| Summary: | xcursor-themes missing symlinks (firefox/kwin issues) | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | satmd <satmd> |
| Component: | New packages | Assignee: | Gentoo X packagers <x11> |
| Status: | RESOLVED NEEDINFO | ||
| Severity: | major | Keywords: | Inclusion |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | patched ebuild | ||
|
Description
satmd
2006-06-14 05:54:21 UTC
Created attachment 89145 [details]
patched ebuild
While I don't believe this is the best coding style (incoporated an existing shell script), it works for me.
I wonder if this is something that should be inherited from an eclass.
Yeah, that style is really wacky, but it gives the idea of what needs to be done. So I talked to keithp about this. Those hashes are generated using XcursorImageHash(), which is an exported function in libXcursor. We could write a small app that just takes cursor names, calls XcursorImageHash() and spits out the hash. That would be the more correct way to do this. Donnie, do you remember what's going on here? Thanks This bug was somewhat resolved, but I was waiting for another solution after comment #2 ... and eventually forgot about the bug. I'm still using this patch. I tried the approach Donnie Berkholz mentioned (using XcursorImageHash()), but I ran in to a problem. XCreateImage() expects an XImage*, not an XcursorImage*, which is what for example XcursorFileLoadAllImages() returns. Manually creating an XImage from the pixeldata of an XcursorImage leads to different hashes. This is likely because there is something more going on, because a single cursor can have many XImages. For example sb_v_double_arrow contains 5 XcursorImages. |