On Tue, Jan 08, 2008 at 11:06:07PM +0100, Agustin Martin wrote: > This all means, I will need to carefully study the diff between them. > Unfortunately, unless I am really lucky, I do not expect this to be quick.
I found some time to look into this problem. Noticed that in the buffers showing the problem, flyspell-mode was indeed disabled (was nil) and debugging what was leading to flyspell-word calls in that buffers, things pointed to (flyspell-post-command-hook) being called in that buffers even if flyspell-mode is nil, once it is enabled in other buffer. I have included a patch which just checks that flyspell-mode is enabled before actually calling flyspell-post-command-hook. There is however something left, with just the above, word in the buffer with flyspell-mode disabled is not checked when the buffer is visited, but surprisingly when after that mouse returns to the buffer with flyspell-mode enabled. Making `flyspell-pre-point' buffer-local seems to address this. I have built a package including that patch and put it in our alioth repository at http://dict-common.alioth.debian.org/testing It is still minimally tested, but has not yet shown drawbacks, so you may want to give it a try (changes file is signed, if you want to gpg check it). Above is mostly a workaround, and I am unhappy that I could not properly identify what is making emacs and xemacs behave differently. However what the patch does seems to me sane. Will upload in a few days if no problems appear. -- Agustin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
