Ar an chéad lá de mí Aibréan, scríobh Stephen J. Turnbull:
That said, there are two issues that I'd like insight into (no
on your patch, as they apply to the original version too). (They're
flagged with #### in the patch below.)
1. I wonder if there might be efficiency implications to the
alignment of text_storage, since in many cases we'll be storing
shorts or integers into it.
2. I wonder what happens if `ccl_setup_program' fails? Aidan, you're
the expert on CCL, any ideas?
If setup_ccl_program fails, we don’t call the CCL program, and garbage gets
displayed (more exactly, our internal coding mapped to the positions in the
font). We’ll only crash if displaying that garbage is something the X server
can’t manage, but most of them do, especially since that XFree86 bug fix
involving zero-width tab characters was integrated.
Also, I'd like to suggest the patch below (to be applied *on top
your patch). It does the following:
1. Updates the big header comments.
2. Add modest header comments to the Xft/no-Mule and no-Xft/Mule
3. Cleans up a couple of inline comments.
4. Removes some trailing whitespace.
5. Always initialize prev_charset to Qunbound, and nuke the comment
(I think it's unnecessary and potentially could be misleading if
(for example) Qnil came to represent the internal charset
ichar_charset returns Vcharset_ascii, defined to Qnil on non-Mule. The
comment was wrong, though still vaguely helpful; it’s better gone.
6. Reorganize separate_textual_runs_mule and add comments to
what's going on.
Where you’ve rewritten comments, your rewrites are more opaque to me than
what was there before. But then I wrote what was there before (still long
enough ago to need to understand it though), so maybe that’s not surprising.
On the quay of the little Black Sea port, where the rescued pair came once
more into contact with civilization, Dobrinton was bitten by a dog which was
assumed to be mad, though it may only have been indiscriminating. (Saki)
XEmacs-Patches mailing list