lots of
places to need to pack and unpack data to call into other functions.
This has been the subject of some past cleanups (see e.g. commit
0270a07ad0 ("merge-recursive: remove final remaining caller of
merge_file_one()", 2018-09-19)), but let's just remove the underlying
mess altogether by switching to use diff_filespec.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren e2d563dfa9 merge-recursive: cleanup handle_rename_* function signatures
Instead of passing various bits and pieces of 'ci', just pass it
directly.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren c336ab8593 merge-recursive: track branch where rename occurred in rename struct
We previously tracked the branch associated with a rename in a separate
field in rename_conflict_info, but since it is directly associated with
the rename it makes more sense to move it into the rename struct.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 3f9c92ec99 merge-recursive: remove ren[12]_other fields from rename_conflict_info
The ren1_other and ren2_other fields were synthesized from information
in ren1->src_entry and ren2->src_entry.  Since we already have the
necessary information in ren1 and ren2, just use those.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren e9cd1b5ca4 merge-recursive: shrink rename_conflict_info
The rename_conflict_info struct used both a pair and a stage_data which
were taken from a rename struct.  Just use the original rename struct.
This will also allow us to start making other simplifications to the
code.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 967d6be725 merge-recursive: move some struct declarations together
These structs are related and reference each other, so move them
together to make it easier for folks to determine what they hold and
what their purpose is.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 043622b2e9 merge-recursive: use 'ci' for rename_conflict_info variable name
We used a couple different names, but used 'ci' the most.  Use the same
variable name throughout for a little extra consistency.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 93a02c5553 merge-recursive: rename locals 'o' and 'a' to 'obuf' and 'abuf'
Since we want to replace oid,mode pairs with a single diff_filespec,
we will soon want to be able to use the names 'o', 'a', and 'b' for
the three different file versions.  Rename some local variables in
blob_unchanged() that would otherwise conflict.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren e3de888ca0 merge-recursive: rename diff_filespec 'one' to 'o'
In the previous commit, we noted that several places throughout merge
recursive both had a reason to use 'o'; some for a merge_options struct,
and others for a diff_filespec struct.  Some places had both, forcing
one of the two to be renamed, though the choice was inconsistent.  Now
that the merge_options struct has been renamed to 'opt' everywhere, we
can replace the few places that used 'one' for the diff_filespec to 'o'.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 259ccb6cc3 merge-recursive: rename merge_options argument from 'o' to 'opt'
The name 'o' was used for the merge_options struct pointer taken by many
functions, but in a few places it was named 'opt'.  Several functions
that didn't need merge_options instead used 'o' for a diff_filespec
argument or local.  Some functions needed both an inconsistently either
renamed the merge_options to 'opt' or the diff_filespec to 'one'.  I
want to remove the weird split in the codebase between using a
diff_filespec and a pair of (oid,mode) values in favor of using a
diff_filespec everywhere, but that dramatically increases the number of
cases where we want to use 'o' as a diff_filespec.  Rename the
merge_options argument to 'opt' to make room.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
Elijah Newren 5ec1e72823 Use 'unsigned short' for mode, like diff_filespec does
struct diff_filespec defines mode to be an 'unsigned short'.  Several
other places in the API which we'd like to interact with using a
diff_filespec used a plain unsigned (or unsigned int).  This caused
problems when taking addresses, so switch to unsigned short.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-08 16:02:07 +09:00
brian m. carlson db1ba2a230 submodule: avoid hard-coded constants
Instead of using hard-coded 40-based constants, express these values in
terms of the_hash_algo and GIT_MAX_HEXSZ.

Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-04-01 11:57:37 +09:00
Junio C Hamano b0e7fb2e5c Merge branch 'nd/completion-more-parameters'
The command line completion (in contrib/) has been taught to
complete more subcommand parameters.

* nd/completion-more-parameters:
  completion: add more parameter value completion
2019-03-07 09:59:58 +09:00
Nguyễn Thái Ngọc Duy 5a59a2301f completion: add more parameter value completion
This adds value completion for a couple more paramters. To make it
easier to maintain these hard coded lists, add a comment at the original
list/code to remind people to update git-completion.bash too.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-02-20 12:31:56 -08:00
Jeff King b53c502807 merge-recursive: drop several unused parameters
There are a few functions related to directory renames that have unused
parameters. After consulting with the author in [1], these seem to be
leftover cruft from the development process, and not signs of any bug.
Let's drop them.

[1] https://public-inbox.org/git/CABPp-BHobf8wbBsXF97scNQCzkxQukziODRXq6JOOWq61cAd9g@mail.gmail.com/

Helped-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-02-14 15:26:15 -08:00
Junio C Hamano 7589e63648 Merge branch 'nd/the-index-final'
The assumption to work on the single "in-core index" instance has
been reduced from the library-ish part of the codebase.

* nd/the-index-final:
  cache.h: flip NO_THE_INDEX_COMPATIBILITY_MACROS switch
  read-cache.c: remove the_* from index_has_changes()
  merge-recursive.c: remove implicit dependency on the_repository
  merge-recursive.c: remove implicit dependency on the_index
  sha1-name.c: remove implicit dependency on the_index
  read-cache.c: replace update_index_if_able with repo_&
  read-cache.c: kill read_index()
  checkout: avoid the_index when possible
  repository.c: replace hold_locked_index() with repo_hold_locked_index()
  notes-utils.c: remove the_repository references
  grep: use grep_opt->repo instead of explict repo argument
2019-02-06 22:05:23 -08:00
Junio C Hamano d6f05a435f Merge branch 'nd/attr-pathspec-in-tree-walk'
The traversal over tree objects has learned to honor
":(attr:label)" pathspec match, which has been implemented only for
enumerating paths on the filesystem.

* nd/attr-pathspec-in-tree-walk:
  tree-walk: support :(attr) matching
  dir.c: move, rename and export match_attrs()
  pathspec.h: clean up "extern" in function declarations
  tree-walk.c: make tree_entry_interesting() take an index
  tree.c: make read_tree*() take 'struct repository *'
2019-01-14 15:29:28 -08:00
Nguyễn Thái Ngọc Duy 150fe065f7 read-cache.c: remove the_* from index_has_changes()
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-01-14 12:13:05 -08:00
Nguyễn Thái Ngọc Duy d7cf3a96e9 merge-recursive.c: remove implicit dependency on the_repository
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-01-14 12:13:04 -08:00
Nguyễn Thái Ngọc Duy 0d6caa2d08 merge-recursive.c: remove implicit dependency on the_index
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-01-14 12:13:04 -08:00
Nguyễn Thái Ngọc Duy e1ff0a32e4 read-cache.c: kill read_index()
read_index() shares the same problem as hold_locked_index(): it
assumes $GIT_DIR/index. Move all call sites to repo_read_index()
instead. read_index_preload() and read_index_unmerged() are also
killed as a consequence.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-01-14 12:13:04 -08:00
Nguyễn Thái Ngọc Duy 3a95f31d1c repository.c: replace hold_locked_index() with repo_hold_locked_index()
hold_locked_index() assumes the index path at $GIT_DIR/index. This is
not good for places that take an arbitrary index_state instead of
the_index, which is basically everywhere except builtin/.

Replace it with repo_hold_locked_index(). hold_locked_index() remains
as a wrapper around repo_hold_locked_index() to reduce changes in builtin/

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-01-14 12:13:04 -08:00
Junio C Hamano ac193e0e0a Merge branch 'en/merge-path-collision'
Updates for corner cases in merge-recursive.

* en/merge-path-collision:
  t6036: avoid non-portable "cp -a"
  merge-recursive: combine error handling
  t6036, t6043: increase code coverage for file collision handling
  merge-recursive: improve rename/rename(1to2)/add[/add] handling
  merge-recursive: use handle_file_collision for add/add conflicts
  merge-recursive: improve handling for rename/rename(2to1) conflicts
  merge-recursive: fix rename/add conflict handling
  merge-recursive: new function for better colliding conflict resolutions
  merge-recursive: increase marker length with depth of recursion
  t6036, t6042: testcases for rename collision of already conflicting files
  t6042: add tests for consistency in file collision conflict handling
2019-01-04 13:33:32 -08:00
Nguyễn Thái Ngọc Duy e092073d64 tree.c: make read_tree*() take 'struct repository *'
These functions call tree_entry_interesting() which will soon require
a 'struct index_state *' to be passed in. Instead of just changing the
function signature to take an index, update to take a repo instead
because these functions do need object database access.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-19 10:50:33 +09:00
Derrick Stolee 80cee6e321 merge-recursive: combine error handling
In handle_rename_rename_1to2(), we have duplicated error handling
around colliding paths. Specifically, when we want to write out
the file and there is a directory or untracked file in the way,
we need to create a temporary file to hold the contents. This has
some special output to alert the user, and this output is
duplicated for each side of the conflict.

