With MyST relative references to other components in firefox-source-docs are broken
Categories
(Developer Infrastructure :: Source Documentation, defect)
Tracking
(firefox-esr91 unaffected, firefox95 unaffected, firefox96 unaffected, firefox97 wontfix, firefox98 fixed)
Tracking | Status | |
---|---|---|
firefox-esr91 | --- | unaffected |
firefox95 | --- | unaffected |
firefox96 | --- | unaffected |
firefox97 | --- | wontfix |
firefox98 | --- | fixed |
People
(Reporter: whimboo, Assigned: jgraham)
References
(Regressed 1 open bug, Regression)
Details
(Keywords: regression)
Attachments
(10 files)
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review |
Since bug 1746782 has been landed all the relative references at least in the /remote/doc
and /testing/geckodriver/doc
files are broken and links are no longer existent on rendered pages.
For example check mozconfig
at:
https://firefox-source-docs.mozilla.org/testing/marionette/Building.html
There are a lot more references that are affected. Sadly it wasn't noticed when landing bug 1746782 that these are busted now. I wonder how more widespread this issue is.
Comment 1•3 years ago
|
||
We could likely backout bug 1746782, though I wouldn't expect anyone to attempt re-landing (doc infrastructure is unowned).. so we'd lose out on the benefits myst
provides and be on the deprecated recommonmark.
Afaict, the only other alternative is to fix the regressions. I don't think the fixes would be too hard, but I don't have time for it.. and if no one else has time, backing out might be the only option.
I can't resist mentioning that if I were maintaining docs in markdown.. I'd be very excited about myst
:p
Reporter | ||
Comment 2•3 years ago
|
||
Andrew, could you maybe repeat the proposed solutions as you mentioned on Element? Maybe you have some more information or an example too? I assume there is no solution that works for both? If we are going to update our own docs others might still be busted, and reverting the patch would render our changes as broken then.
Comment 3•3 years ago
|
||
The options are outlined in comment 1.
As for examples, it depends on what you're trying to do, but there are some examples here:
https://myst-parser.readthedocs.io/en/latest/sphinx/use.html
Looks like you can link to docs like:
[link](path/to/doc.md)
You could also probably use the :doc:
role (e.g in in rst) or the :ref:
role (which is more robust than manually linking to a doc as it will survive renames).
But I've never used myst
myself, so you'll have to experiment.
Comment 4•3 years ago
|
||
Set release status flags based on info from the regressing bug 1746782
Assignee | ||
Comment 5•3 years ago
|
||
Updated•3 years ago
|
Assignee | ||
Comment 6•3 years ago
|
||
These were broken since the update to MyST.
Assignee | ||
Comment 7•3 years ago
|
||
Assignee | ||
Comment 8•3 years ago
|
||
Assignee | ||
Comment 9•3 years ago
|
||
Assignee | ||
Comment 10•3 years ago
|
||
Assignee | ||
Comment 11•3 years ago
|
||
These were broken since we adopted MyST.
Assignee | ||
Comment 12•3 years ago
|
||
These were broken since the update to MyST.
Assignee | ||
Comment 13•3 years ago
|
||
sphinx-js is apparently trying to link these to another object. In the interests
of removing all 'reference target not found' type warnings, I just removed the
@see annotations, since it wasn't trivial to work out if there was something
they could point at the would work in the generated docs.
Assignee | ||
Comment 14•3 years ago
|
||
The idea is to capture the warnings in a temporary file, and then
apply a set of regex to find any that should be treated as fatal.
This allows us to fix warnings one type at a time, and prevents us
regressing the warnings that are already fixed.
The "reference target not count" warning is added to the initial
forbidden list, so we can ensure we don't end up with internal links
pointing to nowhere.
Updated•3 years ago
|
Reporter | ||
Updated•3 years ago
|
Comment 15•3 years ago
|
||
Comment 16•3 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/d9b01f388d27
https://hg.mozilla.org/mozilla-central/rev/72207858d8c7
https://hg.mozilla.org/mozilla-central/rev/dc15c73282cf
https://hg.mozilla.org/mozilla-central/rev/33babc343579
https://hg.mozilla.org/mozilla-central/rev/e6b91230a0c0
https://hg.mozilla.org/mozilla-central/rev/334f96cdbe67
https://hg.mozilla.org/mozilla-central/rev/db2f5380d783
https://hg.mozilla.org/mozilla-central/rev/8bf9a5e9580b
https://hg.mozilla.org/mozilla-central/rev/d742452c2851
https://hg.mozilla.org/mozilla-central/rev/6a3d6ade80e6
Comment 17•3 years ago
|
||
Is this something we need on Beta too or is riding 98 to release sufficient?
Assignee | ||
Comment 18•3 years ago
|
||
I think riding the trains to release is fine; this only really affects https://firefox-source-docs.mozilla.org/ and not anything we ship to end users.
Updated•3 years ago
|
Updated•3 years ago
|
Updated•2 years ago
|
Description
•