| Age | Commit message (Collapse) | Author |
|---|
|
Preparatory commit for implementation of '--clone' option.
|
|
|
|
Marking a source repository 'as plain' means that the whole source
repository tree will get imported as a repository type
corresponding to the source type. In this case, additional pragmas
than those supported by the inndividual imports might need to be
set.
Solve this by supporting the just-mr-style 'pragma' field also in
the source description, for all sources also accepting the
'as plain' field. Currently support only the 'special' pragma.
Document change and add test for plain imports that checks this
feature.
|
|
In particular, any transitive 'file'-type repository will inherit
any given '{to_git: true}' pragma in the import description
objects. Note that this technically can only happen for transitive
'file' repositories imported from a 'file' source, so in all other
cases such a pragma would not have any effect.
Document change and extend the import from 'file' source test to
check this feature.
|
|
Update CHANGELOG accordingly.
|
|
Transitive 'file'-type repositories should inherit any pragmas
suported by the new repository type they are rewritten as.
Extend import tests to avoid regressing on this issue in the
future.
|
|
...before trying remotes.
|
|
While there, improve docstrings of methods operating on Git
repositories to specify when the underlying Git repositories are
expected to be present.
|
|
Based on f0473730 and 22ef0a16
|
|
Based on 76ce2123
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
...below the GitCAS and CAS utilities, in order for those to be
available for use when implementing the other import source types.
|
|
Enforce 'subdir' field of 'archive' import source to also be
non-upward.
Normalize subdirs during rewrite of dependent 'file'-type
repositories.
|
|
|
|
|
|
The rewrite logic assumes that for repositories taken as layers we
will not have certain fields and so they are unconditionally
reassigned.
Fix this by extending the existing check for bindings to include
also alternative roots and root files.
|
|
Use local CAS to cache fetched archives. Take the storage garbage
collector shared lock to guard against unwanted generation
rotations.
|
|
Add file locking logic; guard critical Git operations with the
same approach as in Just and also take the repositories garbage
collector shared lock to guard against unwanted generation
rotations.
|
|
When commit to fetch is known, check mirrors before going to the
main URL.
While there, ensure the context for logging in the checkout method is
not replaced, but appended to.
|
|
|
|
...to better reflect intent. In particular, improve the name of the
variable storing the partial description to be used during rewrite
of dependent 'file'-type repositories brought in by an import.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Use the RGB values of the predefined fmt library colors that both
just and just-mr use.
|
|
While there, also improve check for early return from import
method.
|
|
Able to handle imports from Git repositories and deduplicate the
resulting configuration.
|
|
|
|
|
|
As per documentation, if a location is given for the configuration
file in the foreign repository then it must be taken relative to
its root. However, in our implementation this location is taken
relative to the current directory.
Fix this.
|
|
The NoReturn type hint should be used to ensure the return type of
methods calling no-return methods are not falsely enforced to an
Optional return type.
Add the NoReturn type hint as needed and clear up existing Optional
returns together with any corresponding casts affected by the
above.
While there, also fix formatting.
|
|
Add a basic implementation of computed roots. So far we neither
enforce nor make good use of the fact that the underlying target
of a computed root has to be an export target of a content-fixed
repository; instead, we always install the root to a temporary
directory and pick it up from there. Still, the basic functionality
is there, and it is minimally useful.
|
|
|
|
... by recursively copying and resolving sybolic links. In this
way, references within a tree (e.g., #include "../Something.hpp")
do not confuse the compiler, as opposed to directory symlinks.
Neverthess, by doing this copying only for tree constructors, we
still keep the overhead acceptable.
|
|
The Python script used for the first stage of bootstrapping just
uses the Git index to create trees from directories (be it file
repositories, unpacked archives, or distfiles), therefore it has
the limitations of Git itself in committing trees that contain
entries ignored by Git, such as empty directories, the .git folder,
.gitignore files and files referenced there, or other entries with
Git-specific magic names.
This commit updates the Python script to replace the use of the
Git index for importing directories to directly writing the needed
blobs and trees to the object database, then commit the resulting
top tree explicitly.
While there, fix a typing issue from our relaxed approach in using
os.environ to set the subprocess env when running commands. As on
POSIX the type is _Environ, not simple Dict[str, str], use implicit
dictionary merging (Python v3.5+) to set the Git envars. The issue
was initially flagged by pyright.
|
|
|
|
...with a default implementation.
Although it is not used directly, it will be needed for instantiation of std::optional.
|
|
While just-deduplcate-repos only identifies repositories that are
semantically indistinguishable, there still is the pragma to consider.
- If the repository in one use requires to be present, we have to
fetch it, hence cannot treat it as absent.
- If a repository in one place has to be transformed to a git root,
this also has to happen for the copy we keep.
|
|
This allows to simply specify the direct dependencies desired
to be absent at import where the dependency structure is handled
anyway, rather than afterwards reconstructing which repositories
should be absent.
|
|
When importing a repository via just-import-git, allow to
specify the value for the "inherit env" property for the
repository just being imported.
|
