November 2013 - Xquartz-dev - lists.macosforge.org

focus-related regression with XQuartz 2.7.3+
by Marc Bejarano 10 May '14

10 May '14

[this is a resend because i don't think my first attempt will go through as it was sent with a non-subscribed email address] hello all, i opened http://xquartz.macosforge.org/trac/ticket/636 back in september and jeremyhu suggested i write here to try and get the issue addressed. from the ticket: -- ever since upgrading from 2.7.2 to 2.7.3, i now see X11.app stop bringing it's windows to the foreground when switching back to it. this doesn't happen immediately. if i restart X11.app, things work for a while, but then eventually break. once broken, the bad behavior is reproducible every time. i only run one X application (Visual SlickEdit). using the <CMD>-2 keyboard shortcut will bring the window back to the foreground. sometimes, like right now, that won't even work and i'll need to minimize the window and i'll need to use Window -> Bring All to Front to get it to come back up. -- and: -- another data point: exiting and restarting the X application is another way to work around the problem (presumably temporarily) -- help? tia, marc

2 3

XQuartz works only with sudo
by Christof Wolf 25 Nov '13

25 Nov '13

Hi Jeremy, I am running XQuartz 2.7.5 under Maverick, but I can't get it work properly. (I am on a Retina 11,3 Macbook Pro) First of all, it takes a long time to start, xauth always gives me: xauth: timeout in locking authority file /Users/Wolf/.Xauthority it does not matter, if I remove .Xauthority or not, then it does give me: Invalid MIT-MAGIC-COOKIE-1 key waiting for X sever to begin accepting connections MIT-MAGIC-COOKIE-1 key.. MIT-MAGIC-COOKIE-1 key.. MIT-MAGIC-COOKIE-1 key.. an nothing ... the only way to get it working is if I start it with sudo in a terminal, I get: NeoIX:~ Wolf$ sudo startx font_cache: Scanning user font directories to generate X11 font caches font_cache: Updating FC cache font_cache: Done xauth: file /Users/Wolf/.serverauth.3832 does not exist /opt/X11/bin/xinit: XFree86_VT property unexpectedly has 0 items instead of 1 and kde starts working properly. not sure, what's wrong. Something wrong with the rights of certain files? best christof -- Munich School of Philosophy Kaulbachstr. 31a 80539 Munich office: +49.89.2386.2418 cell: +49.163.637.6450 http://www.hfph.mwn.de/lehrkoerper/lehrende/c_wolf For secure communication, extract the public key at http://chwolf.userweb.mwn.de/ch.wolf.asc Please respect my privacy and do not make my contact information available to third parties. This email is UNCLASSIFIED.

1 0

XQuartz 2.7.5
by Jeremy Huddleston Sequoia 15 Nov '13

15 Nov '13

XQuartz 2.7.5 is available for install. The main difference between it and rc4 is that I fixed a possible hang when quitting XQuartz and reverted a commit to fontconfig which caused the font cache to be rebuilt every time. http://xquartz.macosforge.org/trac/wiki/X112.7.5 --Jeremy

8 22

Re: [Xquartz-dev] "xquartz" load average reached 90%
by Jeremy Huddleston Sequoia 13 Nov '13

13 Nov '13

It should be just: /downloads/sparkle/release.xml /trac/wiki/X112.7.5 /downloads/SL/XQuartz-2.7.5.dmg Sent from my iPhone... > On Nov 13, 2013, at 13:35, Shreeraj Karulkar <skarulkar(a)apple.com> wrote: > > Jeremy > Let me figure out some thing, give me some time. BTW, are there any other directories which maybe hit this way? I can move them off this server as well? Let me know. > > Shree > >> On Nov 13, 2013, at 1:30 PM, Jeremy Huddleston Sequoia <jeremyhu(a)apple.com> wrote: >> >> If users are running old versions of XQuartz, they will get an update notification upon launch of the old version. This will trigger a GET for /downloads/sparkle/release.xml and then /trac/wiki/X112.7.5 (and related assets). >> >> 2.7.5 was just released this weekend, so I would expect increased traffic for a while until most users have upgraded. >> >> --Jeremy >> >>> On Nov 13, 2013, at 13:25, Shreeraj Karulkar <skarulkar(a)apple.com> wrote: >>> >>> Jeremy >>> >>> At about 1 PM the xquartz.macosforge’s load average reached 90% and we started getting the same problems on other projects. I have shutdown the server for now. Do you know anything that might have triggered this? I see a lot of GET requests for /trac/wiki/.. >>> >>> Shree >

