Subject: RE: dogfood feedback -- Smart Quotes
From: Paul Rohr (paul@abisource.com)
Date: Thu Feb 22 2001 - 16:07:03 CST
At 01:23 PM 2/22/01 -0800, WJCarpenter wrote:
>I basically agree about the UNDO granularity. The problem is about
>insertion point motion and undo operations. There is a fuller
>explanation of the problems I ran into here:
>
> http://www.abisource.com/mailinglists/abiword-dev/00/August/0402.html
>
>I agree that this is not good, but without the equivalent of
>(save-excursion ...), I was sort of stuck. I'd be happy if someone
>figured this out (either by fixing the UNDO system in AbiWord or by
>figuring out how I had wrongly used it).
Sorry I missed that the first go-around. Could you give more details on
your undo problem -- focusing on approaches which didn't work?
We should be able to come up with ways to do what you want here, but I'm not
sure how much work it'll be to do the Right Thing. Essentially, you want a
user-atomic change from the old-quoted state to the new-quoted state, with
no change to the IP. Right?
The cheap & obvious hack would be to replace the entire quoted passage,
instead of just the quote character(s), but I haven't thought through the
side effects this might cause. (For example, would it mess up the existing
coalescing algorithms?)
>Sorry, but I can't reproduce this (I just tried it on the unnumbered
>win32 0.7.13 binary). "aren't" gets no red squiggle, "arren't" gets a
>red squiggle. I went to a certain amount of trouble to preserve the
>apostrophe case (and even fixed a few existing glitches), so I know
>the goal state well.
As I mentioned in the note to Vlad, this might just be a side effect of the
imp/exp glyph substitution stuff below.
I was dogfooding an important document, so I was saving a *lot*. ;-)
>I *can* repro this on win32, with single or double quote marks, but
>I'm pretty sure it's a im/ex problem and not a smart quote problem.
>(The problem still occurs if you turn smart quotes off.) On win32 to
>repro the problem, just save either of Paul's or Vlad's examples into
>"sq.abw", open it, see the glyphs correctly, save as "sq1.abw", open
>that, and see the glyphs incorrectly. (You can also see that they're
>different things in the *.abw files.)
OK. Could you work with Vlad to isolate what's happening here?
Thanks,
Paul
This archive was generated by hypermail 2b25 : Thu Feb 22 2001 - 15:59:32 CST