_get_mmds_from_requires does not do anything against db.session. Hence,
remove parameter session from it and
get_mmds_required_by_module_recursively.
Signed-off-by: Chenxiong Qi <cqi@redhat.com>
TestUtilsModuleStreamExpansion._make_module is moved to
tests/__init__.py so that other tests could reuse this to create a
module build.
Signed-off-by: Chenxiong Qi <cqi@redhat.com>
Modularity team needs to be able to define multiple streams for single "platform"
module. They need it to express that "platform:el8.1.0" also provides "platform:el8".
This needs changes in a way how MMDResolver resolves dependencies between modules.
For example, if we are building module against platform:el8.1.0, the MMDResolver must
not return buildrequired module built against platform > el8.1.0, but it can for example
return (cherry-pick) buildrequired module from platform:el8.0.0 if there is no such
module built for platform:el8.1.0.
The way how it is implemented is following:
- MMDResolver reads list of virtual streams from xmd["mbs"]["virtual_streams"] when
creating Solvable from MMD and adds additional Provides for these virtual streams.
We expect these to be set mainly on base modules. The versions of such provides
are based on "stream version" number, so we can compare them.
- The base module ("platform") buildrequires of MMDs added to MMDResolver are overriden
to mark particular platform "stream version". For example, if module "foo" buildrequires
"platform:el8" in MMD file, but was in fact built against platform:el8.1.0, it will
be treated as if it would really buildrequire "platform:el8.1.0". This is needed
to not include buildrequired modules built against newer platform than what we want.
- MMDResolver resolves all the valid combinations of buildrequires, but we are only
interested in the one with latest versions of buildrequired module. Therefore the
solve() method is changed to find out combinations which have the latest versions
for each alternative name:stream buildrequires.
This reverts commit 8173040ea6. Since it didn't
provide any real benefit after talking with sgallagh, it seems like we'd better
off not introduce more risk to the next deployment.
MBS has been using the libmodulemd API incorrectly by assuming that
methods like `get_rpm_components` return the actual object in memory
and not a copy. This has been true but wasn't something sgallagh
intended. We found this out after he had me test his new version of
libmodulemd. This PR removes those assumptions.
I discovered that local builds have been broken by recent (good) changes in
dnf. At the end of every batch, we regenerate the local repo with createrepo
and we also call modifyrepo to include the modulemd file as we progress. This
was added in #467.
What we really want, is for the modulemd file to be present at the *end* of the
build.
Recently, dnf started respecting the modulemd file natively, such that builds
we built in batch0 would not show up in batch1. They would be present in the
repo, but they would be marked as belonging to a module which was not enabled,
and so could not be pulled in. Every module build would fail because
module-srpm-macros was in a disabled module (the module being built at the
time).
This change makes it so that the module metadata is only added to the repo at
the very end of the build. I moved it into a `finalize` method on the builder
which the copr builder was using, and for symmetry's sake I moved the koji
content generator code to the same method on that builder.
There was a circular import issue to solve between the koji module builder and
the koji content generator modules that generated a larger diff, but is mostly
cosmetic and mock changes.
1. Changed ModuleBuild's context property to db column, it's
non-nullable and default value is '00000000' to keep it consistent
with previous behaviour.
2. Changed ModuleBuild.contexts_from_mmd to return a tuple of
(ref_build_context, build_context, runtime_context, context).
3. Updated tests affected by this change.
koji now supports tags with max length of 256, we can use
more informative tag name instead of the hash one.
The new format of koji tag name is:
module-<name>-<stream>-<version>-<context>
However when the generated tag's length > (256 - len('build')), we
fallback to the old way of name in hash format (module-<hash>).
In this change, koji tag is always generated from MBS itself, even
with pdc resolver.
FIXES: #918#925
