Re: [blfs-dev] lxdm killing processes on logout
On Oct 10, 2020 4:32 PM, "Pierre Labastie via blfs-dev" < blfs-dev@lists.linuxfromscratch.org> wrote: > > FWIIW, neither lightdm, nor gdm kill the screen process... Thanks for confirming the issue. I've switched to lightdm. However I'm now seeing an issue with pcmanfm when starting LXDE with lightdm and gdm. I'll post a new thread once I finish testing. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm killing processes on logout
On Wed, 2020-10-07 at 23:33 +0200, Pierre Labastie via blfs-dev wrote: > On Wed, 2020-10-07 at 22:48 +0200, Pierre Labastie via blfs-dev > wrote: > > On Sun, 2020-09-20 at 17:22 -0400, Joe Locash via blfs-dev wrote: > > > I've been battling this problem and would appreciate a sanity > > > check. > > > In my testing of lxde any screen session I start and then detach > > > from > > > gets killed when I logout. If I bypass lxdm by changing run level > > > to > > > 3 and using startx to start lxde and then logout the screen > > > session > > > survives. > > > > > > This also happens in a systemd build. > > > > > > It's not a config issue. I build both elogind and systemd to not > > > kill > > > user processes on logout. > > > > > > Thoughts? > > > > Been pinged privately, but answering publicly. Sorry for not doing > > that > > earlier. I usually use gdm, and have never tried to keep any > > personal > > process open when loging out. > > But now, as you may have seen from the ticket, I've built lxdm > > (under > > gtk+-3), so I can try to reproduce. > > If I understand correctly, you start a screen session under lxde > > (started with lxdm), detach from that session (I'm not sure how to > > do > > that), then logout, which means you get back to the lxdm loging > > screen. > > But all your processes are killed then (including screen). You test > > that through a ssh connection or so? > > > > the other way, you start lxde from a linux console, you sart screen > > and > > detach, then you exit lxde, then you logout from the linux console. > > Or > > are you doing something different? > > > > I'll try and get back when I can confirm that behavior. > > > > I seem to observe the same behavior. Will look at it in more details > tomorrow. > FWIIW, neither lightdm, nor gdm kill the screen process... Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm killing processes on logout
On Wed, 2020-10-07 at 22:48 +0200, Pierre Labastie via blfs-dev wrote: > On Sun, 2020-09-20 at 17:22 -0400, Joe Locash via blfs-dev wrote: > > I've been battling this problem and would appreciate a sanity > > check. > > In my testing of lxde any screen session I start and then detach > > from > > gets killed when I logout. If I bypass lxdm by changing run level > > to > > 3 and using startx to start lxde and then logout the screen session > > survives. > > > > This also happens in a systemd build. > > > > It's not a config issue. I build both elogind and systemd to not > > kill > > user processes on logout. > > > > Thoughts? > > Been pinged privately, but answering publicly. Sorry for not doing > that > earlier. I usually use gdm, and have never tried to keep any personal > process open when loging out. > But now, as you may have seen from the ticket, I've built lxdm (under > gtk+-3), so I can try to reproduce. > If I understand correctly, you start a screen session under lxde > (started with lxdm), detach from that session (I'm not sure how to do > that), then logout, which means you get back to the lxdm loging > screen. > But all your processes are killed then (including screen). You test > that through a ssh connection or so? > > the other way, you start lxde from a linux console, you sart screen > and > detach, then you exit lxde, then you logout from the linux console. > Or > are you doing something different? > > I'll try and get back when I can confirm that behavior. > I seem to observe the same behavior. Will look at it in more details tomorrow. Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm killing processes on logout
