A minor change, it should probably be:
"$TM_MATE" --line 32:3 --mark "error:unexpected identifier"
/path/to/file
This way we can use the same option for errors, warnings, and set
bookmarks or search marks from ‘mate’.
Yeah, good point. It should also support passing multiple --mark flags,
to allow setting multiple marks for a line.
You mean that the code that segfaults could be your custom paragraph_t
function? That does sound plausible if you copy/pasted the code from
layout_t::line_record_for to get the paragraph_t object (since that code
should work).
Currently I'm using "iterator_at", but that's probably not right.
The layout_t holds a list (tree) of paragraph_t objects which each
correspond to a single line in the source buffer.
The paragraph_t has a list of paragraph_t::node_t objects, and
paragraph_t::iterator_at returns a pointer to the node corresponding to
a particular offset into the line. The line is broken up into nodes to
handle things like soft wrap, tabs, or folding markers.
The ct::line_t object held by the paragraph_t::node_t object is a
shallow wrapper for a CoreText CTLineRef and lazily created (and
disposed if the source buffer gets updated).
So you do not want to drill all the way down to ct::line_t
I manged to drill down to ct::line_t. I kind of works, but not entirely
as I want to. As soon as I edit a marked line the background color is
removed.
As of such, I do think this actually belongs in buffer_t and then
layout_t will read the proper attributes from the source buffer (during
drawing or when constructing/updating the layout).
Probably a better idea. I did had some problem with the drawing, see above.
But if you want to make it an attribute of the layout_t then it should
probably be set on paragraph_t and this object will pass it as an
argument when it calls draw_background on its children.
Ok, I see. Thanks.
--
/Jacob Carlborg
_______________________________________________
textmate-dev mailing list
textmate-dev@lists.macromates.com
http://lists.macromates.com/listinfo/textmate-dev