Simplify the call by generating this new path in a helper
function.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:54 +09:00
Elijah Newren 48c9cb9d6d merge-recursive: improve rename/rename(1to2)/add[/add] handling
When we have a rename/rename(1to2) conflict, each of the renames can
collide with a file addition.  Each of these rename/add conflicts suffered
from the same kinds of problems that normal rename/add suffered from.
Make the code use handle_file_conflicts() as well so that we get all the
same fixes and consistent behavior between the different conflict types.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Elijah Newren dcf2815098 merge-recursive: use handle_file_collision for add/add conflicts
This results in no-net change of behavior, it simply ensures that all
file-collision conflict handling types are being handled the same by
calling the same function.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Elijah Newren bbafc9c44a merge-recursive: improve handling for rename/rename(2to1) conflicts
This makes the rename/rename(2to1) conflicts use the new
handle_file_collision() function.  Since that function was based
originally on the rename/rename(2to1) handling code, the main
differences here are in what was added.  In particular:

  * Instead of storing files at collide_path~HEAD and collide_path~MERGE,
    the files are two-way merged and recorded at collide_path.

  * Instead of recording the version of the renamed file that existed
    on the renamed side in the index (thus ignoring any changes that
    were made to the file on the side of history without the rename),
    we do a three-way content merge on the renamed path, then store
    that at either stage 2 or stage 3.

  * Note that since the content merge for each rename may have conflicts,
    and then we have to merge the two renamed files, we can end up with
    nested conflict markers.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Elijah Newren 7f8671656f merge-recursive: fix rename/add conflict handling
This makes the rename/add conflict handling make use of the new
handle_file_collision() function, which fixes several bugs and improves
things for the rename/add case significantly.  Previously, rename/add
would:

  * Not leave any higher order stage entries in the index, making it
    appear as if there were no conflict.
  * Would place the rename file at the colliding path, and move the
    added file elsewhere, which combined with the lack of higher order
    stage entries felt really odd.  It's not clear to me why the
    rename should take precedence over the add; if one should be moved
    out of the way, they both probably should.
  * In the recursive case, it would do a two way merge of the added
    file and the version of the renamed file on the renamed side,
    completely excluding modifications to the renamed file on the
    unrenamed side of history.

Use the new handle_file_collision() to fix all of these issues.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Elijah Newren 37b65ce36b merge-recursive: new function for better colliding conflict resolutions
There are three conflict types that represent two (possibly entirely
unrelated) files colliding at the same location:
  * add/add
  * rename/add
  * rename/rename(2to1)

These three conflict types already share more similarity than might be
immediately apparent from their description: (1) the handling of the
rename variants already involves removing any entries from the index
corresponding to the original file names[*], thus only leaving entries
in the index for the colliding path; (2) likewise, any trace of the
original file name in the working tree is also removed.  So, in all
three cases we're left with how to represent two colliding files in both
the index and the working copy.

[*] Technically, this isn't quite true because rename/rename(2to1)
conflicts in the recursive (o->call_depth > 0) case do an "unrename"
since about seven years ago.  But even in that case, Junio felt
compelled to explain that my decision to "unrename" wasn't necessarily
the only or right answer -- search for "Comment from Junio" in t6036 for
details.

My initial motivation for looking at these three conflict types was that
if the handling of these three conflict types is the same, at least in
the limited set of cases where a renamed file is unmodified on the side
of history where the file is not renamed, then a significant performance
improvement for rename detection during merges is possible.  However,
while that served as motivation to look at these three types of
conflicts, the actual goal of this new function is to try to improve the
handling for all three cases, not to merely make them the same as each
other in that special circumstance.

=== Handling the working tree ===

The previous behavior for these conflict types in regards to the
working tree (assuming the file collision occurs at 'foo') was:
  * add/add does a two-way merge of the two files and records it as 'foo'.
  * rename/rename(2to1) records the two different files into two new
    uniquely named files (foo~HEAD and foo~$MERGE), while removing 'foo'
    from the working tree.
  * rename/add records the two different files into two different
    locations, recording the add at foo~$SIDE and, oddly, recording
    the rename at foo (why is the rename more important than the add?)

So, the question for what to write to the working tree boils down to
whether the two colliding files should be two-way merged and recorded in
place, or recorded into separate files.  As per discussion on the git
mailing lit, two-way merging was deemed to always be preferred, as that
makes these cases all more like content conflicts that users can handle
from within their favorite editor, IDE, or merge tool.  Note that since
renames already involve a content merge, rename/add and
rename/rename(2to1) conflicts could result in nested conflict markers.

=== Handling of the index ===

For a typical rename, unpack_trees() would set up the index in the
following fashion:
           old_path  new_path
   stage1: 5ca1ab1e  00000000
   stage2: f005ba11  00000000
   stage3: 00000000  b0a710ad
And merge-recursive would rewrite this to
           new_path
   stage1: 5ca1ab1e
   stage2: f005ba11
   stage3: b0a710ad
Removing old_path from the index means the user won't have to `git rm
old_path` manually every time a renamed path has a content conflict.
It also means they can use `git checkout [--ours|--theirs|--conflict|-m]
new_path`, `git diff [--ours|--theirs]` and various other commands that
would be difficult otherwise.

This strategy becomes a problem when we have a rename/add or
rename/rename(2to1) conflict, however, because then we have only three
slots to store blob sha1s and we need either four or six.  Previously,
this was handled by continuing to delete old_path from the index, and
just outright ignoring any blob shas from old_path.  That had the
downside of deleting any trace of changes made to old_path on the other
side of history.  This function instead does a three-way content merge of
the renamed file, and stores the blob sha1 for that at either stage2 or
stage3 for new_path (depending on which side the rename came from).  That
has the advantage of bringing information about changes on both sides and
still allows for easy resolution (no need to git rm old_path, etc.), but
does have the downside that if the content merge had conflict markers,
then what we store in the index is the sha1 of a blob with conflict
markers.  While that is a downside, it seems less problematic than the
downsides of any obvious alternatives, and certainly makes more sense
than the previous handling.  Further, it has a precedent in that when we
do recursive merges, we may accept a file with conflict markers as the
resolution for the merge of the merge-bases, which will then show up in
the index of the outer merge at stage 1 if a conflict exists at the outer
level.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Elijah Newren b2a7942b8b merge-recursive: increase marker length with depth of recursion
Later patches in this series will modify file collision conflict
handling (e.g. from rename/add and rename/rename(2to1) conflicts) so
that multiply nested conflict markers can arise even before considering
conflicts in the virtual merge base.  Including the virtual merge base
will provide a way to get triply (or higher) nested conflict markers.
This new way to get nested conflict markers will force the need for a
more general mechanism to extend the length of conflict markers in order
to differentiate between different nestings.

Along with this change to conflict marker length handling, we want to
make sure that we don't regress handling for other types of conflicts
with nested conflict markers.  Add a more involved testcase using
merge.conflictstyle=diff3, where not only does the virtual merge base
contain conflicts, but its virtual merge base does as well (i.e. a case
with triply nested conflict markers).  While there are multiple
reasonable ways to handle nested conflict markers in the virtual merge
base for this type of situation, the easiest approach that dovetails
well with the new needs for the file collision conflict handling is to
require that the length of the conflict markers increase with each
subsequent nesting.

Subsequent patches which change the rename/add and rename/rename(2to1)
conflict handling will modify the extra_marker_size flag appropriately
for their new needs.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-08 14:23:53 +09:00
Junio C Hamano ff8e25e99e Merge branch 'en/merge-cleanup-more'
Further clean-up of merge-recursive machinery.

* en/merge-cleanup-more:
  merge-recursive: avoid showing conflicts with merge branch before HEAD
  merge-recursive: improve auto-merging messages with path collisions
2018-11-03 00:53:57 +09:00
Junio C Hamano 11877b9ebe Merge branch 'nd/the-index'
Various codepaths in the core-ish part learn to work on an
arbitrary in-core index structure, not necessarily the default
instance "the_index".