1 0

Re: [Xquartz-dev] "xquartz" load average reached 90%
by Jeremy Huddleston Sequoia 13 Nov '13

13 Nov '13

If users are running old versions of XQuartz, they will get an update notification upon launch of the old version. This will trigger a GET for /downloads/sparkle/release.xml and then /trac/wiki/X112.7.5 (and related assets). 2.7.5 was just released this weekend, so I would expect increased traffic for a while until most users have upgraded. --Jeremy On Nov 13, 2013, at 13:25, Shreeraj Karulkar <skarulkar(a)apple.com> wrote: > Jeremy > > At about 1 PM the xquartz.macosforge’s load average reached 90% and we started getting the same problems on other projects. I have shutdown the server for now. Do you know anything that might have triggered this? I see a lot of GET requests for /trac/wiki/.. > > Shree >

1 0

I am having problems with 2.7.5_rc3 and _rc4 (rc2 & before were ok)
by SciFi 12 Nov '13

12 Nov '13

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, (yeah I'm still here, still running 10.6.8 with all security updates etc. <g>) The latest two rc's for 2.7.5 (rc3 and rc4, so far) seem to have problems here. (rc2 & before did not exhibit any of this) I use the CLI installer (/usr/sbin/installer) under a 'login root' terminal. Let me snip its log to show exactly where I think the problems stem from: > # installer -dumplog -target / -pkg /Volumes/XQuartz-2.7.5_rc4/XQuartz.pkg > […] > installer[25130]: PackageKit: Executing script "./postinstall" in /private/tmp/PKInstallSandbox.ApMrWM/Scripts/org.macosforge.xquartz.pkg.oK4S7s > installer[25130]: ./postinstall: lockfile: Forcing lock on "/opt/X11/var/run/font_cache.lock" > installer[25130]: ./postinstall: font_cache: Scanning system font directories to generate X11 font caches > installer[25130]: ./postinstall: font_cache: Making fonts.dir for updated directories. > […] > installer[25130]: ./postinstall: font_cache: Updating FC cache > installer[25130]: ./postinstall: /opt/X11/share/fonts: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/100dpi: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/75dpi: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/OTF: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/Speedo: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/TTF: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/Type1: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/cyrillic: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/encodings: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/encodings/large: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/misc: failed to write cache > installer[25130]: ./postinstall: /opt/X11/share/fonts/util: failed to write cache > [… and hundreds more such lines, one for each subdir under the various font paths …] > [… I've collected tons of fonts over the (long) years since 10.3-Tiger <g> …] > installer[25130]: ./postinstall: /opt/X11/share/fonts: failed to write cache > installer[25130]: ./postinstall: /usr/X11R6/lib/X11/fonts: failed to write cache > installer[25130]: ./postinstall: /System/Library/Fonts: failed to write cache > installer[25130]: ./postinstall: /Library/Fonts: failed to write cache > installer[25130]: ./postinstall: font_cache: Done > installer[25130]: PackageKit: Registered bundle file://localhost/Applications/Utilities/XQuartz.app/ > installer[25130]: Installed "XQuartz 2.7.5_rc4" () > installer[25130]: PackageKit: ----- End install ----- > installer[25130]: Running install actions > installer[25130]: Removing temporary directory "/var/folders/zz/zzzivhrRnAmviuee+++++++++++/-Tmp-//Install.251301fxWVS" > installer[25130]: Finalize disk "WhiteCat" > installer[25130]: Notifying system of updated components > installer[25130]: > installer[25130]: **** Summary Information **** > installer[25130]: Operation Elapsed time > installer[25130]: ----------------------------- > installer[25130]: disk 0.02 seconds > installer[25130]: script 0.02 seconds > installer[25130]: zero 0.02 seconds > installer[25130]: install 426.42 seconds > installer[25130]: -total- 426.48 seconds > installer[25130]: > installer: The upgrade was successful. > > # _ I'll launch XQuartz the normal way (use Finder to double-click the app icon under /Applications/Utilities). It will start up, but with an unusually long period running the various font tools (again) before it's ready for use. I will start various X11 apps in the terminal. I notice that the cli app takes another unusually long time before it shows on the screen. It is as if the system needs to "relink" the dylibs and whatnot -- but now it does this every time I start an X11 app, no matter what. Since these are mostly apps based on Glib/GTK+ (and requisites etc.), this takes a _really_ long time. Once the app is 'finished' in this realm, it seems the app will run as expected. (I have seen some awesome changes in the look of the fonts on-screen, better than the earlier XQuartz builds to be honest.) As a specific example, I start Pan up to three times -- one for each NNTP server, in different configs -- simultaneously. Each task is now taking this same long time before it is ready, every time I start one up. With 2.7.5_rc2 and before, all is fine, with the exact same setup here. I know the "relink" phase might be needed when the app is first started after such an update to XQuartz, but thereafter the app does not need another "relink" and starts up 'faster' -- even with the three Pan tasks as just mentioned. FWIW, thank you for keeping 10.6 alive in this project for us who cannot afford new hardware for upgrading the o.s. itself (I have an 'only' model iMac6,1 here [32-bit EFI/BIOS/kernel but its Core-2-Duo can run 64-bit user space], still disabled / retired, barely enough "monthly allotment" for the rent/food/etc. and some [cheap] hobbies). -----BEGIN PGP SIGNATURE----- iQEcBAEBAgAGBQJSd4HuAAoJEKkeWNKet7+K6xMIAJPTxpJiefYB+ppISf82JqmV X128gYztOXYqWoy3gXQ0BnROJBM+/anLRo/lmzQaQsQiczcqYuVMtX3d3Bgm1UCP fTDi6vnSfD0m/BT5mgoz0P+51+LPxRFdSKnmlYGvwXyo6u6cKL7ptYA9mEnW2yID +JG6TItSSsUI1HfRxVmUuJR1ZLvUcWAf1sXrkLlg5Tm0xxh/2XnjHZGI5kwhoLZ+ F7Xz3vaxneCbPxIESlmhgPRF7RHoGYBq1v2He7XPRj8qNLQBipQYIsytE3ErNaOj 0hhFeSGNxYP714Hwa8x3abmo3oMY6FthLB1nrSUs5mICSKN6puP3qsY14GKxFdU= =yLHq -----END PGP SIGNATURE-----

