Re: [PATCH] Highlight ANSI sequences in the whole buffer (was [PATCH] ANSI color on example blocks and fixed width elements)

[Nathaniel Nicandro]

Hello,

I've gotten around to working on this again.  Sorry for the long
periods between updates.

Ihor Radchenko <yantar92@posteo.net> writes:

>> You want me to remove the code that is responsible for keeping the
>> sequences visible according to `org-fold-show-set-visibility` and
>> according to whether or not the sequence is currently being edited
>> then?
>
>Yes.

Done.

> For now, let's default `org-ansi-hide-sequences' to nil.

Done.

With this patch, I have implemented the extend the font-lock region
idea to account for editing around elements that have highlights which
need to be maintained across edits, e.g. due to a split in a
paragraph.

The additional idea to the previous patch, other than the extend
region stuff, is the packing of the `org-ansi-context` variable,
essentially the same as the `ansi-color-context-region` variable, into
an integer representation and storing it as the `org-ansi-context`
text property (discussed in a previous email) of the highlighted
regions.  This is done to make that text property `eq` comparable for
determining the extent of the highlight at point
(`org-ansi-extent-of-context`).

One other thing to note is that the ANSI sequences don't really play
well with `flyspell-mode`.  If you have something like

[42mparagraph

The end byte of the sequence, the 'm', causes flyspell to think the
word is mparagraph instead of paragraph.  Haven't looked into this at
all yet.  Wondering if anyone is more familiar with flyspell has an
ideas as to a solution.

Please review the patch and tell me what you think.  I'm happy with
the solution that I've come up with so far.  There could be more tests
for the extend region functionality, but I think its simple enough.  I
have extensive tests for the essential highlighting feature though,
they could be a little cleaner though.  There are not too many tests
for the maintenance of highlighting across edits, developing unit
tests for ensuring the highlights are proper after an edit is a little
cumbersome given my understanding of font-lock.  I wonder if there is
any better approach.

0001-Highlight-ANSI-escape-sequences.patch
Description: patch

-- 
Nathaniel