* nd/the-index: (23 commits)
  revision.c: reduce implicit dependency the_repository
  revision.c: remove implicit dependency on the_index
  ws.c: remove implicit dependency on the_index
  tree-diff.c: remove implicit dependency on the_index
  submodule.c: remove implicit dependency on the_index
  line-range.c: remove implicit dependency on the_index
  userdiff.c: remove implicit dependency on the_index
  rerere.c: remove implicit dependency on the_index
  sha1-file.c: remove implicit dependency on the_index
  patch-ids.c: remove implicit dependency on the_index
  merge.c: remove implicit dependency on the_index
  merge-blobs.c: remove implicit dependency on the_index
  ll-merge.c: remove implicit dependency on the_index
  diff-lib.c: remove implicit dependency on the_index
  read-cache.c: remove implicit dependency on the_index
  diff.c: remove implicit dependency on the_index
  grep.c: remove implicit dependency on the_index
  diff.c: remove the_index dependency in textconv() functions
  blame.c: rename "repo" argument to "r"
  combine-diff.c: remove implicit dependency on the_index
  ...
2018-10-19 13:34:02 +09:00
Elijah Newren 4f44545313 merge-recursive: avoid showing conflicts with merge branch before HEAD
We want to load unmerged entries from HEAD into the index at stage 2 and
from MERGE_HEAD into stage 3.  Similarly, folks expect merge conflicts
to look like

    <<<<<<<< HEAD
    content from our side
    ========
    content from their side
    >>>>>>>> MERGE_HEAD

not

    <<<<<<<< MERGE_HEAD
    content from their side
    ========
    content from our side
    >>>>>>>> HEAD

The correct order usually comes naturally and for free, but with renames
we often have data in the form {rename_branch, other_branch}, and
working relative to the rename first (e.g. for rename/add) is more
convenient elsewhere in the code.  Address the slight impedance
mismatch by having some functions re-call themselves with flipped
arguments when the branch order is reversed.

Note that setup_rename_conflict_info() has one asymmetry in it, in
setting dst_entry1->processed=0 but not doing similarly for
dst_entry2->processed.  When dealing with rename/rename and similar
conflicts, we do not want the processing to happen twice, so the
desire to only set one of the entries to unprocessed is intentional.
So, while this change modifies which branch's entry will be marked as
unprocessed, that dovetails nicely with putting HEAD first so that we
get the index stage entries and conflict markers in the right order.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-10-18 14:45:39 +09:00
Elijah Newren 2b168ef3ff merge-recursive: improve auto-merging messages with path collisions
Each individual file involved in a rename could have also been modified
on both sides of history, meaning it may need to have content merges.
If two such files are renamed into the same location, then on top of the
two natural auto-merging messages we also have to two-way merge the
result, giving us messages that look like

  Auto-merging somefile.c (was somecase.c)
  Auto-merging somefile.c (was somefolder.c)
  Auto-merging somefile.c

However, despite the fact that I was the one who put the "(was %s)"
portions into the messages (and just a few months ago), I was still
initially confused when running into a rename/rename(2to1) case and
wondered if somefile.c had been merged three times.  Update this to
instead be:

  Auto-merging version of somefile.c from somecase.c
  Auto-merging version of somefile.c from someportfolio.c
  Auto-merging somefile.c

This is an admittedly long set of messages for a single path, but you
only get all three messages when dealing with the rare case of a
rename/rename(2to1) conflict where both sides of both original files
were also modified, in conflicting ways.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-10-18 13:54:36 +09:00
Junio C Hamano 74bb46354f Merge branch 'en/merge-cleanup'
Code clean-up.

* en/merge-cleanup:
  merge-recursive: rename merge_file_1() and merge_content()
  merge-recursive: remove final remaining caller of merge_file_one()
  merge-recursive: avoid wrapper function when unnecessary and wasteful
  merge-recursive: set paths correctly when three-way merging content
2018-10-16 16:16:00 +09:00
Junio C Hamano ae109a9789 Merge branch 'en/double-semicolon-fix'
Code clean-up.

* en/double-semicolon-fix:
  Remove superfluous trailing semicolons
2018-09-24 10:30:47 -07:00
Nguyễn Thái Ngọc Duy 2abf350385 revision.c: remove implicit dependency on the_index
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-21 09:51:19 -07:00
Nguyễn Thái Ngọc Duy 32eaa46883 ll-merge.c: remove implicit dependency on the_index
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-21 09:48:10 -07:00
Nguyễn Thái Ngọc Duy e675765235 diff.c: remove implicit dependency on the_index
A new variant repo_diff_setup() is added that takes 'struct repository *'
and diff_setup() becomes a thin macro around it that is protected by
NO_THE_REPOSITORY_COMPATIBILITY_MACROS, similar to NO_THE_INDEX_....
The plan is these macros will always be defined for all library files
and the macros are only accessible in builtin/

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-21 09:48:10 -07:00
Elijah Newren d957355675 merge-recursive: rename merge_file_1() and merge_content()
Summary:
  merge_file_1()  -> merge_mode_and_contents()
  merge_content() -> handle_content_merge()

merge_file_1() is a very unhelpful name.  Rename it to
merge_mode_and_contents() to reflect what it does.

merge_content() calls merge_mode_and_contents() to do the main part of
its work, but most of this function was about higher level stuff, e.g.
printing out conflict messages, updating skip_worktree bits, checking
for ability to avoid updating the working tree or for D/F conflicts
being in the way, etc.  Since there are several handle_*() functions for
similar levels of checking and handling in merge-recursive.c (e.g.
handle_change_delete(), handle_rename_rename_2to1()), let's rename this
function to handle_content_merge().

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-20 12:33:57 -07:00
Elijah Newren 0270a07ad0 merge-recursive: remove final remaining caller of merge_file_one()
The function names merge_file_one() and merge_file_1() aren't
particularly intuitive function names, especially since there is no
associated merge_file() function that these are related to.  The
previous commit showed that merge_file_one() was prone to be called when
merge_file_1() should be, and since it is just a thin wrapper around
merge_file_1() anyway and only has one caller left, let's just remove
merge_file_one() entirely.

(It also turns out that the one remaining caller of merge_file_one()
has very broken code that needs to be completely rewritten, but that's
the subject of a future patch series; for now, we're just translating
it into a merge_file_1() call.)

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-20 12:33:57 -07:00
Elijah Newren 75f3fa79c3 merge-recursive: avoid wrapper function when unnecessary and wasteful
merge_file_one() is a convenience function taking a bunch of oids and
modes, combining each pair into a diff_filespec, and then calling
merge_file_1().  When we already start with diff_filespec's, we can
just call merge_file_1() directly instead of splitting out the oids
and modes for the wrapper to recombine into what we already had.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-20 12:33:57 -07:00
Elijah Newren 52396e1d3d merge-recursive: set paths correctly when three-way merging content
merge_3way() has code to mark different sides of the conflict with info
about where the content comes from.  If the names of the files involved
match, it simply uses the branch name.  If the names of the files do not
match, it uses branchname:filename.  Unfortunately, merge_content()
previously always called it with one.path = a.path = b.path.  Granted,
it didn't have other path information available to it for years, but
that was corrected by passing rename_conflict_info in commit
3c217c077a ("merge-recursive: Provide more info in conflict markers
with file renames", 2011-08-11).  In that commit, instead of just fixing
the bug with the pathnames, it created fake branch names incorporating
both the branch name and file name.

This "fake branch" workaround was extended further when I pulled that
logic out into a special function in commit dac4741554
("merge-recursive: Create function for merging with branchname:file
markers", 2011-08-11), and a number of other sites outside of
merge_content() have been added which call into that.  However, this
Rube-Goldberg-esque setup is not merely duplicate code and unnecessary
work, it also risked having other callsites invoke it in a way that
would result in markers of the form branchname:filename:filename (i.e.
with the filename repeated).

Fix this whole mess by:
  - setting one.path, a.path, and b.path appropriately
  - calling merge_file_1() directly
  - deleting the merge_file_special_markers() workaround wrapper

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-20 12:33:57 -07:00
Junio C Hamano 769af0fd9e Merge branch 'jk/cocci'
spatch transformation to replace boolean uses of !hashcmp() to
newly introduced oideq() is added, and applied, to regain
performance lost due to support of multiple hash algorithms.

* jk/cocci:
  show_dirstat: simplify same-content check
  read-cache: use oideq() in ce_compare functions
  convert hashmap comparison functions to oideq()
  convert "hashcmp() != 0" to "!hasheq()"
  convert "oidcmp() != 0" to "!oideq()"
  convert "hashcmp() == 0" to hasheq()
  convert "oidcmp() == 0" to oideq()
  introduce hasheq() and oideq()
  coccinelle: use <...> for function exclusion
2018-09-17 13:53:57 -07:00
Junio C Hamano 1b7a91da71 Merge branch 'ds/reachable'
The code for computing history reachability has been shuffled,
obtained a bunch of new tests to cover them, and then being
improved.

