Open dustdfg opened 6 months ago
Also at least the selection
colorgroup is still missing.
BTW I've updated the colorgroups documentation in PR #3203.
Yeah, current approach in micro.yaml
is not quite scalable. Moreover, micro actually allows extending basic syntax colorgroups (type
, identifier
etc) with arbitrary subgroups (type.keyword
, identifier.foo.bar
etc), and that is documented. So we should probably highlight any of those abritrary subgroups, not just a fixed limited set of subgroups. For example, instead of:
- identifier: "\\b(identifier(\\.(class|macro|var))?)\\b"
just:
- identifier: "\\b(identifier(\\.\\S*)?)\\b"
I agree with - identifier: "\\b(identifier(\\.\\S*)?)\\b"
moreover I would say we need to highlight something like \\b(i\\S(\\.\\S*)?)\\b
moreover I would say we need to highlight something like
\\b(i\\S(\\.\\S*)?)\\b
I can't see why?
Wouldn't user want to create custom colors? You can think that all the possible situations are covered and won't new top-level names exist but then someone just wants to create git highlight and then new colorgroup is added. I mean just to highlight any valid by syntax colorgroup
One more rock in the side of the color groups. They aren't coherent in my opinion. diff-added
, diff-modified
3 different top level things and not diff.added
...
As like for error
and tab-error
. Why not error.tab
?
Wouldn't user want to create custom colors? You can think that all the possible situations are covered and won't new top-level names exist but then someone just wants to create git highlight and then new colorgroup is added. I mean just to highlight any valid by syntax colorgroup
Well, perhaps... OTOH then completely unsupported top-level names (e.g. if the user makes a typo) would be highlighted as well.
One more rock in the side of the color groups. They aren't coherent in my opinion.
diff-added
,diff-modified
3 different top level things and notdiff.added
...
Perhaps, but changing them now would mean at least breaking compatibility.
As like for
error
andtab-error
. Why noterror.tab
?
Hmm, is error
used at all?
Speaking of specifically diff-added
, diff-modified
etc: making them subgroups of diff
would make sense if there was a use case for highlighting them all with the same color (so then specifying just the diff
color would be enough). But I doubt there are users that would want that. :)
Hmm, is
error
used at all?
Looks like no. Grep said about gutter-error
, tab-error
, error-message
but no plain error
in go files
Honestly I don't think current approach is flexible