The Koji tag extra options used to be hard-coded and to change them,
we had to release new MBS version.
We do not change them often, but right now fedora-infra is requesting
to use new `mock.new_chroot` option and we need to release new MBS
because of that.
This commit makes such changes easier in the future.
When importing a base module, we must ensure the value that will be
used in the RPM disttags doesn't contain a dash since a dash isn't
allowed in the release field of the NVR.
MBS uses the base module's stream that was buildrequired by the module
in the RPM disttags for that module build. The stream name may not be
ideal for all situations, so now this is customizable by setting the
xmd['mbs']['disttag_marking'] in the base module's modulemd.
Currently, the Koji tags have properly set the scrmod- prefix, but
the build target still sets module- prefix even for scratch builds.
In this commit, build target has the scrmod- prefix too.
The current code reads the data, converts them to unicode string and
then uses the `len()` of that string as filesize. This is wrong,
because Koji expects filesize to be really number of bytes, not number
of characters.
Therefore, in this commit, the filesize is computed from raw data (bytes).
The `ImportModuleAPI` calls the `auth.get_user()` which auths the
user using Kerberos. the `ImportModuleAPI` later calls `SCMHandler`
which in its `__init__` method calls the `auth.get_user()` again.
This leads to traceback in GSSAPI, because the user is already
authed.
This commit fixes this by caching the auth results in `flask.g`,
which is reset after each request based on the Note in
http://flask.pocoo.org/docs/1.0/appcontext/#storing-data.
This commit also marks mutual auth as OPTIONAL in `mbs-cli`,
because MBS server currently does not do mutual auth.
The `KojiContentGenerator.finalize()` needs the `ModuleBuild.time_completed`.
We currently set `time_completed` once the module build transitions
into `done` state. But we have moved the `KojiContentGenerator` call
to end of `build` state, so right now it is called before the `time_completed`
is set. This leads to traceback.
In this commit, the `time_completed` is set before the `KojiContentGenerator`
call, so it is defined properly.
- Keep scratch module builds in the 'done' state.
- Make koji tagging for scratch modules unique so the same
commit can be resubmitted.
- Use alternate prefix for scratch module build components so they can
be identified later.
- Prevent scratch build components from being reused.
- Assorted code and comment cleanup.
Signed-off-by: Merlin Mathesius <mmathesi@redhat.com>
We encountered an issue where start_next_batch_build was called twice on a
module build and it caused the module build to fail because Brew failed
one of the components since it was submitted twice. This will help us narrow
down the issue if it happens again.
In case there is lot of components in a module build or there are some
networking issues and we need to retry the "git clone" commands, the
`format_mmd` method can take long time.
If it takes more than 10 minutes, the poller can produce fake event,
because it seems the module build is stuck. This is wrong, because
it would lead to another unexpected init handler call.
In this commit, the `format_mmd` updates the `ModuleBuild.time_modified`
method regularly to prevent poller from sending the unexpected fake
init message.
This is required for monitoring use-cases, where we can have a Kerberos principal for a
service account but no associated account in LDAP to check group membership.
This commit introduces new to_text_type helper method and calls it
for return value of all mmd.dumps() calls. That way, we always
end up with proper unicode string represntation on both python
major versions.
This commit also adds unicode character to description of all
the yaml files we use in the tests so we can be sure MBS can
handle unicode characters properly.
This might be temporary fix, depending on the result of discussion
at https://github.com/fedora-modularity/libmodulemd/issues/184.
This reverts commit ae79b711d8.
This breaks MBS with python3 and fixes only single occurence of this
issue. We want to fix this in libmodulemd or find a better way how
to fix this to work on both python2 and python3.
We have libmodulemd PR open to address this issue on libmodulemd
level: https://github.com/fedora-modularity/libmodulemd/issues/184.
We always set the "context" to DEFAULT_MODULE_CONTEXT from the historical
reasons - the context was not stored in the database before and we stored
just the build_context/runtime_context. But this is no longer true for
some time.
In this commit, the context is respected and stored in the database when
importing module using the `import_mmd` method. If the context is not set
in the imported MMD, the DEFAULT_MODULE_CONTEXT is used.
This is needed for Fedora branching, but it is generally useful.
For example, there is a module buildrequiring "platform:[]". In the time
this module has been built, only platform:f29 existed, so it has been built
just against platform:f29.
After a while, the platform:f30 is released and the maintainer needs
to rebuild the module against platform:f30. Right now, he needs to create
new commit in the module and submit the build, but this will result in useless
rebuild of the module also against platform:f29.
In this commit, MBS allows to resubmit the module build in a case
there are new MSE builds to build. MBS will send all the module builds
back to the user - so the existing builds will be already marked as
"ready" and the newly submitted builds will have the "init" state in
the REST API response.
However, in case when there are no new MSE builds to build, MBS still
sends back the Conflict error as it used to. This is done for backwards
compatibility and also to not confuse the users in case when no new build
has been submitted.
With this command, admins can retire module builds that should no longer
be used as a dependency for other module builds.
Fixes#1021
Signed-off-by: Luiz Carvalho <lucarval@redhat.com>
Due to a logic error, if record_filtered_rpms() was called on a module which
already had filtered_rpms set, the buildrequires xmd information was discarded.
This broke all local builds.
When encoutering a Windows end of line (^M), io.open and open in Python 3
will convert those to UNIX end of lines by default. When reading logs
to compute the checksum, it's important those new lines aren't converted,
to ensure the checksum is correct. This caused issues in Fedora staging
because when cloning down a repo, the repoSpanner output had Windows end
of lines, and this would end up in build.log. The solution is to just read
it as binary so that Python doesn't perform these conversions.