On Sun, 2020-09-20 at 17:22 -0400, Joe Locash via blfs-dev wrote: > I've been battling this problem and would appreciate a sanity check. > In my testing of lxde any screen session I start and then detach from > gets killed when I logout. If I bypass lxdm by changing run level to > 3 and using startx to start lxde and then logout the screen session > survives. > > This also happens in a systemd build. > > It's not a config issue. I build both elogind and systemd to not kill > user processes on logout. > > Thoughts? Been pinged privately, but answering publicly. Sorry for not doing that earlier. I usually use gdm, and have never tried to keep any personal process open when loging out. But now, as you may have seen from the ticket, I've built lxdm (under gtk+-3), so I can try to reproduce. If I understand correctly, you start a screen session under lxde (started with lxdm), detach from that session (I'm not sure how to do that), then logout, which means you get back to the lxdm loging screen. But all your processes are killed then (including screen). You test that through a ssh connection or so? the other way, you start lxde from a linux console, you sart screen and detach, then you exit lxde, then you logout from the linux console. Or are you doing something different? I'll try and get back when I can confirm that behavior. Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] lxdm killing processes on logout
I've been battling this problem and would appreciate a sanity check. In my testing of lxde any screen session I start and then detach from gets killed when I logout. If I bypass lxdm by changing run level to 3 and using startx to start lxde and then logout the screen session survives. This also happens in a systemd build. It's not a config issue. I build both elogind and systemd to not kill user processes on logout. Thoughts? -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Sun, Dec 04, 2016 at 04:20:54PM -0600, Douglas R. Reno wrote: > > We do have a ticket. I'm working on it. Thanks. Hope that either it works for you, or that you manage to fix it. Me, I'm going back to fonts - I actually *paid* for one yesterday! ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Dec 4, 2016 4:18 PM, "Ken Moffat"wrote: On Sun, Dec 04, 2016 at 08:32:10PM +, Ken Moffat wrote: > On Sun, Dec 04, 2016 at 05:33:52PM +, Ken Moffat wrote: > > On Sun, Dec 04, 2016 at 12:10:44PM +0100, Pierre Labastie wrote: > > > Have you tried valgrind? > > > > But whatever I do, lxdm-binary seems to have already died before I > get to gdb. For me, debugging X or even just desktop programs is a > mystery. > I then tried reverting to libXfont-1.5.1 (from 1.5.2 - this older system is xorg-server-1.18.4), rebuilding xorg-server, glib, lxdm. No obvious difference. I'd already noted that lxdm-git didn't make any difference. The only other likely difference to fedora-25 is that fedora have glib-2.50.2, we have 2.50.0 (and no ticket). So I tried that (it's stable, didn't rebuild anything else). No change. Giving up. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page We do have a ticket. I'm working on it. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Sun, Dec 04, 2016 at 08:32:10PM +, Ken Moffat wrote: > On Sun, Dec 04, 2016 at 05:33:52PM +, Ken Moffat wrote: > > On Sun, Dec 04, 2016 at 12:10:44PM +0100, Pierre Labastie wrote: > > > Have you tried valgrind? > > > > But whatever I do, lxdm-binary seems to have already died before I > get to gdb. For me, debugging X or even just desktop programs is a > mystery. > I then tried reverting to libXfont-1.5.1 (from 1.5.2 - this older system is xorg-server-1.18.4), rebuilding xorg-server, glib, lxdm. No obvious difference. I'd already noted that lxdm-git didn't make any difference. The only other likely difference to fedora-25 is that fedora have glib-2.50.2, we have 2.50.0 (and no ticket). So I tried that (it's stable, didn't rebuild anything else). No change. Giving up. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Sun, Dec 04, 2016 at 05:33:52PM +, Ken Moffat wrote: > On Sun, Dec 04, 2016 at 12:10:44PM +0100, Pierre Labastie wrote: > > >> > > Have you tried valgrind? > > > > That was suggested - but that system is already very close to 100% > full (even with /opt packages moved to a different partition), which > is why I decided to use a different system. That one was only > partly-built. I built lxdm, then realised that I needed lxsession to > replicate the setup. Too many missing deps, built next chunk of my > standard desktop but haven't yet got back to lxde. > Tried and failed - no memory leaks, no errors. I tried running it from a graphic term, to see what would happen. Got a black screen, then the desktop reappeared. Log shows ==11921== TO DEBUG THIS PROCESS USING GDB: start GDB like this ==11921== /path/to/gdb /usr/sbin/lxdm-binary ==11921== and then give GDB the following command ==11921== target remote | /usr/lib/valgrind/../../bin/vgdb --pid=11921 ==11921== --pid is optional if only one valgrind process is running ==11921== So I tried it again - by preparing command lines on two ttys (as root) for valgrind and for gdb - without the --pid. First time, gdb didn't seem to like the ../../ (no such file for vgdb) so tried again specifying /usr/bin/vgdb. Also tried starting gdb (to load the symbols) first. But whatever I do, lxdm-binary seems to have already died before I get to gdb. For me, debugging X or even just desktop programs is a mystery. