Date: Sat, 07 Feb 2009 18:00:59 -0500 From: Alex Goncharov <alex-goncharov@comcast.net> To: freebsd-ports@freebsd.org Subject: libX11-XCB [Was: X drivers depend on Gnome?] Message-ID: <E1LVwAV-000NVx-83@daland.home> In-Reply-To: <E1LUja7-0000w7-SU@daland.home> (message from Alex Goncharov on Wed, 04 Feb 2009 10:22:27 -0500) References: <E1LUQYl-000Cir-UT@daland.home> <op.uostqymy9aq2h7@localhost> <E1LUVek-000EoV-0p@daland.home> <20090204083740.12061iw6kf4c9vwg@webmail.leidinger.net> <E1LUhXO-0000cF-PB@daland.home> <a01628140902040521y6e396dfbhd80a53cd0455a021@mail.gmail.com> <E1LUiDw-0000iN-7q@daland.home> <a01628140902040600y5dd209cfo83dab015015ff21e@mail.gmail.com> <E1LUja7-0000w7-SU@daland.home>
next in thread | previous in thread | raw e-mail | index | archive | help
,--- I/Alex (Wed, 04 Feb 2009 10:22:27 -0500) ----* | ,--- I/Alex (Wed, 04 Feb 2009 08:55:28 -0500) ----* | | ,--- You/Florent (Wed, 4 Feb 2009 13:21:15 +0000) ----* | | | Please send patches. Kthxbye. | | | | Does this mean that you agree with me in principle and if I send you | | patches, there is a high probability they will he accepted (provided | | they work, of course)? | ,--- You/Florent (Wed, 4 Feb 2009 14:00:25 +0000) ----* | | In principle yes. | | Good, thank you. | | | Now you do have to understand that this is a lot of boring work and | | you'll need approval from portmgr@, not just me. | | I understand both and it remains to be seen if I handle the task (not | in the next three weeks, at least, I think). But I don't want to go | for a lot of boring work without knowing that the change would be | welcome in principle, at least by some of the involved people. I spent some time playing with this. Yes, that's a lot of work -- and a potential build destabilization. This said, I created a bsd.xorg.mk without GNOME dependencies, and modified a few lower level X components make files accordingly. The work is not done yet but I am able to build some protocols and libraries now and am usefully improving my understanding of both the ports build system and the modern X structure. I don't think I am going to submit my patches, even when I am done, because I see the far-reaching scope of the changes and an extra destabilization of current X in ports would not be a good idea. But I am thinking of building an X build structure for my own consumption. Essentially, I want a minimalist X build/install -- I run not that many X applications. E.g. I haven't seen any benefit of using HAL yet -- so I am happily running HAL-less and intend to continue to do so. Now another new (about two-years-old, I think) addition to X caught my attention -- XCB. I've done some reading and didn't see any argument for using it by a non-X-developer (while the architectural argument sounds good, no doubt.) This is good recent post expressing one view on XCB usefulness -- http://mail-index.netbsd.org/pkgsrc-users/2008/09/07/msg007999.html: Re: Why disabling XCB in x11/lix11 ? ... So far nothing really needed it which made attacking the issues a less interesting way to proceed. And I also see plenty of XCB-enable/disable clauses in ports. E.g. graphics/cairo/Makefile:.ifdef(WITH_XCB) graphics/libGL/bsd.mesalib.mk:.if defined(WITHOUT_XCB) [ Three (grumbling) side observations here: 1. Why does bsd.mesalib.mk live in graphics/libGL, rather than in the standard Mk? 2. Why are there two different switches for the same purpose: WITH_XCB and WITHOUT_XCB? 3. Why neither of the switches is documented (as far as the search in Mk shows)? ] So, I built my libX11 without XCB and will try to build other components without it. My question to X experts is: will I lose anything without using XCB? (Again, not caring about X application development.) Thanks, -- Alex -- alex-goncharov@comcast.net --
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E1LVwAV-000NVx-83>