git/t/t6426-merge-skip-unneeded-updates.sh

780 lines
18 KiB
Bash
Raw Normal View History

#!/bin/sh
test_description="merge cases"
# The setup for all of them, pictorially, is:
#
# A
# o
# / \
# O o ?
# \ /
# o
# B
#
# To help make it easier to follow the flow of tests, they have been
# divided into sections and each test will start with a quick explanation
# of what commits O, A, and B contain.
#
# Notation:
# z/{b,c} means files z/b and z/c both exist
# x/d_1 means file x/d exists with content d1. (Purpose of the
# underscore notation is to differentiate different
# files that might be renamed into each other's paths.)
. ./test-lib.sh
. "$TEST_DIRECTORY"/lib-merge.sh
###########################################################################
# SECTION 1: Cases involving no renames (one side has subset of changes of
# the other side)
###########################################################################
# Testcase 1a, Changes on A, subset of changes on B
# Commit O: b_1
# Commit A: b_2
# Commit B: b_3
# Expected: b_2
test_setup_1a () {
test_create_repo 1a_$1 &&
(
cd 1a_$1 &&
test_write_lines 1 2 3 4 5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 10.5 >b &&
git add b &&
test_tick &&
git commit -m "A" &&
git checkout B &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '1a-L: Modify(A)/Modify(B), change on B subset of A' '
test_setup_1a L &&
(
cd 1a_L &&
git checkout A^0 &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 b >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
# Make sure b was NOT updated
test-tool chmtime --get b >new-mtime &&
test_cmp old-mtime new-mtime &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:b &&
git rev-parse >expect A:b &&
test_cmp expect actual &&
git hash-object b >actual &&
git rev-parse A:b >expect &&
test_cmp expect actual
)
'
test_expect_success '1a-R: Modify(A)/Modify(B), change on B subset of A' '
test_setup_1a R &&
(
cd 1a_R &&
git checkout B^0 &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 b >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive A^0 >out 2>err &&
# Make sure b WAS updated
test-tool chmtime --get b >new-mtime &&
test $(cat old-mtime) -lt $(cat new-mtime) &&
test_must_be_empty err &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:b &&
git rev-parse >expect A:b &&
test_cmp expect actual &&
git hash-object b >actual &&
git rev-parse A:b >expect &&
test_cmp expect actual
)
'
###########################################################################
# SECTION 2: Cases involving basic renames
###########################################################################
# Testcase 2a, Changes on A, rename on B
# Commit O: b_1
# Commit A: b_2
# Commit B: c_1
# Expected: c_2
test_setup_2a () {
test_create_repo 2a_$1 &&
(
cd 2a_$1 &&
test_seq 1 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_seq 1 11 >b &&
git add b &&
test_tick &&
git commit -m "A" &&
git checkout B &&
git mv b c &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '2a-L: Modify/rename, merge into modify side' '
test_setup_2a L &&
(
cd 2a_L &&
git checkout A^0 &&
test_path_is_missing c &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive B^0 >out 2>err &&
test_path_is_file c &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:c &&
git rev-parse >expect A:b &&
test_cmp expect actual &&
git hash-object c >actual &&
git rev-parse A:b >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:b &&
test_path_is_missing b
)
'
test_expect_success '2a-R: Modify/rename, merge into rename side' '
test_setup_2a R &&
(
cd 2a_R &&
git checkout B^0 &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 c >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive A^0 >out 2>err &&
# Make sure c WAS updated
test-tool chmtime --get c >new-mtime &&
test $(cat old-mtime) -lt $(cat new-mtime) &&
test_must_be_empty err &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:c &&
git rev-parse >expect A:b &&
test_cmp expect actual &&
git hash-object c >actual &&
git rev-parse A:b >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:b &&
test_path_is_missing b
)
'
# Testcase 2b, Changed and renamed on A, subset of changes on B
# Commit O: b_1
# Commit A: c_2
# Commit B: b_3
# Expected: c_2
test_setup_2b () {
test_create_repo 2b_$1 &&
(
cd 2b_$1 &&
test_write_lines 1 2 3 4 5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 10.5 >b &&
git add b &&
git mv b c &&
test_tick &&
git commit -m "A" &&
git checkout B &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '2b-L: Rename+Mod(A)/Mod(B), B mods subset of A' '
test_setup_2b L &&
(
cd 2b_L &&
git checkout A^0 &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 c >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
# Make sure c WAS updated
test-tool chmtime --get c >new-mtime &&
test_cmp old-mtime new-mtime &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:c &&
git rev-parse >expect A:c &&
test_cmp expect actual &&
git hash-object c >actual &&
git rev-parse A:c >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:b &&
test_path_is_missing b
)
'
test_expect_success '2b-R: Rename+Mod(A)/Mod(B), B mods subset of A' '
test_setup_2b R &&
(
cd 2b_R &&
git checkout B^0 &&
test_path_is_missing c &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive A^0 >out 2>err &&
# Make sure c now present (and thus was updated)
test_path_is_file c &&
test_must_be_empty err &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual HEAD:c &&
git rev-parse >expect A:c &&
test_cmp expect actual &&
git hash-object c >actual &&
git rev-parse A:c >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:b &&
test_path_is_missing b
)
'
