When testing the freshness of a build command, the fingerprint of an entire
package is generated. The current method of fingerprinting a path source is to
just stat() the entire tree and look at mtimes. This works fine when the build
command places *all* output in the build directory.
Some systems, like autotools, place *most* output in the build directory and
some output in the source directory, however (see spidermonkey). In this case
the coarse-grained "consider all files in a directory as input" is too painful
for the build command as the package will forever be rebuilt.
This commit adds support for filtering the list of files in a directory
considered to be part of a package by using `git ls-files`. This git-specific
logic can be extended to other VCSs when cargo grows support for them.
Closes#379
This commit changes the hash of Profile to only take into account
flags/variables that affect the actual output file itself (as opposed to its
location), and then changes cargo {test, build, doc} to all use the same
directory of output (in order to share deps).
This will cause a `cargo build` to remove all of the tests generated by `cargo
test`, but it speeds up the cycle of `cargo test` followed by a `cargo build` by
not needing to rebuild all dependencies.
Additionally, `cargo bench` now shares the same directory as
`cargo build --release` for the same reasons as above.
Closes#348
Previously an invocation of `cargo build` would generate a lockfile *without*
dev dependencies, but an invocation of `cargo test` would generate a lockfile
*with* dependencies. This causes odd problems and diffs with the lockfile.
This commit switches the resolve-to-be-a-lockfile to using all dependencies of a
package, while the resolve-to-be-compiled continues to use just the dependencies
needed for the current build.
The syntax was originally chosen to perhaps allow multiple libraries in the future, but that is less and less likely to happen at this point. This commit deprecates the now-misleading `[[lib]]` in favor of `[lib]` to indicate that only one can be present.
This does not start allowing `[bin]` or `[example]` and such as I felt that it was too many ways to specify what's essentially the same thing. This also as a bonus allows a top-level `bin = []` to completely opt-out of binary inference (as well as for tests and examples).
Closes#327
We observed that after updating the git repo in a test, unless we waited
for 1s, we got intermittent cases where a subsequent update of the git
source would not pick up the changes.
We observed this both with the git CLI and the branch using libgit2.
There may be a better solution, but we did not find it.
I customized my git repo template to default to the standard pre-commit
hook that checks for e.g. trailing whitespace and refuses to commit if
any trailing whitespace was found. This causes some of cargo's tests to
fail.
To be independent of the user's git template, create the repo for the
test without using any, by specifying an empty template.
The fingerprinting code was erroneously using all sources from a manifest when
calculating the fingerprint which meant that sources not yet downloaded were
attempted to be fingerprinted.
The correct source to fingerprint is located in the SourceId field of the
resolved PackageId.
Closes#259
This ends up serving the same purpose, but a critical change is that it
canonicalizes the relevant git url (if one is used) to ensure that the same
package from two slightly different locations is always built the same way.
Sadly I'm not quite sure how to add a test for this as it involves using remote
git urls which are unusable during tests.
TOML recently grew multiline string literals as well as literal strings with no
escaping, which can be used instead of the hokey escape_path() function.
This allows the dependency queue to properly handle packages with the same
name but from different sources.
A test was added which exercieses this functionality by depending on two
different revs of the same git repo.
* Add a convenience method bin() for generating the name of a binary. On windows
this remembers to append `.exe`.
* Stop executing relative paths to binaries and relying on PATH. This is
suffering from rust-lang/rust#15149 and failing to spawn processes on windows.
Additionally, this allows the tests to work with a pre-installed cargo becuase
the freshly built executables are precisely specified.
* A new function, escape_path(), was added for tests. When generated source
files with paths, this function needs to be called to properly escape the
\-character that appears in windows path names. Without this function we would
be generating invalid TOML and rust.
This commit adds support for recognizing "fingerprints" of upstream
dependencies. When a dependency's fingerprint change, it must be rebuilt.
Currently the fingerprint unconditionally includes the version of the compiler
you're using as well as a specialized version depending on the type of source
you're compiling from:
- git sources return their fingerprint as the current SHA. This will
disregard any local changes.
- path sources return their fingerprint as the maximum mtime of any file found
at the location. This is a little too coarse and may rebuild packages too
often (due to sub-packages), but this should get the job done for now.
When executing `cargo compile`, dependencies are not rebuilt if their
fingerprint remained constant.