6 22

Re: [Xquartz-dev] I am having problems with 2.7.5_rc3 and _rc4 (rc2 & before were ok)
by David Ryskalczyk 12 Nov '13

12 Nov '13

> Because in the shells of X clients and non-X11 applications DISPLAY is set to different values. When I do echo $DISPLAY in non-X11 Terminal.app and in the X11 xterm, I get the same value, which is the path to the xquartz socket: % echo $DISPLAY /tmp/launch-PScxyX/org.macosforge.xquartz:0 --Dave

2 1

Re: [Xquartz-dev] Xquartz
by Jeremy Huddleston Sequoia 12 Nov '13

12 Nov '13

On Nov 12, 2013, at 10:24, Shreeraj Karulkar <skarulkar(a)apple.com> wrote: > Hi > We’ve had to shutdown apache on the Xquartz server. We will let you know when it is back up. > Shree I just put out a release yesterday. If you shut it down due to "oddly high traffic", please just bring it back up as it is expected due to automatic updates. --Jeremy

1 0

Re: [Xquartz-dev] Another XQuartz 2.7.5 rc4 issue
by Tom Lane 11 Nov '13

11 Nov '13

Brandon Allbery <allbery.b(a)gmail.com> writes: > On Sun, Nov 10, 2013 at 11:12 AM, Tom Lane <tgl(a)sss.pgh.pa.us> wrote: >> I've noticed that when using exmh (a Tcl/Tk mail reader), each time >> I summon and dismiss a window such as the search box, the window >> reappears a bit lower on the screen --- it looks like the displacement >> is exactly the height of the title bar. I can't swear how long this > For what it's worth, I'm not seeing this with xchat on 10.8.5. But xchat > remembers its window location and may be overriding it. Well, exmh remembers its window location(s), too, which is why this is surprising. It's entirely possible that this is a Tk bug at bottom, but *something* has changed about the behavior of the window manager. regards, tom lane

2 1

XQuartz 2.7.5 rc4
by Jeremy Huddleston Sequoia 10 Nov '13

10 Nov '13

Hi all, I just uploaded XQuartz 2.7.5 rc4. The only changes from 2.7.5 rc3 are the update to xorg-server 1.14.4 (from 1.14.4 rc1) and a fix for a crash regression (http://xquartz.macosforge.org/trac/ticket/525) If no further release-blocking issues are found in the next week or so, this should become the 2.7.5. http://xquartz.macosforge.org/downloads/SL/XQuartz-2.7.5_rc3.dmg Thanks, Jeremy

2 2