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On 04/12/2016 18:33, Ken Moffat wrote: On Sun, Dec 04, 2016 at 12:10:44PM +0100, Pierre Labastie wrote: Everything still works with (sed'ed) kernel 4.8.12, and up to date BLFS. OTOH, LFS is 7.10, except that glibc is 2.24 (I recompiled it to get debug symbols when debugging the firefoc/ICU segfault). I have looked more closely at your gdb backtrace. Glib asks iconv to convert a message from UTF-8 to ANSI_X3.4-1968 (a.k.a ASCII). I do not see how it could be related to anything about X (fonts or whatever). It'd be interesting to understand why glib wants the conversion, since UTF-8 seems to be the default encoding used by glib... Anyway, the bug seems that _int_malloc segfaults. The gdb trace does not have details on what is passed to malloc, but I'd say that it should be guarded against wrong parameters and return an error rather than segfaulting. Yeah, I've not yet found a likely explanation for the (presumably unset) locale. With gdb, I have seen that the exact same conversion is asked on my computer, so either the locale is not set as well, or this is the normal behaviour. The message to be converted is a log message. Actually, my /var/log/lxdm.conf file begins with: - ** Message: find greeter (nil) ** Message: find idle (nil) ** Message: 1480870181: add xserver watch [ then the usual xorg-server messages ] - Looks like the message which triggers the bug on your computer is the third one. Regards Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Sun, Dec 04, 2016 at 12:10:44PM +0100, Pierre Labastie wrote: > >> > > Everything still works with (sed'ed) kernel 4.8.12, and up to date BLFS. > > OTOH, > > LFS is 7.10, except that glibc is 2.24 (I recompiled it to get debug symbols > > when debugging the firefoc/ICU segfault). > > > > I have looked more closely at your gdb backtrace. Glib asks iconv to > > convert a > > message from UTF-8 to ANSI_X3.4-1968 (a.k.a ASCII). I do not see how it > > could > > be related to anything about X (fonts or whatever). It'd be interesting to > > understand why glib wants the conversion, since UTF-8 seems to be the > > default > > encoding used by glib... Anyway, the bug seems that _int_malloc segfaults. > > The > > gdb trace does not have details on what is passed to malloc, but I'd say > > that > > it should be guarded against wrong parameters and return an error rather > > than > > segfaulting. Yeah, I've not yet found a likely explanation for the (presumably unset) locale. Hmm, not sure if I checked from straight login as root, or only from su. > > > > That's all I can say for now. > > > > Sincerely, > > Pierre > > > Well, > > Thinking (and reading) more about it, failures in malloc are most likely to > come from overwriting some information in the memory heap, that is, some other > allocated memory is overflowded. Have you tried valgrind? > > Regards, > Pierre That was suggested - but that system is already very close to 100% full (even with /opt packages moved to a different partition), which is why I decided to use a different system. That one was only partly-built. I built lxdm, then realised that I needed lxsession to replicate the setup. Too many missing deps, built next chunk of my standard desktop but haven't yet got back to lxde. Thanks for looking. