We have a fallback route in `meson subprojects download` and friends,
which tries to retrieve wrapdb urls via http, if Python was not built
with SSL support.
Stop doing this. Replace it with a command line option to specify that
insecure downloads are wanted, and reference it in the error message if
downloading fails due to SSL issues.
In some cases, init variables that accept None as a sentinel and
immediately overwrite with [], are migrated to dataclass field
factories. \o/
Note: dataclasses by default cannot provide eq methods, as they then
become unhashable. In the future we may wish to opt into declaring them
frozen, instead/additionally.
This will re-apply the meson.build patch overlay, ensuring it is up to
date. Also take the opportunity offered by this infrastructure to
repatch when performing `update --reset` since internally this will run
`git stash` and thus cause the (possibly locally modified) meson.build
files to disappear.
This reverts commit 4568482316.
As it turns out, the rationale for this was completely bogus. This
command doesn't re-apply the patch_directory etc. and in fact there is
no command whatsoever that does this. So, this command does not have two
purposes, and we are not making one of them more robust -- instead it
has one purpose, and we are making it lie about whether it failed.
Instead of trying to freeload off of this command, we will just add
another command to properly apply patch overlays.
This command is useful to e.g. update a cloned subproject which does not
have its packagefiles merged, or which has updated packagefiles. This
does not strictly require internet, so if we can satisfy the ref, simply
log a warning and continue.
This enables a workflow where for network-isolated builds, the
subproject is cloned and moved into place and later on the build
orchestrator runs `meson subprojects update` without a network interface
to initialize the subproject.
If the command fails for one subproject we should still continue with
others. Failed subprojects are reported at the end.
This issue became more problematic when doing parallel tasks because the
reason the command fails was completely hidden by other parallel tasks.
Logs are printed only when the task is done to not interleave logs while
running multiple tasks in parallel. That means that nothing is printed
until the task is done and that could take time.
This moves all the code into a class and call its run() method in a
thread. The class queues all logs to print them at the end to avoid
mixing output of multiple actions.
This will help facilitate cache busting in certain situations, and
replaces hand-rolled solutions of writing a length command to remove
various files/folders within the subprojects directory.
When using --reset we should guarantee that next reconfigure will pick
the latest code. For wrap-file we have no way to know if the revision
changed, so we have to delete the source tree and extract again.
It is unlikely that user has local changes in non-git subprojects, and
--reset is known to be dangerous.
Otherwise it's not possible to share git subprojects via worktrees
when creating a worktree of a git repository that uses meson
subprojects.
The downside is that the user needs to be careful while adding commits
to each tree's index.
wraps from subprojects are now merged into the list of wraps from main
project, so they can be used to download dependencies of dependencies
instead of having to promote wraps manually. If multiple projects
provides the same wrap file, the first one to be configured wins.
This also fix usage of sub-subproject that don't have wrap files. We can
now configure B when its source tree is at
`subprojects/A/subprojects/B/`. This has the implication that we cannot
assume that subproject "foo" is at `self.subproject_dir / 'foo'` any
more.
It was done to include them in `meson subprojects foreach` without
--types argument, but it's better to special case missing --types and
include wraps that have type=None too. It was a bad idea because that
was messing them in `meson subprojects update`, now they are ignored by
that command.
If revision is a tag that does not exist locally, `git fetch origin
<revision>` won't create it and checkout will fail. Using --refmap
ensures that references exists locally.
Besides refactoring code into smaller functions:
- Makes the --rebase behaviour the default for consistency: it was
already rebasing when current branch and revision are the same, it is
less confusing to rebase when they are different too.
- Add --reset mode that checkout the new branch and hard reset that
branch to remote commit. This new mode guarantees that every
subproject are exactly at the wrap's revision.
- Local changes are always stashed first to avoid any data loss. In the
worst case scenario the user can always check reflog and stash list to
rollback.
Fixes: #7526
If the command fails on some subprojects continue with the rest but
return non-0 code. This is useful for CI scripts to ensure it tests
latest code instead of old cached code in case of network error or
something.
D lang compilers have an option -release (or similar) which turns off
asserts, contracts, and other runtime type checking. This patch wires
that up to the b_ndebug flag.
Fixes#7082
Eli Schwartz
Paweł Marczewski
Tristan Partin
Andrei Alexeyev
Xavier Claessens
Nirbheek Chauhan
Dylan Baker
Daniel Mensinger