On Tue, Apr 20, 2004 at 06:27:17AM +0200, Fabio Massimo Di Nitto wrote: > I have only a few changes that i am not fully convinced about and that i > would like to discuss with the team. From debian/changelog: > > * If defoma-app and fontconfig's defoma script are available, invoke > "defoma-app update fontconfig" on package configuration and removal; > otherwise, use fc-cache as before. (Closes: #238797) > > This sounds a bit of a workaround since the problem might be located > somewhere else. Did i miss something from the bug report or do we want to > prevent the problem in advance? Well, no, the idea is simply that we take advantage of defoma if it is installed, since we can trust it to run fc-cache. If it's not, we run fc-cache ourselves. This is not intended as a workaround. If there's anything else defoma wants done with the fonts, it can do it. > * Implement IsolateDevice and SingleCard layout options for XF86Config, and > -isolateDevice command-line flag for XFree86 X server. Useful for > multi-card setups where different X servers run on each card (thanks, > Andreas Schuldei). (Closes: #207543) > > This looks a bit dangerous to my eyes. I am not 100% happy to include it > but it for sure needs a more global roll out. I would say it goes in for > -8, but it will fly away at the first problem related to it and if Andreas > cannot fix it in a reasonable amount of time. (it would be the only > dangerous bit in this release) It looksed dangerous to me, but I'm not quite as nervous about it now that I've read the code. I agree that it should be the first thing to go if this release has weird problems in the X server/PCI department. > * Update xdm.options(5) and xfs.options(5) manual pages to instruct the user > to use the invoke-rc.d command instead of invoking packages' init scripts > directly. > > Is there any specific motivation for this change? Just curious, I don't > disagree. People shouldn't run "/etc/init.d/$SCRIPT" anymore; they should use "invoke-rc.d $SCRIPT" instead. There have been many threads on -devel about this. > * Fix problem with DVI displays connected to ATI Radeon 9200 cards not > initializing properly every other time the X server is started (patch by > Hui Yu of ATI, from XFree86 Bugzilla #673). (Closes: #243575) > > Ok. Are we licence safe? I didn't see any note about it. Yes, we should be. 1) The patch comes from Hui Yu of ATI, not XFree86 CVS. 2) It was submitted to XFree86's Bugzilla on 2003-09-11, before the relicensing of XFree86 (2004-02-13), and even before the X-Oz license first reared its head (2003-10-08). 3) We could have taken the patch from XFree86 CVS anyway, as it was committed there on 2003-10-07, before the relicensing and the X-Oz stuff. Some of the earlier stuff I committed does need to be re-reviewed to ensure that it is licensing-clean, and documented thus. Recently word has gone around that all CVS commits by David Dawes should be regarded by him as being under the XFree86 1.1 license, even those that didn't contain a message to that effect in the commit log (i.e., contrast with [1]). I don't know how far back that regard should go, and given the extreme difficulty I've had in getting straight answers to simple questions from him, I'm not going to bother asking, and assume it goes back to 2004-02-13, when the codebase was relicensed. How exactly such an assertion of copyright is supposed to work over changes like this[2] or [3], I'm not sure, but it's probably better to just avoid XFree86 CVS commits after 2004-02-13 on general principle. Most changes that would be of interest us were submitted by external people to their Bugzilla, and/or are authored by people who don't mind using the traditional license. If you have any more questions or concerns about the pending changes, please let me know. Otherwise, if I have addressed your questions satisfactorily, give me the green light and I'll do the merge onto branches/4.3.0/sid. > >From the TODO list, the following fixes should be included/done for -8 in > order of priority: > > * #242865: remove (unused) sourceless binary exectables from source tree > > * #239991: Fix SEGV in xc/lib/X11/imDefIc.c:_XimRegisterKeyPressFilter(). > > * #147898: xlibs: select() hangs in XDrawString16() when using TrueType with > > * #222804: xlibs: [xkb] proposed symbol mapping for Compaq Evo > > * #241034: xlibs: Meta not working and AltGr cancels Compose with > > and all of them have patches/fixes ready but included in trunk. *Not* included, right. Okay. I will work on those items in that order of priority. [1] http://marc.theaimsgroup.com/?l=xfree-cvs&m=107959393514924&w=2 [2] http://marc.theaimsgroup.com/?l=xfree-cvs&m=107923290610864&w=2 [3] http://marc.theaimsgroup.com/?l=xfree-cvs&m=107956770516652&w=2 -- G. Branden Robinson | If God had intended for man to go Debian GNU/Linux | about naked, we would have been branden@debian.org | born that way. http://people.debian.org/~branden/ |
Attachment:
signature.asc
Description: Digital signature