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On 04/12/2016 10:38, Pierre Labastie wrote: > On 03/12/2016 18:45, Pierre Labastie wrote: >> On 03/12/2016 17:29, Ken Moffat wrote: >>> On Sat, Dec 03, 2016 at 08:49:47AM -0700, Roger Koehler wrote: On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffatwrote: > If I can't make any progress in the next few days, I guess I'd > better raise a BLFS ticket asserting that this is broken. Could be the kernel. I just rebuilt my system with all the latest (SVN) packages except for the kernel. I built with linux 3.16.39, and lxdm works fine. I am on an intel x86_64 laptop. >>> Could be, at this point all I know is that it's broken for me, and I >>> don't know when it last worked (i.e. I forget why I gave up on it >>> after BLFS-7.6) but fedora-25 will not be using anything that old. >>> >>> ĸen >> Hi, >> I know how "worksforme" is frustrating. But I have a working version of lxdm >> with kernel 4.8.6 (needs to update, I know). >> I am building the recent versions of the gtk stack (was still with the one >> from the end of october or so). Will let you know whether it still runs after >> that. >> >> Pierre >> >> >> > Everything still works with (sed'ed) kernel 4.8.12, and up to date BLFS. OTOH, > LFS is 7.10, except that glibc is 2.24 (I recompiled it to get debug symbols > when debugging the firefoc/ICU segfault). > > I have looked more closely at your gdb backtrace. Glib asks iconv to convert a > message from UTF-8 to ANSI_X3.4-1968 (a.k.a ASCII). I do not see how it could > be related to anything about X (fonts or whatever). It'd be interesting to > understand why glib wants the conversion, since UTF-8 seems to be the default > encoding used by glib... Anyway, the bug seems that _int_malloc segfaults. The > gdb trace does not have details on what is passed to malloc, but I'd say that > it should be guarded against wrong parameters and return an error rather than > segfaulting. > > That's all I can say for now. > > Sincerely, > Pierre > Well, Thinking (and reading) more about it, failures in malloc are most likely to come from overwriting some information in the memory heap, that is, some other allocated memory is overflowded. Have you tried valgrind? Regards, Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On 03/12/2016 18:45, Pierre Labastie wrote: > On 03/12/2016 17:29, Ken Moffat wrote: >> On Sat, Dec 03, 2016 at 08:49:47AM -0700, Roger Koehler wrote: >>> On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffatwrote: If I can't make any progress in the next few days, I guess I'd better raise a BLFS ticket asserting that this is broken. >>> Could be the kernel. I just rebuilt my system with all the latest >>> (SVN) packages except for the kernel. I built with linux 3.16.39, and >>> lxdm works fine. I am on an intel x86_64 laptop. >> Could be, at this point all I know is that it's broken for me, and I >> don't know when it last worked (i.e. I forget why I gave up on it >> after BLFS-7.6) but fedora-25 will not be using anything that old. >> >> ĸen > Hi, > I know how "worksforme" is frustrating. But I have a working version of lxdm > with kernel 4.8.6 (needs to update, I know). > I am building the recent versions of the gtk stack (was still with the one > from the end of october or so). Will let you know whether it still runs after > that. > > Pierre > > > Everything still works with (sed'ed) kernel 4.8.12, and up to date BLFS. OTOH, LFS is 7.10, except that glibc is 2.24 (I recompiled it to get debug symbols when debugging the firefoc/ICU segfault). I have looked more closely at your gdb backtrace. Glib asks iconv to convert a message from UTF-8 to ANSI_X3.4-1968 (a.k.a ASCII). I do not see how it could be related to anything about X (fonts or whatever). It'd be interesting to understand why glib wants the conversion, since UTF-8 seems to be the default encoding used by glib... Anyway, the bug seems that _int_malloc segfaults. The gdb trace does not have details on what is passed to malloc, but I'd say that it should be guarded against wrong parameters and return an error rather than segfaulting. That's all I can say for now. Sincerely, Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
> From: Roger Koehler <roger.o.koeh...@gmail.com> > Date: Sat, 3 Dec 2016 08:49:47 -0700 > Subject: Re: [blfs-dev] lxdm > > On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffat <zarniwh...@ntlworld.com> wrote: > > On Tue, Nov 29, 2016 at 01:32:18AM +, Ken Moffat wrote: > >> On Mon, Nov 28, 2016 at 01:59:26PM -0700, Roger Koehler wrote: > >> > On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Reno <renodr2...@gmail.com> > >> > wrote: > >> > > > >> > > > >> > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler > >> > > <roger.o.koeh...