* ds/reachable:
  commit-reach: correct accidental #include of C file
  commit-reach: use can_all_from_reach
  commit-reach: make can_all_from_reach... linear
  commit-reach: replace ref_newer logic
  test-reach: test commit_contains
  test-reach: test can_all_from_reach_with_flags
  test-reach: test reduce_heads
  test-reach: test get_merge_bases_many
  test-reach: test is_descendant_of
  test-reach: test in_merge_bases
  test-reach: create new test tool for ref_newer
  commit-reach: move can_all_from_reach_with_flags
  upload-pack: generalize commit date cutoff
  upload-pack: refactor ok_to_give_up()
  upload-pack: make reachable() more generic
  commit-reach: move commit_contains from ref-filter
  commit-reach: move ref_newer from remote.c
  commit.h: remove method declarations
  commit-reach: move walk methods from commit.c
2018-09-17 13:53:52 -07:00
Elijah Newren c3b9bc94b9 Remove superfluous trailing semicolons
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-09-05 10:21:24 -07:00
Junio C Hamano ca676b9bd3 Merge branch 'en/directory-renames-nothanks'
Recent addition of "directory rename" heuristics to the
merge-recursive backend makes the command susceptible to false
positives and false negatives.  In the context of "git am -3",
which does not know about surrounding unmodified paths and thus
cannot inform the merge machinery about the full trees involved,
this risk is particularly severe.  As such, the heuristic is
disabled for "git am -3" to keep the machinery "more stupid but
predictable".

* en/directory-renames-nothanks:
  am: avoid directory rename detection when calling recursive merge machinery
  merge-recursive: add ability to turn off directory rename detection
  t3401: add another directory rename testcase for rebase and am
2018-09-04 14:31:38 -07:00
Elijah Newren 5fdddd9b75 merge-recursive: add ability to turn off directory rename detection
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-30 07:58:59 -07:00
Jeff King 4a7e27e957 convert "oidcmp() == 0" to oideq()
Using the more restrictive oideq() should, in the long run,
give the compiler more opportunities to optimize these
callsites. For now, this conversion should be a complete
noop with respect to the generated code.

The result is also perhaps a little more readable, as it
avoids the "zero is equal" idiom. Since it's so prevalent in
C, I think seasoned programmers tend not to even notice it
anymore, but it can sometimes make for awkward double
negations (e.g., we can drop a few !!oidcmp() instances
here).

This patch was generated almost entirely by the included
coccinelle patch. This mechanical conversion should be
completely safe, because we check explicitly for cases where
oidcmp() is compared to 0, which is what oideq() is doing
under the hood. Note that we don't have to catch "!oidcmp()"
separately; coccinelle's standard isomorphisms make sure the
two are treated equivalently.

I say "almost" because I did hand-edit the coccinelle output
to fix up a few style violations (it mostly keeps the
original formatting, but sometimes unwraps long lines).

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-29 11:32:49 -07:00
Junio C Hamano dc0f6f9e1d Merge branch 'nd/no-the-index'
The more library-ish parts of the codebase learned to work on the
in-core index-state instance that is passed in by their callers,
instead of always working on the singleton "the_index" instance.

* nd/no-the-index: (24 commits)
  blame.c: remove implicit dependency on the_index
  apply.c: remove implicit dependency on the_index
  apply.c: make init_apply_state() take a struct repository
  apply.c: pass struct apply_state to more functions
  resolve-undo.c: use the right index instead of the_index
  archive-*.c: use the right repository
  archive.c: avoid access to the_index
  grep: use the right index instead of the_index
  attr: remove index from git_attr_set_direction()
  entry.c: use the right index instead of the_index
  submodule.c: use the right index instead of the_index
  pathspec.c: use the right index instead of the_index
  unpack-trees: avoid the_index in verify_absent()
  unpack-trees: convert clear_ce_flags* to avoid the_index
  unpack-trees: don't shadow global var the_index
  unpack-trees: add a note about path invalidation
  unpack-trees: remove 'extern' on function declaration
  ls-files: correct index argument to get_convert_attr_ascii()
  preload-index.c: use the right index instead of the_index
  dir.c: remove an implicit dependency on the_index in pathspec code
  ...
2018-08-20 11:33:53 -07:00
Junio C Hamano e4095da40e Merge branch 'en/merge-recursive-skip-fix'
When the sparse checkout feature is in use, "git cherry-pick" and
other mergy operations lost the skip_worktree bit when a path that
is excluded from checkout requires content level merge, which is
resolved as the same as the HEAD version, without materializing the
merge result in the working tree, which made the path appear as
deleted.  This has been corrected by preserving the skip_worktree
bit (and not materializing the file in the working tree).

* en/merge-recursive-skip-fix:
  merge-recursive: preserve skip_worktree bit when necessary
  t3507: add a testcase showing failure with sparse checkout
2018-08-15 15:08:26 -07:00
Nguyễn Thái Ngọc Duy 7f944e264e convert.c: remove an implicit dependency on the_index
Make the convert API take an index_state instead of assuming the_index
in convert.c. All external call sites are converted blindly to keep
the patch simple and retain current behavior. Individual call sites
may receive further updates to use the right index instead of
the_index.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-13 14:14:42 -07:00
Junio C Hamano c18ac30e9e Merge branch 'en/dirty-merge-fixes'
The recursive merge strategy did not properly ensure there was no
change between HEAD and the index before performing its operation,
which has been corrected.

* en/dirty-merge-fixes:
  merge: fix misleading pre-merge check documentation
  merge-recursive: enforce rule that index matches head before merging
  t6044: add more testcases with staged changes before a merge is invoked
  merge-recursive: fix assumption that head tree being merged is HEAD
  merge-recursive: make sure when we say we abort that we actually abort
  t6044: add a testcase for index matching head, when head doesn't match HEAD
  t6044: verify that merges expected to abort actually abort
  index_has_changes(): avoid assuming operating on the_index
  read-cache.c: move index_has_changes() from merge.c
2018-08-02 15:30:45 -07:00
Junio C Hamano ae533c4a92 Merge branch 'jm/cache-entry-from-mem-pool'
For a large tree, the index needs to hold many cache entries
allocated on heap.  These cache entries are now allocated out of a
dedicated memory pool to amortize malloc(3) overhead.

* jm/cache-entry-from-mem-pool:
  block alloc: add validations around cache_entry lifecyle
  block alloc: allocate cache entries from mem_pool
  mem-pool: fill out functionality
  mem-pool: add life cycle management functions
  mem-pool: only search head block for available space
  block alloc: add lifecycle APIs for cache_entry structs
  read-cache: teach make_cache_entry to take object_id
  read-cache: teach refresh_cache_entry to take istate
2018-08-02 15:30:43 -07:00
Junio C Hamano 3a2a1dc170 Merge branch 'sb/object-store-lookup'
lookup_commit_reference() and friends have been updated to find
in-core object for a specific in-core repository instance.

* sb/object-store-lookup: (32 commits)
  commit.c: allow lookup_commit_reference to handle arbitrary repositories
  commit.c: allow lookup_commit_reference_gently to handle arbitrary repositories
  tag.c: allow deref_tag to handle arbitrary repositories
  object.c: allow parse_object to handle arbitrary repositories
  object.c: allow parse_object_buffer to handle arbitrary repositories
  commit.c: allow get_cached_commit_buffer to handle arbitrary repositories
  commit.c: allow set_commit_buffer to handle arbitrary repositories
  commit.c: migrate the commit buffer to the parsed object store
  commit-slabs: remove realloc counter outside of slab struct
  commit.c: allow parse_commit_buffer to handle arbitrary repositories
  tag: allow parse_tag_buffer to handle arbitrary repositories
  tag: allow lookup_tag to handle arbitrary repositories
  commit: allow lookup_commit to handle arbitrary repositories
  tree: allow lookup_tree to handle arbitrary repositories
  blob: allow lookup_blob to handle arbitrary repositories
  object: allow lookup_object to handle arbitrary repositories
  object: allow object_as_type to handle arbitrary repositories
  tag: add repository argument to deref_tag
  tag: add repository argument to parse_tag_buffer
  tag: add repository argument to lookup_tag
  ...
2018-08-02 15:30:42 -07:00
Elijah Newren 2b75fb601c merge-recursive: preserve skip_worktree bit when necessary
merge-recursive takes any files marked as unmerged by unpack_trees,
tries to figure out whether they can be resolved (e.g. using renames
or a file-level merge), and then if they can be it will delete the old
cache entries and writes new ones.  This means that any ce_flags for
those cache entries are essentially cleared when merging.

