Page MenuHomePhabricator

[GTWL] Investigation: increasing the length of the edit summary
Closed, ResolvedPublic2 Estimated Story Points


As there is already existing code, we should find out:

What is the current status?
What can we do there to help?

Event Timeline

KasiaWMDE raised the priority of this task from to Medium.
KasiaWMDE updated the task description. (Show Details)

It looks like this is all also blocked on ?

As for see below for what is still to be done there.
Quotes in here are from bawolff (owner of the original patch).


Another rebase of course!

Whether or not everything has to be fixed in the first version of the
patch is debatable, so some of them could conceivably be delayed for
follow-up fixes later. Although it would be nice to convert everything
at once.

Each of these should be trivial (2/3pt)?

  • File delete form doesn't use new limit in UI - In FileDeleteForm::showForm()
  • Confirm delete not fixed in UI - input box in Article::confirmDelete()
  • revdelete not fixed in UI - inputbox in SpecialRevisionDelete::showForm() needs the new limit
  • special:userrights not fixed in UI - input box in UserrightsPage::switchForm() needs the new limit


May need to think through behaviour of auto-summaries/section title more (in backend)

I think all I meant by that, is that auto-summaries will often put
part of the article contents in the edit summary (And section titles
put /* title */). Should we change the cut-off for how much of the
article we put in the edit summary, now that we have longer summaries,
or does the current cut off still make sense as we don't want super
long auto-edit summaries.

Would it be more appropriate EditPage::MAX_SUMMARY_BYTES to actually be in a different class (e.g. Content).

This is a bit bike shed-y. The constant might make more sense being part of a different class.

Maybe maxlength attribute should actually be min( bytes, codepoint )

e.g. The actual attribute in html, as a fallback for non-js browsers.
maxlength is in units of UTF-16 code points (e.g. Astral characters =
2, other stuff = 1), so its not really clear what the best fallback
is, as we only care about utf-8.

Done per the above.
This task was just about investigation.