# Testcase 2c, Changes on A, rename on B
# Commit O: b_1
# Commit A: b_2, c_3
# Commit B: c_1
# Expected: rename/add conflict c_2 vs c_3
#
# NOTE: Since A modified b_1->b_2, and B renamed b_1->c_1, the threeway
# merge of those files should result in c_2. We then should have a
# rename/add conflict between c_2 and c_3. However, if we note in
# merge_content() that A had the right contents (b_2 has same
# contents as c_2, just at a different name), and that A had the
# right path present (c_3 existed) and thus decides that it can
# skip the update, then we're in trouble. This test verifies we do
# not make that particular mistake.
test_setup_2c () {
test_create_repo 2c &&
(
cd 2c &&
test_seq 1 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_seq 1 11 >b &&
echo whatever >c &&
git add b c &&
test_tick &&
git commit -m "A" &&
git checkout B &&
git mv b c &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '2c: Modify b & add c VS rename b->c' '
test_setup_2c &&
(
cd 2c &&
git checkout A^0 &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 c >old-mtime &&
GIT_MERGE_VERBOSITY=3 &&
export GIT_MERGE_VERBOSITY &&
test_must_fail git merge -s recursive B^0 >out 2>err &&
test_i18ngrep "CONFLICT (.*/add):" out &&
test_must_be_empty err &&
# Make sure c WAS updated
test-tool chmtime --get c >new-mtime &&
test $(cat old-mtime) -lt $(cat new-mtime)
# FIXME: rename/add conflicts are horribly broken right now;
# when I get back to my patch series fixing it and
# rename/rename(2to1) conflicts to bring them in line with
# how add/add conflicts behave, then checks like the below
# could be added. But that patch series is waiting until
# the rename-directory-detection series lands, which this
# is part of. And in the mean time, I do not want to further
# enforce broken behavior. So for now, the main test is the
# one above that err is an empty file.
#git ls-files -s >index_files &&
#test_line_count = 2 index_files &&
#git rev-parse >actual :2:c :3:c &&
#git rev-parse >expect A:b A:c &&
#test_cmp expect actual &&
#git cat-file -p A:b >>merged &&
#git cat-file -p A:c >>merge-me &&
#>empty &&
#test_must_fail git merge-file \
# -L "Temporary merge branch 1" \
# -L "" \
# -L "Temporary merge branch 2" \
# merged empty merge-me &&
#sed -e "s/^\([<=>]\)/\1\1\1/" merged >merged-internal &&
#git hash-object c >actual &&
#git hash-object merged-internal >expect &&
#test_cmp expect actual &&
#test_path_is_missing b
)
'