Unfortunately, if a file was marked as skip_worktree and it needs a
file-level merge but the merge results in the same version of the file
that was found in HEAD, we skip updating the worktree (because the
file was unchanged) but clear the skip_worktree bit (because of the
delete-cache-entry-and-write-new-one).  This makes git treat the file
as having a local change in the working copy, namely a delete, when it
should appear as unchanged despite not being present.  Avoid this
problem by copying the skip_worktree flag in this case.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-27 11:15:20 -07:00
Derrick Stolee 6404355657 commit.h: remove method declarations
These methods are now declared in commit-reach.h. Remove them from
commit.h and add new include statements in all files that require these
declarations.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-20 15:38:54 -07:00
Junio C Hamano 00624d608c Merge branch 'sb/object-store-grafts'
The conversion to pass "the_repository" and then "a_repository"
throughout the object access API continues.

* sb/object-store-grafts:
  commit: allow lookup_commit_graft to handle arbitrary repositories
  commit: allow prepare_commit_graft to handle arbitrary repositories
  shallow: migrate shallow information into the object parser
  path.c: migrate global git_path_* to take a repository argument
  cache: convert get_graft_file to handle arbitrary repositories
  commit: convert read_graft_file to handle arbitrary repositories
  commit: convert register_commit_graft to handle arbitrary repositories
  commit: convert commit_graft_pos() to handle arbitrary repositories
  shallow: add repository argument to is_repository_shallow
  shallow: add repository argument to check_shallow_file_for_update
  shallow: add repository argument to register_shallow
  shallow: add repository argument to set_alternate_shallow_file
  commit: add repository argument to lookup_commit_graft
  commit: add repository argument to prepare_commit_graft
  commit: add repository argument to read_graft_file
  commit: add repository argument to register_commit_graft
  commit: add repository argument to commit_graft_pos
  object: move grafts to object parser
  object-store: move object access functions to object-store.h
2018-07-18 12:20:28 -07:00
Junio C Hamano 473b8bb3aa Merge branch 'en/merge-recursive-cleanup'
Code cleanup.

* en/merge-recursive-cleanup:
  merge-recursive: add pointer about unduly complex looking code
  merge-recursive: rename conflict_rename_*() family of functions
  merge-recursive: clarify the rename_dir/RENAME_DIR meaning
  merge-recursive: align labels with their respective code blocks
  merge-recursive: fix numerous argument alignment issues
  merge-recursive: fix miscellaneous grammar error in comment
2018-07-18 12:20:27 -07:00
Elijah Newren eddd1a411d merge-recursive: enforce rule that index matches head before merging
builtin/merge.c says that when we are about to perform a merge:

    ...the index must be in sync with the head commit.  The strategies are
    responsible to ensure this.

merge-recursive has always relied on unpack_trees() to enforce this
requirement, except in the case of an "Already up to date!" merge.
unpack-trees.c does not actually enforce this requirement, though.  It
allows for a pair of exceptions, in cases which it refers to as #14(ALT)
and #2ALT.  Documentation/technical/trivial-merge.txt can be consulted for
the precise meanings of the various case numbers and their meanings for
unpack-trees.c, but we have a high-level description of the intent behind
these two exceptions in a combined and summarized form in
Documentation/git-merge.txt:

    ...[merge will] abort if there are any changes registered in the index
    relative to the `HEAD` commit.  (One exception is when the changed index
    entries are in the state that would result from the merge already.)

While this high-level description does describe conditions under which it
would be safe to allow the index to diverge from HEAD, it does not match
what is actually implemented.  In particular, unpack-trees.c has no
knowledge of renames, and these two exceptions were written assuming that
no renames take place.  Once renames get into the mix, it is no longer
safe to allow the index to not match for #2ALT.  We could modify
unpack-trees to only allow #14(ALT) as an exception, but that would be
more strict than required for the resolve strategy (since the resolve
strategy doesn't handle renames at all).  Therefore, unpack_trees.c seems
like the wrong place to fix this.

Further, if someone fixes the combination of break and rename detection
and modifies merge-recursive to take advantage of the combination, then it
will also no longer be safe to allow the index to not match for #14(ALT)
when the recursive strategy is in use.  Therefore, leaving one of the
exceptions in place with the recursive merge strategy feels like we are
just leaving a latent bug in the code for folks in the future to stumble
across.

It may be possible to fix both unpack-trees and merge-recursive in a way
that implements the exception as stated in Documentation/git-merge.txt,
but it would be somewhat complex, possibly also buggy at first, and
ultimately, not all that valuable.  Instead, just enforce the requirement
stated in builtin/merge.c; error out if the index does not match the HEAD
commit, just like the 'ours' and 'octopus' strategies do.

Some testcase fixups were in order:
  t7611: had many tests designed to show that `git merge --abort` could
	 not always restore the index and working tree to the state they
	 were in before the merge started.  The tests that were associated
	 with having changes in the index before the merge started are no
         longer applicable, so they have been removed.
  t7504: had a few tests that had stray staged changes that were not
         actually part of the test under consideration
  t6044: We no longer expect stray staged changes to sometimes result
         in the merge continuing.  Also, fix a case where a merge
         didn't abort but should have.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-11 09:38:36 -07:00
Elijah Newren e1f8694f33 merge-recursive: fix assumption that head tree being merged is HEAD
`git merge-recursive` does a three-way merge between user-specified trees
base, head, and remote.  Since the user is allowed to specify head, we can
not necesarily assume that head == HEAD.

Modify index_has_changes() to take an extra argument specifying the tree
to compare against.  If NULL, it will compare to HEAD.  We then use this
from merge-recursive to make sure we compare to the user-specified head.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-11 09:38:36 -07:00
Elijah Newren 92702392ce merge-recursive: make sure when we say we abort that we actually abort
In commit 65170c07d4 ("merge-recursive: avoid incorporating uncommitted
changes in a merge", 2017-12-21), it was noted that there was a special
case when merge-recursive didn't rely on unpack_trees() to enforce the
index == HEAD requirement, and thus that it needed to do that enforcement
itself.  Unfortunately, it returned the wrong exit status, signalling that
the merge completed but had conflicts, rather than that it was aborted.
Fix the return code, and while we're at it, change the error message to
match what unpack_trees() would have printed.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-11 09:38:36 -07:00
Elijah Newren 1b9fbefbe0 index_has_changes(): avoid assuming operating on the_index
Modify index_has_changes() to take a struct istate* instead of just
operating on the_index.  This is only a partial conversion, though,
because we call do_diff_cache() which implicitly assumes work is to be
done on the_index.  Ongoing work is being done elsewhere to do the
remainder of the conversion, and thus is not duplicated here.  Instead,
a simple check is put in place until that work is complete.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-03 13:13:18 -07:00
Jameson Miller a849735bfb block alloc: add lifecycle APIs for cache_entry structs
It has been observed that the time spent loading an index with a large
number of entries is partly dominated by malloc() calls. This change
is in preparation for using memory pools to reduce the number of
malloc() calls made to allocate cahce entries when loading an index.

Add an API to allocate and discard cache entries, abstracting the
details of managing the memory backing the cache entries. This commit
does actually change how memory is managed - this will be done in a
later commit in the series.

This change makes the distinction between cache entries that are
associated with an index and cache entries that are not associated with
an index. A main use of cache entries is with an index, and we can
optimize the memory management around this. We still have other cases
where a cache entry is not persisted with an index, and so we need to
handle the "transient" use case as well.

To keep the congnitive overhead of managing the cache entries, there
will only be a single discard function. This means there must be enough
information kept with the cache entry so that we know how to discard
them.

A summary of the main functions in the API is:

make_cache_entry: create cache entry for use in an index. Uses specified
                  parameters to populate cache_entry fields.

make_empty_cache_entry: Create an empty cache entry for use in an index.
                        Returns cache entry with empty fields.

make_transient_cache_entry: create cache entry that is not used in an
                            index. Uses specified parameters to populate
                            cache_entry fields.

make_empty_transient_cache_entry: create cache entry that is not used in
                                  an index. Returns cache entry with
                                  empty fields.

discard_cache_entry: A single function that knows how to discard a cache
                     entry regardless of how it was allocated.

Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-03 10:58:27 -07:00
Jameson Miller 825ed4d9a0 read-cache: teach make_cache_entry to take object_id
Teach make_cache_entry function to take object_id instead of a SHA-1.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-03 10:58:15 -07:00
Jameson Miller 768d796506 read-cache: teach refresh_cache_entry to take istate
Refactor refresh_cache_entry() to work on a specific index, instead of
implicitly using the_index. This is in preparation for making the
make_cache_entry function apply to a specific index.

Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-03 10:58:15 -07:00
Stefan Beller a74093da5e tag: add repository argument to deref_tag
Add a repository argument to allow the callers of deref_tag
to be more specific about which repository to act on. This is a small
mechanical change; it doesn't change the implementation to handle
repositories other than the_repository yet.

