On Thu, May 10, 2012 at 6:37 PM, Eli Zaretskii <e...@gnu.org> wrote: >> From: Lennart Borgman <lennart.borg...@gmail.com> >> Date: Thu, 10 May 2012 11:50:58 +0200 >> Cc: help-emacs-windows@gnu.org >> >> I just realized that I have only told about one of the problems I have >> seen. There are two: >> >> - The menus does not work as they should. This problem can only happen >> the way I see it in my patched Emacs, but I am not sure it is really >> related to any of my patches. What I see is that the "&"-signs that I >> use to get underlined chars in the menus now shows up as just "&" >> chars instead. Fixing the underline is done by the w32 API so they get >> wrong information somehow here. > > Do those menus include non-ASCII characters?
No, it is all ASCII chars. And I see the problem only now and then. In those cases Emacs shows the menus perfectly well from the beginning, but then later that strange behavior above appears. > There's a bug in all versions of Emacs older than the current trunk, > whereby whenever some call to AppendMenuW returns an error indication, > Emacs assumes that it is running on Windows 9X, and switches to > calling AppendMenuA instead. Menu entries that include non-ASCII > characters not from the current locale will display as blanks or > question marks or just gibberish after that switch. Maybe what you > see is some manifestation of that bug. Thanks. I guess there could be a problem in the call of the w32 menu API, but I have not been able to locate it. (BTW, this was one of the reasons why I wanted all system API-calls to be checked for return values.)
