[Ben: live-icon doesn't work. Can you take a quick look and see if there's anything obvious? Recipe below.] Thanks for the followups. >>>>> "Dan" == Dan Espen <dane(a)mk.telcordia.com&gt; writes: Dan> live-icon doesn't work with fvwm2, twm, and dtwm. Thats all Dan> the wm's I have. It used to work with all of them. Thank you. Dan> I also tried this: [...] And got this: (omitted) Dan> I noticed that the ":file" stuff is still in the glyph and Dan> "xemacs-icon.xpm" is the icon I see. Yep. Dan> The data from live-icon doesn't seem to be there... Well, you _can_ see the color specs (chartreuse is _not_ in the default icon!!) and the :data is in the ellipsis. In *scratch*, I did ;;; Ben, this demonstrates the bug (specifier-instance (glyph-image frame-icon-glyph)) #<image-instance (color-pixmap) on #<x-device on ":0.0" 0x60c> "xemacs-icon.xpm" 48x48x24 (0x800031) 0x1b0b> (require 'live-icon) (live-icon-one-frame) (specifier-instance (glyph-image frame-icon-glyph)) #<image-instance (color-pixmap) on #<x-device on ":0.0" 0x60c> "xemacs-icon.xpm" 48x48x24 (0x800031) 0x1b0b> (specifier-specs (glyph-image frame-icon-glyph) 'all) I've removed all the output except from "specifier-instance" which confirms your report at the Lisp level. If you do the same, you'll see the data. What appears to be happening is that having multiple options that specify data in the same spec doesn't give an error, but instead picks one arbitrarily. So what live-icon should be doing is clearing out the spec somehow, then adding the new spec. Why this is not happening, I don't know yet. The default behavior for specifiers is "remove existing specs for the tagset and prepend to the list", which seems sensible. I'm copying xemacs-beta(a)xemacs.org and Ben, where something might get done about this. If you're not on xemacs-beta and don't need to be otherwise, you can follow the thread with at most 24 hours lag by going to http://list-archive.xemacs.org/xemacs-beta and searching for "live-icon". -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.

>>>>> "Ben" == Ben Wing <ben(a)666.com&gt; writes: Ben> i don't see extra specs. What I'm referring to is the fact that my understanding was that (1) we clear out all the specs that match the tagset, then (2) we prepend the new spec.

should not be picked up and used in preference to the :data spec which the user (via live-icon) has specified. But that seems to be what's happening.

Maybe I'm just confused. Specifiers can be pretty confusing. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.

Does anyone have a test case for the bug?

andy > -----Original Message----- > From: xemacs-beta-admin(a)xemacs.org > [mailto:xemacs-beta-admin＠xemacs.org]On Behalf Of Ben Wing > Sent: Sunday, November 11, 2001 4:10 PM > To: Stephen J. Turnbull > Cc: xemacs-beta(a)xemacs.org; Andy Piper > Subject: Re: live-icon and XEmacs 21.4.5 > > > > > "Stephen J. Turnbull" wrote: > > > > >>>>> "Ben" == Ben Wing <ben(a)666.com&gt; writes: > > > > Ben> i don't see extra specs. > > > > What I'm referring to is the fact that my understanding was that (1) > > we clear out all the specs that match the tagset, then (2) we prepend > > the new spec. > > there's some confusion here. your sentences above describe changing the > specifiers ... > > That would imply that the :file spec from the fallback > > should not be picked up and used in preference to the :data spec which > > the user (via live-icon) has specified. But that seems to be what's > > happening. > > ... but these describe the process of instantiation. these are > different. > "clear out all the specs that match the tagset" occurs only in > the same locale; > if i add a global spec that's restricted to `mswindows', it makes > no sense to > clear out buffer-local specs also restricted to `mswindows'. > > furthermore, fallback specs are never changed -- that's the > nature of fallback. > > but there does appear to be a bug in the instantiation process, > as i said. this > may be related to andy's widget changes. in particular, andy > added a bunch of > caches of various sorts, and it's quite possible that a cache > isn't getting > properly cleared. > > andy, could you take a look? > > > > > > Maybe I'm just confused. Specifiers can be pretty confusing. > > > > -- > > Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp > University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN > Don't ask how you can "do" free software business; > ask what your business can "do for" free software.