As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-29 10:43:39 -07:00
Stefan Beller 2122f6754c commit: add repository argument to lookup_commit_reference
Add a repository argument to allow callers of lookup_commit_reference
to be more specific about which repository to handle. This is a small
mechanical change; it doesn't change the implementation to handle
repositories other than the_repository yet.

As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-29 10:43:39 -07:00
Stefan Beller f86bcc7b2c tree: add repository argument to lookup_tree
Add a repository argument to allow the callers of lookup_tree
to be more specific about which repository to act on. This is a small
mechanical change; it doesn't change the implementation to handle
repositories other than the_repository yet.

As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-29 10:43:38 -07:00
Stefan Beller 109cd76dd3 object: add repository argument to parse_object
Add a repository argument to allow the callers of parse_object
to be more specific about which repository to act on. This is a small
mechanical change; it doesn't change the implementation to handle
repositories other than the_repository yet.

As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-29 10:43:38 -07:00
Junio C Hamano b16b60f71b Merge branch 'sb/object-store-grafts' into sb/object-store-lookup
* sb/object-store-grafts:
  commit: allow lookup_commit_graft to handle arbitrary repositories
  commit: allow prepare_commit_graft to handle arbitrary repositories
  shallow: migrate shallow information into the object parser
  path.c: migrate global git_path_* to take a repository argument
  cache: convert get_graft_file to handle arbitrary repositories
  commit: convert read_graft_file to handle arbitrary repositories
  commit: convert register_commit_graft to handle arbitrary repositories
  commit: convert commit_graft_pos() to handle arbitrary repositories
  shallow: add repository argument to is_repository_shallow
  shallow: add repository argument to check_shallow_file_for_update
  shallow: add repository argument to register_shallow
  shallow: add repository argument to set_alternate_shallow_file
  commit: add repository argument to lookup_commit_graft
  commit: add repository argument to prepare_commit_graft
  commit: add repository argument to read_graft_file
  commit: add repository argument to register_commit_graft
  commit: add repository argument to commit_graft_pos
  object: move grafts to object parser
  object-store: move object access functions to object-store.h
2018-06-29 10:43:28 -07:00
Junio C Hamano 110240588d Merge branch 'sb/object-store-alloc'
The conversion to pass "the_repository" and then "a_repository"
throughout the object access API continues.

* sb/object-store-alloc:
  alloc: allow arbitrary repositories for alloc functions
  object: allow create_object to handle arbitrary repositories
  object: allow grow_object_hash to handle arbitrary repositories
  alloc: add repository argument to alloc_commit_index
  alloc: add repository argument to alloc_report
  alloc: add repository argument to alloc_object_node
  alloc: add repository argument to alloc_tag_node
  alloc: add repository argument to alloc_commit_node
  alloc: add repository argument to alloc_tree_node
  alloc: add repository argument to alloc_blob_node
  object: add repository argument to grow_object_hash
  object: add repository argument to create_object
  repository: introduce parsed objects field
2018-06-25 13:22:38 -07:00
Junio C Hamano b3b2aaf0fd Merge branch 'nd/commit-util-to-slab'
The in-core "commit" object had an all-purpose "void *util" field,
which was tricky to use especially in library-ish part of the
code.  All of the existing uses of the field has been migrated to a
more dedicated "commit-slab" mechanism and the field is eliminated.

* nd/commit-util-to-slab:
  commit.h: delete 'util' field in struct commit
  merge: use commit-slab in merge remote desc instead of commit->util
  log: use commit-slab in prepare_bases() instead of commit->util
  show-branch: note about its object flags usage
  show-branch: use commit-slab for commit-name instead of commit->util
  name-rev: use commit-slab for rev-name instead of commit->util
  bisect.c: use commit-slab for commit weight instead of commit->util
  revision.c: use commit-slab for show_source
  sequencer.c: use commit-slab to associate todo items to commits
  sequencer.c: use commit-slab to mark seen commits
  shallow.c: use commit-slab for commit depth instead of commit->util
  describe: use commit-slab for commit names instead of commit->util
  blame: use commit-slab for blame suspects instead of commit->util
  commit-slab: support shared commit-slab
  commit-slab.h: code split
2018-06-25 13:22:35 -07:00
Junio C Hamano f72432d64e Merge branch 'en/rename-directory-detection'
Newly added codepath in merge-recursive had potential buffer
overrun, which has been fixed.

* en/rename-directory-detection:
  merge-recursive: use xstrdup() instead of fixed buffer
2018-06-18 10:18:44 -07:00
René Scharfe 9da2d0379e merge-recursive: use xstrdup() instead of fixed buffer
Paths can be longer than PATH_MAX.  Avoid a buffer overrun in
check_dir_renamed() by using xstrdup() to make a private copy safely.

Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-14 08:56:35 -07:00
Junio C Hamano 8d0d53a8cf Merge branch 'sb/submodule-merge-in-merge-recursive'
Finishing touches to a topic that already is in 'master'.

* sb/submodule-merge-in-merge-recursive:
  merge-submodule: reduce output verbosity
2018-06-13 12:50:45 -07:00
Elijah Newren 5b26c3c941 merge-recursive: add pointer about unduly complex looking code
handle_change_delete() has a block of code displaying one of four nearly
identical messages.  Each contains about half a dozen variable
interpolations, which use nearly identical variables as well.  Someone
trying to parse this may be slowed down trying to parse the differences
and why they are here; help them out by adding a comment explaining the
differences.

Further, point out that this code structure isn't collapsed into something
more concise and readable for the programmer, because we want to keep full
messages intact in order to make translators' jobs much easier.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:57 -07:00
Elijah Newren 8ebe7b057a merge-recursive: rename conflict_rename_*() family of functions
These functions were added because processing of these conflicts needed
to be deferred until process_entry() in order to get D/F conflicts and
such right.  The number of these has grown over time, and now include
some whose name is misleading:
  * conflict_rename_normal() is for handling normal file renames; a
    typical rename may need content merging, but we expect conflicts
    from that to be more the exception than the rule.
  * conflict_rename_via_dir() will not be a conflict; it was just an
    add that turned into a move due to directory rename detection.
    (If there was a file in the way of the move, that would have been
    detected and reported earlier.)
  * conflict_rename_rename_2to1 and conflict_rename_add (the latter
    of which doesn't exist yet but has been submitted before and I
    intend to resend) technically might not be conflicts if the
    colliding paths happen to match exactly.
Rename this family of functions to handle_rename_*().

Also rename handle_renames() to detect_and_process_renames() both to make
it clearer what it does, and to differentiate it as a pre-processing step
from all the handle_rename_*() functions which are called from
process_entry().

Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:57 -07:00
Elijah Newren 5455c33839 merge-recursive: clarify the rename_dir/RENAME_DIR meaning
We had an enum of rename types which included RENAME_DIR; this name felt
misleading since it was not about an entire directory but was a status for
each individual file add that occurred within a renamed directory.

Since this type is for signifying that the files in question were being
renamed due to directory rename detection, rename this enum value to
RENAME_VIA_DIR.

Make a similar change to the conflict_rename_dir() function, and add a
comment to the top of that function explaining its purpose (it may not be
quite as obvious as for the other conflict_rename_*() functions).

Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:57 -07:00
Elijah Newren 9366536583 merge-recursive: align labels with their respective code blocks
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:57 -07:00
Elijah Newren d90e759fd5 merge-recursive: fix numerous argument alignment issues
Various refactorings throughout the code have left lots of alignment
issues that were driving me crazy; fix them.

Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:57 -07:00
Elijah Newren 2d6bad918d merge-recursive: fix miscellaneous grammar error in comment
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-12 10:22:50 -07:00
Leif Middelschulte 40aac22b43 merge-submodule: reduce output verbosity
The output shall behave more similar to ordinary file merges' output to provide
a more consistent user experience.

Signed-off-by: Leif Middelschulte <Leif.Middelschulte@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-06-11 11:03:55 -07:00
Junio C Hamano e47dbece39 Merge branch 'ma/unpack-trees-free-msgs'
Leak plugging.

* ma/unpack-trees-free-msgs:
  unpack_trees_options: free messages when done
  argv-array: return the pushed string from argv_push*()
  merge-recursive: provide pair of `unpack_trees_{start,finish}()`
  merge: setup `opts` later in `checkout_fast_forward()`
2018-05-30 21:51:29 +09:00
Junio C Hamano 0821b73063 Merge branch 'sb/submodule-merge-in-merge-recursive'
By code restructuring of submodule merge in merge-recursive,
informational messages from the codepath are now given using the
same mechanism as other output, and honor the merge.verbosity
configuration.  The code also learned to give a few new messages
when a submodule three-way merge resolves cleanly when one side
records a descendant of the commit chosen by the other side.