@gmail.com> > >> > > wrote: > >> > >> > >> > >> It crashes for me, too, with the latest xorg-server. If I go back to > >> > >> the previous version of xorg-server and libXfont, it works. > >> > >> > >> > > > >> > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. > >> > > I do > >> > > have the latest libXfont though. > >> > > >> > That may be the issue. Try libXfont 1.5.2. > >> I guess you meant "try libXfont-1.5.1 ?" The new font lib is > >> libXfont2. > >> > >> I assume both need to be old - with 1.5.1 and 1.19.0 it is still > >> broken for me. > >> > >> Raised upstream at https://sourceforge.net/p/lxde/bugs/842/ > >> > > I haven't made any progress on this. According to upstream, lxdm > > doesn't reference libXfont, and it works on fedora25 with > > libXfont-1.5.2 and xorg-server-1.19.0. > > > > The segfault is after glibc seems to think my locale is > > ANSI_X3.4-1968 and failing to convert the UTF-8 message. Root's > > locale when running lxdm via gdb was en_GB.UTF-8. > > > > Fedora are using lxdm-git : I tried that but it too segfaulted for > > me - I didn't bother to look at a bt for the moment, so I guess it > > might be different. > > > > I'm starting to wonder if something in current glibc broke this, > > but basically I'm flailing around and trying to clutch at straws. > > > > If I can't make any progress in the next few days, I guess I'd > > better raise a BLFS ticket asserting that this is broken. > > Could be the kernel. I just rebuilt my system with all the latest > (SVN) packages except for the kernel. I built with linux 3.16.39, and > lxdm works fine. I am on an intel x86_64 laptop. > If you then bisect with different kernel versions (similar configs), then does it stop working at any point? rgds, akh -- -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On 03/12/2016 17:29, Ken Moffat wrote: On Sat, Dec 03, 2016 at 08:49:47AM -0700, Roger Koehler wrote: On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffatwrote: If I can't make any progress in the next few days, I guess I'd better raise a BLFS ticket asserting that this is broken. Could be the kernel. I just rebuilt my system with all the latest (SVN) packages except for the kernel. I built with linux 3.16.39, and lxdm works fine. I am on an intel x86_64 laptop. Could be, at this point all I know is that it's broken for me, and I don't know when it last worked (i.e. I forget why I gave up on it after BLFS-7.6) but fedora-25 will not be using anything that old. ĸen Hi, I know how "worksforme" is frustrating. But I have a working version of lxdm with kernel 4.8.6 (needs to update, I know). I am building the recent versions of the gtk stack (was still with the one from the end of october or so). Will let you know whether it still runs after that. Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Sat, Dec 03, 2016 at 08:49:47AM -0700, Roger Koehler wrote: > On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffatwrote: > > > > If I can't make any progress in the next few days, I guess I'd > > better raise a BLFS ticket asserting that this is broken. > > Could be the kernel. I just rebuilt my system with all the latest > (SVN) packages except for the kernel. I built with linux 3.16.39, and > lxdm works fine. I am on an intel x86_64 laptop. Could be, at this point all I know is that it's broken for me, and I don't know when it last worked (i.e. I forget why I gave up on it after BLFS-7.6) but fedora-25 will not be using anything that old. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Fri, Dec 2, 2016 at 8:30 PM, Ken Moffatwrote: > On Tue, Nov 29, 2016 at 01:32:18AM +, Ken Moffat wrote: >> On Mon, Nov 28, 2016 at 01:59:26PM -0700, Roger Koehler wrote: >> > On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Reno >> > wrote: >> > > >> > > >> > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler >> > > >> > > wrote: >> > >> >> > >> It crashes for me, too, with the latest xorg-server. If I go back to >> > >> the previous version of xorg-server and libXfont, it works. >> > >> >> > > >> > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. I >> > > do >> > > have the latest libXfont though. >> > >> > That may be the issue. Try libXfont 1.5.2. >> I guess you meant "try libXfont-1.5.1 ?" The new font lib is >> libXfont2. >> >> I assume both need to be old - with 1.5.1 and 1.19.0 it is still >> broken for me. >> >> Raised upstream at https://sourceforge.net/p/lxde/bugs/842/ >> > I haven't made any progress on this. According to upstream, lxdm > doesn't reference libXfont, and it works on fedora25 with > libXfont-1.5.2 and xorg-server-1.19.0. > > The segfault is after glibc seems to think my locale is > ANSI_X3.4-1968 and failing to convert the UTF-8 message. Root's > locale when running lxdm via gdb was en_GB.UTF-8. > > Fedora are using lxdm-git : I tried that but it too segfaulted for > me - I didn't bother to look at a bt for the moment, so I guess it > might be different. > > I'm starting to wonder if something in current glibc broke this, > but basically I'm flailing around and trying to clutch at straws. > > If I can't make any progress in the next few days, I guess I'd > better raise a BLFS ticket asserting that this is broken. Could be the kernel. I just rebuilt my system with all the latest (SVN) packages except for the kernel. I built with linux 3.16.39, and lxdm works fine. I am on an intel x86_64 laptop. