When a issue is moved from one project to another, all associated
Markdown text is rewritten in the context of the new project. If the
note contained a link to a commit URL, `CommitRewriter#rewrite` would
fail because `Commit#link_reference_pattern` would match `nil` `commit`
values in the HTML generated from the Markdown. These `nil` values were
passed along to `Project#commits_by` because `Commit#reference_valid?`
was always returning `true`.
To prevent this issue from happening, we tighten up the check for
`Commit#reference_valid?` to look for valid SHA values.
Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/66666
We've seen a significant performance penalty when using
`BatchLoader#__replace_with!`. This defines methods on the batch loader
that proxy to the 'real' object using send. The alternative is
`method_missing`, which is slower. However, we've noticed that
`method_missing` can be faster if:
1. The objects being loaded have a large interface.
2. We don't call too many methods on the loaded object.
In production, we've seen the rendering times of the merge request
widget increase as a result of loading commit data. BatchLoader attempts
to call replace_methods on the lazy object, but this has a significant
performance penalty. Disabling this mode
(https://github.com/exAspArk/batch-loader/pull/45) appears to cut load
times by about 50% for MergeRequestsController#show.
Relates to https://gitlab.com/gitlab-com/gl-infra/infrastructure/issues/6941
This brings back some of the changes in
https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/20176/diffs.
We discovered another N+1 that hits Gitaly `TreeEntry` via the
`RelativeLinkFilter`:
https://gitlab.com/gitlab-org/gitlab-ce/issues/58657. When a blob is
loaded with many relative links, `TreeEntry` is called for each link to
scan the URI type.
There are multiple paths that hit Gitaly `TreeEntry`, and
https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/25706 did not
cover all cases. This commit covers another common use case.
For users using Gitaly on top of NFS, accessing the Git data directly
via Rugged may be faster than going through than Gitaly. This merge
request introduces the feature flag `rugged_commit_tree_entry` to
activate the Rugged method.
Private commit emails were introduced in !22560, but some parts of
GitLab were not updated to take account of them. This commit adds
support in places that were missed.
Even if it doesn’t save lines of code, since people will tend to use
code they’ve seen. And `SafeRequestStore` is safer since you
don’t have to remember to check `RequestStore.active?`.
When displaying the pipelines of a project we now preload the following
data:
1. Authors of the commits that belong to these pipelines
2. The number of warnings per pipeline, which is used by
Ci::Pipeline#has_warnings?
== Commit Authors
Previously this data was queried for every Commit separately, leading to
20 SQL queries being executed in the worst case. With an average of 3 to
5 milliseconds per SQL query this could result in 100 milliseconds being
spent in _just_ getting Commit authors.
To preload this data Commit#author now uses BatchLoader (through
Commit#lazy_author), and a separate module
Gitlab::Ci::Pipeline::Preloader is used to ensure all authors are loaded
before they are used.
== Number of warnings
This changes Ci::Pipeline#has_warnings? so it supports preloading of the
number of warnings per pipeline. This removes the need for executing a
COUNT(*) query for every pipeline just to see if it has any warnings or
not.
Prior to this change, this was done through unicorn. In theory this
could time out. Workhorse has been sending these raw patches and diffs
for a long time and is stable in doing so.
Added bonus is the fact that `Commit#to_patch` can be removed.
`Commit#to_diff` too, which closes
https://gitlab.com/gitlab-org/gitaly/issues/324
Closes https://gitlab.com/gitlab-org/gitaly/issues/1196
Uses `list_commits_by_oid` on the CommitService, to request the needed
commits for pipelines. These commits are needed to display the user that
created the commit and the commit title.
This includes fixes for tests failing that depended on the commit
being `nil`. However, now these are batch loaded, this doesn't happen
anymore and the commits are an instance of BatchLoader.
GitLab Bot
Stan Hu
Heinrich Lee Yu
Thong Kuah
Kamil Trzciński
Nick Thomas
Rémy Coutable
George Tsiolis
Michael Kozono
Sean McGivern
yisiliang
Jacob Vosmaer
Yorick Peterse
Zeger-Jan van de Weg
Ahmad Sherif
Hiroyuki Sato
blackst0ne
Saverio Miroddi
Alejandro Rodríguez
Robert Speicher