May be relevant to #2358 (which I can't reproduce on Windows). You can view, comment on, or merge this pull request online at:
https://github.com/geany/geany/pull/2360
-- Commit Summary --
* Pascal: Use separate kind for `type` identifiers
-- File Changes --
M ctags/parsers/pascal.c (5) M src/tagmanager/tm_parser.c (1)
-- Patch Links --
https://github.com/geany/geany/pull/2360.patch https://github.com/geany/geany/pull/2360.diff
I need to update bug612019.pas.tags.
@ntrel you need to find why the tests pass on windows and fail on Linux, and that might help #2358 that you can't reproduce on Windows.
Also you need to push upstream, otherwise its likely to get overwritten by the next update from ctags.
Also you need to push upstream, otherwise its likely to get overwritten by the next update from ctags.
If the parser has already diverged from upstream as suggested in the (edited) description, then while it would be nice to sync all the changes back upstream, it probably shouldn't be a blocker for this PR to do it.
Well, if local changes are not pushed upstream it will never get synced, and I said nothing about it being a blocker for this.
You said @ntrel needs to push upstream, but it sounds like it was already diverged, so IMO _he_ doesn't have to do it (inferred "for this PR"), just that ideally "someone" would do it eventually.
(inferred "for this PR")
Any inference that I said this PR should depend on upstream accepting the change is incorrect, I didn't say that at all.
This PR is to fix a crash (see #2358) so a short term local patch is acceptable, but if changes are not pushed upstream then Geany will permanently diverge again.
Any inference that I said this PR should depend on upstream accepting the change is incorrect, I didn't say that at all.
Not just whether or not they accept the changes, the act of someone going through the forked parser and upstream parser and manually syncing any unrelated changes is out of scope for this PR, IMO.
but if changes are not pushed upstream then Geany will permanently diverge again.
The description mentions that they've already diverged, so the Pascal parser was already out of sync and permanently diverged (until someone syncs it).
so the Pascal parser was already out of sync and permanently diverged (until someone syncs it).
And thats why it needs some annoying guy to keep mentioning it :)
This PR is to fix a crash
Not really, it's possible it does but that wasn't the main reason.
if changes are not pushed upstream then Geany will permanently diverge again
Parsing the first identifier after `type` is already a divergence. ATM I'm not confident of its correct parsing, this pull just makes geany see the tag differently.
Updated tests, don't set an arglist & vartype for `type` tags, fix parsing comment before identifier - aae0a21. That commit may fix #2358, perhaps along with the defensive nulling of c46d2df.
I've made https://github.com/universal-ctags/ctags/pull/2241 but that needs ctags tests.
@ntrel, travis seems to like it, what about adding the problem from #2358 to bug612019.pas or making a new test to test it?
github-comments@lists.geany.org