* sb/submodule-merge-in-merge-recursive:
  merge-recursive: give notice when submodule commit gets fast-forwarded
  merge-recursive: i18n submodule merge output and respect verbosity
  submodule.c: move submodule merging to merge-recursive.c
2018-05-30 21:51:27 +09:00
Junio C Hamano 50f08db594 Merge branch 'js/use-bug-macro'
Developer support update, by using BUG() macro instead of die() to
mark codepaths that should not happen more clearly.

* js/use-bug-macro:
  BUG_exit_code: fix sparse "symbol not declared" warning
  Convert remaining die*(BUG) messages
  Replace all die("BUG: ...") calls by BUG() ones
  run-command: use BUG() to report bugs, not die()
  test-tool: help verifying BUG() code paths
2018-05-30 14:04:07 +09:00
Junio C Hamano 6e2ba77bda Merge branch 'bp/merge-rename-config'
With merge.renames configuration set to false, the recursive merge
strategy can be told not to spend cycles trying to find renamed
paths and merge them accordingly.

* bp/merge-rename-config:
  merge: pass aggressive when rename detection is turned off
  merge: add merge.renames config setting
  merge: update documentation for {merge,diff}.renameLimit
2018-05-30 14:04:04 +09:00
Junio C Hamano c67de747f4 Merge branch 'en/rename-directory-detection-reboot'
Rename detection logic in "diff" family that is used in "merge" has
learned to guess when all of x/a, x/b and x/c have moved to z/a,
z/b and z/c, it is likely that x/d added in the meantime would also
want to move to z/d by taking the hint that the entire directory
'x' moved to 'z'.  A bug causing dirty files involved in a rename
to be overwritten during merge has also been fixed as part of this
work.  Incidentally, this also avoids updating a file in the
working tree after a (non-trivial) merge whose result matches what
our side originally had.

* en/rename-directory-detection-reboot: (36 commits)
  merge-recursive: fix check for skipability of working tree updates
  merge-recursive: make "Auto-merging" comment show for other merges
  merge-recursive: fix remainder of was_dirty() to use original index
  merge-recursive: fix was_tracked() to quit lying with some renamed paths
  t6046: testcases checking whether updates can be skipped in a merge
  merge-recursive: avoid triggering add_cacheinfo error with dirty mod
  merge-recursive: move more is_dirty handling to merge_content
  merge-recursive: improve add_cacheinfo error handling
  merge-recursive: avoid spurious rename/rename conflict from dir renames
  directory rename detection: new testcases showcasing a pair of bugs
  merge-recursive: fix remaining directory rename + dirty overwrite cases
  merge-recursive: fix overwriting dirty files involved in renames
  merge-recursive: avoid clobbering untracked files with directory renames
  merge-recursive: apply necessary modifications for directory renames
  merge-recursive: when comparing files, don't include trees
  merge-recursive: check for file level conflicts then get new name
  merge-recursive: add computation of collisions due to dir rename & merging
  merge-recursive: check for directory level conflicts
  merge-recursive: add get_directory_renames()
  merge-recursive: make a helper function for cleanup for handle_renames
  ...
2018-05-23 14:38:19 +09:00
Junio C Hamano c89b6e136e Merge branch 'ds/lazy-load-trees'
The code has been taught to use the duplicated information stored
in the commit-graph file to learn the tree object name for a commit
to avoid opening and parsing the commit object when it makes sense
to do so.

* ds/lazy-load-trees:
  coccinelle: avoid wrong transformation suggestions from commit.cocci
  commit-graph: lazy-load trees for commits
  treewide: replace maybe_tree with accessor methods
  commit: create get_commit_tree() method
  treewide: rename tree to maybe_tree
2018-05-23 14:38:13 +09:00
Martin Ågren 1c41d2805e unpack_trees_options: free messages when done
The strings allocated in `setup_unpack_trees_porcelain()` are never
freed. Provide a function `clear_unpack_trees_porcelain()` to do so and
call it where we use `setup_unpack_trees_porcelain()`. The only
non-trivial user is `unpack_trees_start()`, where we should place the
new call in `unpack_trees_finish()`.

We keep the string pointers in an array, mixing pointers to static
memory and memory that we allocate on the heap. We also keep several
copies of the individual pointers. So we need to make sure that we do
not free what we must not free and that we do not double-free. Let a
separate argv_array take ownership of all the strings we create so that
we can easily free them.

Zero the whole array of string pointers to make sure that we do not
leave any dangling pointers.

Note that we only take responsibility for the memory allocated in
`setup_unpack_trees_porcelain()` and not any other members of the
`struct unpack_trees_options`.

Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-22 11:59:31 +09:00
Nguyễn Thái Ngọc Duy e2e5ac2303 merge: use commit-slab in merge remote desc instead of commit->util
It's done so that commit->util can be removed. See more explanation in
the commit that removes commit->util.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-21 14:07:20 +09:00
Elijah Newren 3f1c1c3600 merge-recursive: provide pair of unpack_trees_{start,finish}()
Rename `git_merge_trees()` to `unpack_trees_start()` and extract the
call to `discard_index()` into a new function `unpack_trees_finish()`.
As a result, these are called early resp. late in `merge_trees()`,
making the resource handling clearer. A later commit will expand on
that, teaching `..._finish()` to free more memory. (So rather than
moving the FIXME-comment, just drop it, since it will be addressed soon
enough.)

Also call `..._finish()` when `merge_trees()` returns early.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-21 06:56:15 +09:00
Leif Middelschulte 76f4212597 merge-recursive: give notice when submodule commit gets fast-forwarded
Inform the user about an automatically fast-forwarded submodule. The
silent merge behavior was introduced by commit 68d03e4a6e ("Implement
automatic fast-forward merge for submodules", 2010-07-07)).

Signed-off-by: Leif Middelschulte <Leif.Middelschulte@gmail.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-18 07:59:13 +09:00
Stefan Beller cbd53a2193 object-store: move object access functions to object-store.h
This should make these functions easier to find and cache.h less
overwhelming to read.

In particular, this moves:
- read_object_file
- oid_object_info
- write_object_file

As a result, most of the codebase needs to #include object-store.h.
In this patch the #include is only added to files that would fail to
compile otherwise.  It would be better to #include wherever
identifiers from the header are used.  That can happen later
when we have better tooling for it.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-16 11:42:03 +09:00
Stefan Beller 14ba97f81c alloc: allow arbitrary repositories for alloc functions
We have to convert all of the alloc functions at once, because alloc_report
uses a funky macro for reporting. It is better for the sake of mechanical
conversion to convert multiple functions at once rather than changing the
structure of the reporting function.

We record all memory allocation in alloc.c, and free them in
clear_alloc_state, which is called for all repositories except
the_repository.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-16 11:16:50 +09:00
Stefan Beller 325f3a8e07 merge-recursive: i18n submodule merge output and respect verbosity
The submodule merge code now uses the output() function that is used by
all the rest of the merge-recursive-code. This allows for respecting
internationalisation as well as the verbosity setting.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-16 10:08:43 +09:00
Stefan Beller 18cfc08866 submodule.c: move submodule merging to merge-recursive.c
In a later patch we want to improve submodule merging by using the output()
function in merge-recursive.c for submodule merges to deliver a consistent
UI to users.

To do so we could either make the output() function globally available
so we can use it in submodule.c#merge_submodule(), or we could integrate
the submodule merging into the merging code. Choose the later as we
generally want to move submodules closer into the core.

Therefore we move any function related to merging submodules
(merge_submodule(), find_first_merges() and print_commit) to
merge-recursive.c.  We'll keep add_submodule_odb() in submodule.c as it
is used by other submodule functions. While at it, add a TODO note that
we do not really like the function add_submodule_odb().

This commit is best viewed with --color-moved.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-16 10:08:43 +09:00
Stefan Beller 8ba0e5ec57 alloc: add repository argument to alloc_commit_node
This is a small mechanical change; it doesn't change the
implementation to handle repositories other than the_repository yet.
Use a macro to catch callers passing a repository other than
the_repository at compile time.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-09 12:12:36 +09:00
Ben Peart 6f10a09e0a merge: pass aggressive when rename detection is turned off
Set aggressive flag in git_merge_trees() when rename detection is turned off.
This allows read_tree() to auto resolve more cases that would have otherwise
been handled by the rename detection.

Reviewed-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:19:41 +09:00
Ben Peart 85b460305c merge: add merge.renames config setting
Add the ability to control rename detection for merge via a config setting.
This setting behaves the same and defaults to the value of diff.renames but only
applies to merge.

