I have attached patches for ebuilds that still require xfree without a condition for using xorg instead. Reproducible: Always Steps to Reproduce: 1. 2. 3.
Created attachment 29161 [details, diff] Patch for blueglass-xcursors-0.4.ebuild
Created attachment 29162 [details, diff] Patch for golden-xcursors-0.8.ebuild
Please use the virtual instead if possible
Created attachment 29163 [details, diff] Patch for mixer_app-1.8.0.ebuild
Created attachment 29164 [details, diff] Patch for nvtv-0.4.6.ebuild
Created attachment 29165 [details, diff] Patch for nxclient-1.3.0.ebuild
Created attachment 29166 [details, diff] Patch for nxserver.eclass
Created attachment 29167 [details, diff] Patch for silver-xcursors-0.4.ebuild
Created attachment 29168 [details, diff] Patch for wmclockmon-0.7.2.ebuild
Created attachment 29169 [details, diff] Patch for yanc-0.2.1.ebuild
you've tested all of these with xorg, shawn?
No, actually, but I have them installed. As far as I can tell Xorg's directory structure is the same as xfree. Further, since xorg is in fact a recent branch of xfree which would have satisfied the DEPs in the ebuilds, so should Xorg methinks.
Sorry guys, I haven't explored the virtual angle. I just wanted to fix some breakage that was trying to pull in the unwanted xfree.
how about using virtual/x11 ?
Well, for some of them it's not appropriate due to version specificity, or at least as implied by the dependancy in the ebuild. For the others, yeah, it'd be great. I have VERY little time left this week and might not get to making new patches, but if you'd like to in my stead, feel unequivocally welcome to do so yourself. I've seen others complain about "xfree mysteriously being brought back in by some dependancy". I just happened to pay attention to the debug...
These need to be compiled against xorg-x11 so they're proven to compile, because it's significantly different from xfree 4.3.0, and many things were never tested against xfree 4.3.99.x. Then the dependency needs to be changed to virtual/x11. Just to sum things up.
Here are some test results after replacing the dependencies for x11-base/xfree with virtual/x11: *-xcursors: These ebuilds install data into /usr/share/cursors/xfree/ which is not recognized by default by xorg-x11. To fix this I just used has_version to look for xorg vs. xfree, example: cursorsdir="xfree" `has_version xorg-x11` && cursorsdir="xorg-x11" mkdir -p ${D}/usr/share/cursors/${cursorsdir}/Silver/cursors/ wmclockmon & mixer_app - compiled and ran without error Yanc: compiles without error, but has XF86Config{,-4} hardcoded in src/Unit1.pas. I don't use QT so I couldn't test the gui.
*** Bug 47740 has been marked as a duplicate of this bug. ***
*** Bug 49643 has been marked as a duplicate of this bug. ***
*** Bug 49465 has been marked as a duplicate of this bug. ***
*** Bug 50367 has been marked as a duplicate of this bug. ***
Created attachment 30956 [details] blueglass-xcursors-0.4-r1.ebuild These ebuilds are completely fixed for use of xorg and xfree, and also fixes the dependency.
Created attachment 30957 [details] golden-xcursors-0.8-r1.ebuild
Created attachment 30958 [details] silver-xcursors-0.4-r1.ebuild
Actually, I'd prefer to use the method to find which is installed that's in ati-drivers-3.7.6-r1. Take a look -- grep it for virtual/x11.
Created attachment 30981 [details] blueglass-xcursors-0.4-r1.ebuild Thanks, these ebuilds use the same system.
Created attachment 30982 [details] golden-xcursors-0.8-r1.ebuild
Created attachment 30983 [details] silver-xcursors-0.4-r1.ebuild
*** Bug 53356 has been marked as a duplicate of this bug. ***
*-xcursors fixed.
*** Bug 53447 has been marked as a duplicate of this bug. ***
Fixed nxserver, nxclient, mixer_app
yanc patched to look for xorg.conf
Fixed all these.