###########################################################################
# SECTION 3: Cases involving directory renames
#
# NOTE:
# Directory renames only apply when one side renames a directory, and the
# other side adds or renames a path into that directory. Applying the
# directory rename to that new path creates a new pathname that didn't
# exist on either side of history. Thus, it is impossible for the
# merge contents to already be at the right path, so all of these checks
# exist just to make sure that updates are not skipped.
###########################################################################
# Testcase 3a, Change + rename into dir foo on A, dir rename foo->bar on B
# Commit O: bq_1, foo/whatever
# Commit A: foo/{bq_2, whatever}
# Commit B: bq_1, bar/whatever
# Expected: bar/{bq_2, whatever}
test_setup_3a () {
test_create_repo 3a_$1 &&
(
cd 3a_$1 &&
mkdir foo &&
test_seq 1 10 >bq &&
test_write_lines a b c d e f g h i j k >foo/whatever &&
git add bq foo/whatever &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_seq 1 11 >bq &&
git add bq &&
git mv bq foo/ &&
test_tick &&
git commit -m "A" &&
git checkout B &&
git mv foo/ bar/ &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '3a-L: bq_1->foo/bq_2 on A, foo/->bar/ on B' '
test_setup_3a L &&
(
cd 3a_L &&
git checkout A^0 &&
test_path_is_missing bar/bq &&
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>
2019-04-05 15:00:26 +00:00
GIT_MERGE_VERBOSITY=3 git -c merge.directoryRenames=true merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
test_path_is_file bar/bq &&
git ls-files -s >index_files &&
test_line_count = 2 index_files &&
git rev-parse >actual HEAD:bar/bq HEAD:bar/whatever &&
git rev-parse >expect A:foo/bq A:foo/whatever &&
test_cmp expect actual &&
git hash-object bar/bq bar/whatever >actual &&
git rev-parse A:foo/bq A:foo/whatever >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:bq HEAD:foo/bq &&
test_path_is_missing bq &&
test_path_is_missing foo/bq &&
test_path_is_missing foo/whatever
)
'
test_expect_success '3a-R: bq_1->foo/bq_2 on A, foo/->bar/ on B' '
test_setup_3a R &&
(
cd 3a_R &&
git checkout B^0 &&
test_path_is_missing bar/bq &&
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>
2019-04-05 15:00:26 +00:00
GIT_MERGE_VERBOSITY=3 git -c merge.directoryRenames=true merge -s recursive A^0 >out 2>err &&
test_must_be_empty err &&
test_path_is_file bar/bq &&
git ls-files -s >index_files &&
test_line_count = 2 index_files &&
git rev-parse >actual HEAD:bar/bq HEAD:bar/whatever &&
git rev-parse >expect A:foo/bq A:foo/whatever &&
test_cmp expect actual &&
git hash-object bar/bq bar/whatever >actual &&
git rev-parse A:foo/bq A:foo/whatever >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:bq HEAD:foo/bq &&
test_path_is_missing bq &&
test_path_is_missing foo/bq &&
test_path_is_missing foo/whatever
)
'
# Testcase 3b, rename into dir foo on A, dir rename foo->bar + change on B
# Commit O: bq_1, foo/whatever
# Commit A: foo/{bq_1, whatever}
# Commit B: bq_2, bar/whatever
# Expected: bar/{bq_2, whatever}
test_setup_3b () {
test_create_repo 3b_$1 &&
(
cd 3b_$1 &&
mkdir foo &&
test_seq 1 10 >bq &&
test_write_lines a b c d e f g h i j k >foo/whatever &&
git add bq foo/whatever &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
git mv bq foo/ &&
test_tick &&
git commit -m "A" &&
git checkout B &&
test_seq 1 11 >bq &&
git add bq &&
git mv foo/ bar/ &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '3b-L: bq_1->foo/bq_2 on A, foo/->bar/ on B' '
test_setup_3b L &&
(
cd 3b_L &&
git checkout A^0 &&
test_path_is_missing bar/bq &&
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>
2019-04-05 15:00:26 +00:00
GIT_MERGE_VERBOSITY=3 git -c merge.directoryRenames=true merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
test_path_is_file bar/bq &&
git ls-files -s >index_files &&
test_line_count = 2 index_files &&
git rev-parse >actual HEAD:bar/bq HEAD:bar/whatever &&
git rev-parse >expect B:bq A:foo/whatever &&
test_cmp expect actual &&
git hash-object bar/bq bar/whatever >actual &&
git rev-parse B:bq A:foo/whatever >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:bq HEAD:foo/bq &&
test_path_is_missing bq &&
test_path_is_missing foo/bq &&
test_path_is_missing foo/whatever
)
'
test_expect_success '3b-R: bq_1->foo/bq_2 on A, foo/->bar/ on B' '
test_setup_3b R &&
(
cd 3b_R &&
git checkout B^0 &&
test_path_is_missing bar/bq &&
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>
2019-04-05 15:00:26 +00:00
GIT_MERGE_VERBOSITY=3 git -c merge.directoryRenames=true merge -s recursive A^0 >out 2>err &&
test_must_be_empty err &&
test_path_is_file bar/bq &&
git ls-files -s >index_files &&
test_line_count = 2 index_files &&
git rev-parse >actual HEAD:bar/bq HEAD:bar/whatever &&
git rev-parse >expect B:bq A:foo/whatever &&
test_cmp expect actual &&
git hash-object bar/bq bar/whatever >actual &&
git rev-parse B:bq A:foo/whatever >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:bq HEAD:foo/bq &&
test_path_is_missing bq &&
test_path_is_missing foo/bq &&
test_path_is_missing foo/whatever
)
'
###########################################################################
# SECTION 4: Cases involving dirty changes
###########################################################################
# Testcase 4a, Changed on A, subset of changes on B, locally modified
# Commit O: b_1
# Commit A: b_2
# Commit B: b_3
# Working copy: b_4
# Expected: b_2 for merge, b_4 in working copy
test_setup_4a () {
test_create_repo 4a &&
(
cd 4a &&
test_write_lines 1 2 3 4 5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 10.5 >b &&
git add b &&
test_tick &&
git commit -m "A" &&
git checkout B &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "B"
)