Reviewed-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Helped-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:19:41 +09:00
Elijah Newren 1de70dbd1a merge-recursive: fix check for skipability of working tree updates
The can-working-tree-updates-be-skipped check has had a long and blemished
history.  The update can be skipped iff:
  a) The merge is clean
  b) The merge matches what was in HEAD (content, mode, pathname)
  c) The target path is usable (i.e. not involved in D/F conflict)

Traditionally, we split b into parts:
  b1) The merged result matches the content and mode found in HEAD
  b2) The merged target path existed in HEAD

Steps a & b1 are easy to check; we have always gotten those right.  While
it is easy to overlook step c, this was fixed seven years ago with commit
4ab9a157d0 ("merge_content(): Check whether D/F conflicts are still
present", 2010-09-20).  merge-recursive didn't have a readily available
way to directly check step b2, so various approximations were used:

  * In commit b2c8c0a762 ("merge-recursive: When we detect we can skip
    an update, actually skip it", 2011-02-28), it was noted that although
    the code claimed it was skipping the update, it did not actually skip
    the update.  The code was made to skip it, but used lstat(path, ...)
    as an approximation to path-was-tracked-in-index-before-merge.

  * In commit 5b448b8530 ("merge-recursive: When we detect we can skip
    an update, actually skip it", 2011-08-11), the problem with using
    lstat was noted.  It was changed to the approximation
       path2 && strcmp(path, path2)
    which is also wrong.  !path2 || strcmp(path, path2) would have been
    better, but would have fallen short with directory renames.

  * In c5b761fb27 ("merge-recursive: ensure we write updates for
    directory-renamed file", 2018-02-14), the problem with the previous
    approximation was noted and changed to
       was_tracked(path)
    That looks close to what we were trying to answer, but was_tracked()
    as implemented at the time should have been named is_tracked(); it
    returned something different than what we were looking for.

  * To make matters more complex, fixing was_tracked() isn't sufficient
    because the splitting of b into b1 and b2 is wrong.  Consider the
    following merge with a rename/add conflict:
       side A: modify foo, add unrelated bar
       side B: rename foo->bar (but don't modify the mode or contents)
    In this case, the three-way merge of original foo, A's foo, and B's
    bar will result in a desired pathname of bar with the same
    mode/contents that A had for foo.  Thus, A had the right mode and
    contents for the file, and it had the right pathname present (namely,
    bar), but the bar that was present was unrelated to the contents, so
    the working tree update was not skippable.

Fix this by introducing a new function:
   was_tracked_and_matches(o, path, &mfi.oid, mfi.mode)
and use it to directly check for condition b.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 05cf21eba2 merge-recursive: make "Auto-merging" comment show for other merges
Previously, merge_content() would print "Auto-merging" whenever the final
content and mode aren't already available from HEAD.  There are a few
problems with this:

  1) There are other code paths doing merges that should probably have the
     same message printed, in particular rename/rename(2to1) which cannot
     call into the normal rename logic.

  2) If both sides of the merge have modifications, then a content merge
     is needed.  It may turn out that the end result matches one of the
     sides (because the other only had a subset of the same changes), but
     the merge was still needed.  Currently, the message will not print in
     that case, though it seems like it should.

Move the printing of this message to merge_file_1() in order to address
both issues.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 277292d5ae merge-recursive: fix remainder of was_dirty() to use original index
was_dirty() uses was_tracked(), which has been updated to use the original
index rather than the current one.  However, was_dirty() also had a
separate call to cache_file_exists(), causing it to still implicitly use
the current index.  Update that to instead use index_file_exists().

Also, was_dirty() had a hack where it would mark any file as non-dirty if
we simply didn't know its modification time.  This was due to using the
current index rather than the original index, because D/F conflicts and
such would cause unpack_trees() to not copy the modification times from
the original index to the current one.  Now that we are using the original
index, we can dispense with this hack.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren a35edc84bd merge-recursive: fix was_tracked() to quit lying with some renamed paths
In commit aacb82de3f ("merge-recursive: Split was_tracked() out of
would_lose_untracked()", 2011-08-11), was_tracked() was split out of
would_lose_untracked() with the intent to provide a function that could
answer whether a path was tracked in the index before the merge.  Sadly,
it instead returned whether the path was in the working tree due to having
been tracked in the index before the merge OR having been written there by
unpack_trees().  The distinction is important when renames are involved,
e.g. for a merge where:

   HEAD:  modifies path b
   other: renames b->c

In this case, c was not tracked in the index before the merge, but would
have been added to the index at stage 0 and written to the working tree by
unpack_trees().  would_lose_untracked() is more interested in the
in-working-copy-for-either-reason behavior, while all other uses of
was_tracked() want just was-it-tracked-in-index-before-merge behavior.

Unsplit would_lose_untracked() and write a new was_tracked() function
which answers whether a path was tracked in the index before the merge
started.

This will also affect was_dirty(), helping it to return better results
since it can base answers off the original index rather than an index that
possibly only copied over some of the stat information.  However,
was_dirty() will need an additional change that will be made in a
subsequent patch.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 2f682e21a6 merge-recursive: avoid triggering add_cacheinfo error with dirty mod
If a cherry-pick or merge with a rename results in a skippable update
(due to the merged content matching what HEAD already had), but the
working directory is dirty, avoid trying to refresh the index as that
will fail.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren bd42380ef1 merge-recursive: move more is_dirty handling to merge_content
conflict_rename_normal() was doing some handling for dirty files that
more naturally belonged in merge_content.  Move it, and rename a
parameter for clarity while at it.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren fd53b7ffd1 merge-recursive: improve add_cacheinfo error handling
Four closely related changes all with the purpose of fixing error handling
in this function:
  - fix reported function name in add_cacheinfo error messages
  - differentiate between the two error messages
  - abort early when we hit the error (stop ignoring return code)
  - mark a test which was hitting this error as failing until we get the
    right fix

In more detail...

In commit 0424138d57 ("Fix bogus error message from merge-recursive
error path", 2007-04-01), it was noted that the name of the function which
the error message claimed it was reported from did not match the actual
function name.  This was changed to something closer to the real function
name, but it still didn't match the actual function name.  Fix the
reported name to match.

Second, the two errors in this function had identical messages, preventing
us from knowing which error had been triggered.  Add a couple words to the
second error message to differentiate the two.

Next, make sure callers do not ignore the return code so that it will stop
processing further entries (processing further entries could result in
more output which could cause the error to scroll off the screen, or at
least be missed by the user) and make it clear the error is the cause of
the early abort.  These errors should never be triggered in production; if
either one is, it represents a bug in the calling path somewhere and is
likely to have resulted in mis-merged content.  The combination of
ignoring of the return code and continuing to print other standard
messages after hitting the error resulted in the following bug report from
Junio: "...the command pretends that everything went well and merged
cleanly in that path...[Behaving] in a buggy and unexplainable way is bad
enough, doing so silently is unexcusable."  Fix this.

Finally, there was one test in the testsuite that did hit this error path,
but was passing anyway.  This would have been easy to miss since it had a
test_must_fail and thus could have failed for the wrong reason, but in a
separate testing step I added an intentional NULL-dereference to the
codepath where these error messages are printed in order to flush out such
cases.  I could modify that test to explicitly check for this error and
fail the test if it is hit, but since this test operates in a bit of a
gray area and needed other changes, I went for a different fix.  The gray
area this test operates in is the following: If the merge of a certain
file results in the same version of the file that existed in HEAD, but
there are dirty modifications to the file, is that an error with a
"Refusing to overwrite existing file" expected, or a case where the merge
should succeed since we shouldn't have to touch the dirty file anyway?
Recent discussion on the list leaned towards saying it should be a
success.  Therefore, change the expected behavior of this test to match.
As a side effect, this makes the failed-due-to-hitting-add_cacheinfo-error
very clear, and we can mark the test as test_expect_failure.  A subsequent
commit will implement the necessary changes to get this test to pass
again.

Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 6e7e027fe5 merge-recursive: avoid spurious rename/rename conflict from dir renames
If a file on one side of history was renamed, and merely modified on the
other side, then applying a directory rename to the modified side gives us
a rename/rename(1to2) conflict.  We should only apply directory renames to
pairs representing either adds or renames.

Making this change means that a directory rename testcase that was
previously reported as a rename/delete conflict will now be reported as a
modify/delete conflict.

Reviewed-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 18797a3b10 merge-recursive: fix remaining directory rename + dirty overwrite cases
Reviewed-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00
Elijah Newren 64b1abe962 merge-recursive: fix overwriting dirty files involved in renames
This fixes an issue that existed before my directory rename detection
patches that affects both normal renames and renames implied by
directory rename detection.  Additional codepaths that only affect
overwriting of dirty files that are involved in directory rename
detection will be added in a subsequent commit.

Reviewed-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-05-08 16:11:00 +09:00