}
# NOTE: For as long as we continue using unpack_trees() without index_only
# set to true, it will error out on a case like this claiming that the locally
# modified file would be overwritten by the merge. Getting this testcase
# correct requires doing the merge in-memory first, then realizing that no
# updates to the file are necessary, and thus that we can just leave the path
# alone.
test_expect_merge_algorithm failure success '4a: Change on A, change on B subset of A, dirty mods present' '
test_setup_4a &&
(
cd 4a &&
git checkout A^0 &&
echo "File rewritten" >b &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 b >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
# Make sure b was NOT updated
test-tool chmtime --get b >new-mtime &&
test_cmp old-mtime new-mtime &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual :0:b &&
git rev-parse >expect A:b &&
test_cmp expect actual &&
git hash-object b >actual &&
echo "File rewritten" | git hash-object --stdin >expect &&
test_cmp expect actual
)
'
# Testcase 4b, Changed+renamed on A, subset of changes on B, locally modified
# Commit O: b_1
# Commit A: c_2
# Commit B: b_3
# Working copy: c_4
# Expected: c_2
test_setup_4b () {
test_create_repo 4b &&
(
cd 4b &&
test_write_lines 1 2 3 4 5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "O" &&
git branch O &&
git branch A &&
git branch B &&
git checkout A &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 10.5 >b &&
git add b &&
git mv b c &&
test_tick &&
git commit -m "A" &&
git checkout B &&
test_write_lines 1 2 3 4 5 5.5 6 7 8 9 10 >b &&
git add b &&
test_tick &&
git commit -m "B"
)
}
test_expect_success '4b: Rename+Mod(A)/Mod(B), change on B subset of A, dirty mods present' '
test_setup_4b &&
(
cd 4b &&
git checkout A^0 &&
echo "File rewritten" >c &&
t6022, t6046: fix flaky files-are-updated checks Several tests wanted to verify that files were actually modified by a merge, which it would do by checking that the mtime was updated. In order to avoid problems with the merge completing so fast that the mtime at the beginning and end of the operation was the same, these tests would first set the mtime of a file to something "old". This "old" value was usually determined as current system clock minus one second, truncated to the nearest integer. Unfortunately, it appears the system clock and filesystem clock are different and comparing across the two runs into race problems resulting in flaky tests. From https://stackoverflow.com/questions/14392975/timestamp-accuracy-on-ext4-sub-millsecond: date will call the gettimeofday system call which will always return the most accurate time available based on the cached kernel time, adjusted by the CPU cycle time if available to give nanosecond resolution. The timestamps stored in the file system however, are only based on the cached kernel time. ie The time calculated at the last timer interrupt. and from https://apenwarr.ca/log/20181113: Does mtime get set to >= the current time? No, this depends on clock granularity. For example, gettimeofday() can return times in microseconds on my system, but ext4 rounds timestamps down to the previous ~10ms (but not exactly 10ms) increment, with the surprising result that a newly-created file is almost always created in the past: $ python -c " import os, time t0 = time.time() open('testfile', 'w').close() print os.stat('testfile').st_mtime - t0 " -0.00234484672546 So, instead of trying to compare across what are effectively two different clocks, just avoid using the system clock. Any new updates to files have to give an mtime at least as big as what is already in the file, so we could define "old" as one second before the mtime found in the file before the merge starts. But, to avoid problems with leap seconds, ntp updates, filesystems that only provide two second resolution, and other such weirdness, let's just pick an hour before the mtime found in the file before the merge starts. Also, clarify in one test where we check the mtime of different files that it really was intentional. I totally forgot the reasons for that and assumed it was a bug when asked. Reported-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-13 20:03:02 +00:00
test-tool chmtime --get -3600 c >old-mtime &&
GIT_MERGE_VERBOSITY=3 git merge -s recursive B^0 >out 2>err &&
test_must_be_empty err &&
# Make sure c was NOT updated
test-tool chmtime --get c >new-mtime &&
test_cmp old-mtime new-mtime &&
git ls-files -s >index_files &&
test_line_count = 1 index_files &&
git rev-parse >actual :0:c &&
git rev-parse >expect A:c &&
test_cmp expect actual &&
git hash-object c >actual &&
echo "File rewritten" | git hash-object --stdin >expect &&
test_cmp expect actual &&
test_must_fail git rev-parse HEAD:b &&
test_path_is_missing b
)
'
test_done