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Fri, Dec 02, 2016 at 09:55:27PM -0600, Douglas R. Reno wrote: > Ken, > > I'll have to build another DE that uses D-Bus when doing GNOME (and for > some other stuff). > > Do you want me to build LXDM to round a complete LXDE installation and > report back to you ASAP? If it suits you. More data always helps, and particularly if it works for you. I remembered I've got another recent system on my test machine, I'm going to try to build (maybe only) lxdm there - not sure of the versions of everything, but I know xorg-server is 1.18.x ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Fri, Dec 2, 2016 at 9:30 PM, Ken Moffatwrote: > On Tue, Nov 29, 2016 at 01:32:18AM +, Ken Moffat wrote: > > On Mon, Nov 28, 2016 at 01:59:26PM -0700, Roger Koehler wrote: > > > On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Reno > wrote: > > > > > > > > > > > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler < > roger.o.koeh...@gmail.com> > > > > wrote: > > > >> > > > >> It crashes for me, too, with the latest xorg-server. If I go back to > > > >> the previous version of xorg-server and libXfont, it works. > > > >> > > > > > > > > I don't have xorg-server-1.19 on my workstation, I'm still at > 1.18.x. I do > > > > have the latest libXfont though. > > > > > > That may be the issue. Try libXfont 1.5.2. > > I guess you meant "try libXfont-1.5.1 ?" The new font lib is > > libXfont2. > > > > I assume both need to be old - with 1.5.1 and 1.19.0 it is still > > broken for me. > > > > Raised upstream at https://sourceforge.net/p/lxde/bugs/842/ > > > I haven't made any progress on this. According to upstream, lxdm > doesn't reference libXfont, and it works on fedora25 with > libXfont-1.5.2 and xorg-server-1.19.0. > > The segfault is after glibc seems to think my locale is > ANSI_X3.4-1968 and failing to convert the UTF-8 message. Root's > locale when running lxdm via gdb was en_GB.UTF-8. > > Fedora are using lxdm-git : I tried that but it too segfaulted for > me - I didn't bother to look at a bt for the moment, so I guess it > might be different. > > I'm starting to wonder if something in current glibc broke this, > but basically I'm flailing around and trying to clutch at straws. > > If I can't make any progress in the next few days, I guess I'd > better raise a BLFS ticket asserting that this is broken. > > Ken, I'll have to build another DE that uses D-Bus when doing GNOME (and for some other stuff). Do you want me to build LXDM to round a complete LXDE installation and report back to you ASAP? -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Tue, Nov 29, 2016 at 01:32:18AM +, Ken Moffat wrote: > On Mon, Nov 28, 2016 at 01:59:26PM -0700, Roger Koehler wrote: > > On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Reno> > wrote: > > > > > > > > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler > > > wrote: > > >> > > >> It crashes for me, too, with the latest xorg-server. If I go back to > > >> the previous version of xorg-server and libXfont, it works. > > >> > > > > > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. I do > > > have the latest libXfont though. > > > > That may be the issue. Try libXfont 1.5.2. > I guess you meant "try libXfont-1.5.1 ?" The new font lib is > libXfont2. > > I assume both need to be old - with 1.5.1 and 1.19.0 it is still > broken for me. > > Raised upstream at https://sourceforge.net/p/lxde/bugs/842/ > I haven't made any progress on this. According to upstream, lxdm doesn't reference libXfont, and it works on fedora25 with libXfont-1.5.2 and xorg-server-1.19.0. The segfault is after glibc seems to think my locale is ANSI_X3.4-1968 and failing to convert the UTF-8 message. Root's locale when running lxdm via gdb was en_GB.UTF-8. Fedora are using lxdm-git : I tried that but it too segfaulted for me - I didn't bother to look at a bt for the moment, so I guess it might be different. I'm starting to wonder if something in current glibc broke this, but basically I'm flailing around and trying to clutch at straws. If I can't make any progress in the next few days, I guess I'd better raise a BLFS ticket asserting that this is broken. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Mon, Nov 28, 2016 at 01:59:26PM -0700, Roger Koehler wrote: > On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Renowrote: > > > > > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler > > wrote: > >> > >> It crashes for me, too, with the latest xorg-server. If I go back to > >> the previous version of xorg-server and libXfont, it works. > >> > > > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. I do > > have the latest libXfont though. > > That may be the issue. Try libXfont 1.5.2. I guess you meant "try libXfont-1.5.1 ?" The new font lib is libXfont2. I assume both need to be old - with 1.5.1 and 1.19.0 it is still broken for me. Raised upstream at https://sourceforge.net/p/lxde/bugs/842/ ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Mon, Nov 28, 2016 at 1:55 PM, Douglas R. Renowrote: > > > On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehler > wrote: >> >> On Mon, Nov 28, 2016 at 1:44 PM, Ken Moffat >> wrote: >> > On my build without libXfont, I reported that lxdm dies. I don't >> > think I posted that it is in malloc in glibc. Since then I've >> > installed old libXfont, bdftopcf, adobe-100dpi bitmap fonts >> > (Courier, Helvetica, New Century Schoolbook, Times). I've just >> > rebuilt lxdm, it still segfaults. >> > >> > This is with LFS svn from 20161009 (r11131) with newer kernel, >> > tzdata, iproute - the base system was to do an LFS kernel update >> > because of the vulnerability. A few things were built at that >> > time, mostly from BLFS r17741 of 2016-09-07, but Xorg and everything >> > else used versions as at r17980 (2016-11-19) or newer. >> > >> > Since I've not had lxdm working since the days of BLFS-7.6, and I >> > don't like it, I'm not minded to look at it. Douglas said it works >> > for him, so maybe this is specific to my machine or how I build it. >> >> It crashes for me, too, with the latest xorg-server. If I go back to >> the previous version of xorg-server and libXfont, it works. >> > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. I do > have the latest libXfont though. That may be the issue. Try libXfont 1.5.2. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Mon, Nov 28, 2016 at 2:48 PM, Roger Koehlerwrote: > On Mon, Nov 28, 2016 at 1:44 PM, Ken Moffat > wrote: > > On my build without libXfont, I reported that lxdm dies. I don't > > think I posted that it is in malloc in glibc. Since then I've > > installed old libXfont, bdftopcf, adobe-100dpi bitmap fonts > > (Courier, Helvetica, New Century Schoolbook, Times). I've just > > rebuilt lxdm, it still segfaults. > > > > This is with LFS svn from 20161009 (r11131) with newer kernel, > > tzdata, iproute - the base system was to do an LFS kernel update > > because of the vulnerability. A few things were built at that > > time, mostly from BLFS r17741 of 2016-09-07, but Xorg and everything > > else used versions as at r17980 (2016-11-19) or newer. > > > > Since I've not had lxdm working since the days of BLFS-7.6, and I > > don't like it, I'm not minded to look at it. Douglas said it works > > for him, so maybe this is specific to my machine or how I build it. > > It crashes for me, too, with the latest xorg-server. If I go back to > the previous version of xorg-server and libXfont, it works. > > I don't have xorg-server-1.19 on my workstation, I'm still at 1.18.x. I do have the latest libXfont though. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Mon, Nov 28, 2016 at 1:44 PM, Ken Moffatwrote: > On my build without libXfont, I reported that lxdm dies. I don't > think I posted that it is in malloc in glibc. Since then I've > installed old libXfont, bdftopcf, adobe-100dpi bitmap fonts > (Courier, Helvetica, New Century Schoolbook, Times). I've just > rebuilt lxdm, it still segfaults. > > This is with LFS svn from 20161009 (r11131) with newer kernel, > tzdata, iproute - the base system was to do an LFS kernel update > because of the vulnerability. A few things were built at that > time, mostly from BLFS r17741 of 2016-09-07, but Xorg and everything > else used versions as at r17980 (2016-11-19) or newer. > > Since I've not had lxdm working since the days of BLFS-7.6, and I > don't like it, I'm not minded to look at it. Douglas said it works > for him, so maybe this is specific to my machine or how I build it. It crashes for me, too, with the latest xorg-server. If I go back to the previous version of xorg-server and libXfont, it works. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm
On Mon, Nov 28, 2016 at 2:44 PM, Ken Moffatwrote: > On my build without libXfont, I reported that lxdm dies. I don't > think I posted that it is in malloc in glibc. Since then I've > installed old libXfont, bdftopcf, adobe-100dpi bitmap fonts > (Courier, Helvetica, New Century Schoolbook, Times). I've just > rebuilt lxdm, it still segfaults. > > This is with LFS svn from 20161009 (r11131) with newer kernel, > tzdata, iproute - the base system was to do an LFS kernel update > because of the vulnerability. A few things were built at that > time, mostly from BLFS r17741 of 2016-09-07, but Xorg and everything > else used versions as at r17980 (2016-11-19) or newer. > > Since I've not had lxdm working since the days of BLFS-7.6, and I > don't like it, I'm not minded to look at it. Douglas said it works > for him, so maybe this is specific to my machine or how I build it. > > ĸen > Don't discount the fact that I said that it's fine. Until I build it on my development system, I can't be 100% sure if its fine or not. I've had issues with Pth in the past, but that was due to a typo in my script. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] lxdm
On my build without libXfont, I reported that lxdm dies. I don't think I posted that it is in malloc in glibc. Since then I've installed old libXfont, bdftopcf, adobe-100dpi bitmap fonts (Courier, Helvetica, New Century Schoolbook, Times). I've just rebuilt lxdm, it still segfaults. This is with LFS svn from 20161009 (r11131) with newer kernel, tzdata, iproute - the base system was to do an LFS kernel update because of the vulnerability. A few things were built at that time, mostly from BLFS r17741 of 2016-09-07, but Xorg and everything else used versions as at r17980 (2016-11-19) or newer. Since I've not had lxdm working since the days of BLFS-7.6, and I don't like it, I'm not minded to look at it. Douglas said it works for him, so maybe this is specific to my machine or how I build it. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
On Sun, Sep 07, 2014 at 12:49:38AM -0500, Bruce Dubbs wrote: Right. There is a boot script for ldxm in svn: bootscripts/blfs/init.d/lxdm. Ah, yes. Added this month, but as of yesterday the book still references 20140827. And should we not suggest installing the bootscript on the lxdm page ? ĸen -- Nanny Ogg usually went to bed early. After all, she was an old lady. Sometimes she went to bed as early as 6 a.m. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
On 07-09-2014 11:42, Ken Moffat wrote: On Sun, Sep 07, 2014 at 12:49:38AM -0500, Bruce Dubbs wrote: Right. There is a boot script for ldxm in svn: bootscripts/blfs/init.d/lxdm. Ah, yes. Added this month, but as of yesterday the book still references 20140827. And should we not suggest installing the bootscript on the lxdm page ? ĸen Sorry, ĸen. Only now I noticed the thread. Yes, we have it in the page (perhaps it is misplaced, should go further up, in the page?): {{{ Boot Script Install the /etc/rc.d/init.d/lxdm init script from the blfs-bootscripts-20140827 package. make install-lxdm }}} I didn't know I had to modify the bootscript version in the book. Other thing I didn't notice, was the kdm page. Yes, I think it could be started from inittab, instead of a bootscript. All being said, I would appreciate you criticizing, or better yet, modifying the page, improving English, etc. Thanks. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
On Sun, Sep 07, 2014 at 12:02:21PM -0300, Fernando de Oliveira wrote: On 07-09-2014 11:42, Ken Moffat wrote: On Sun, Sep 07, 2014 at 12:49:38AM -0500, Bruce Dubbs wrote: Right. There is a boot script for ldxm in svn: bootscripts/blfs/init.d/lxdm. Ah, yes. Added this month, but as of yesterday the book still references 20140827. And should we not suggest installing the bootscript on the lxdm page ? ĸen Sorry, ĸen. Only now I noticed the thread. Yes, we have it in the page (perhaps it is misplaced, should go further up, in the page?): {{{ Boot Script I can see it now - I used search on firefox last night, and somehow did not find it. I think I was up way past my bedtime. Install the /etc/rc.d/init.d/lxdm init script from the blfs-bootscripts-20140827 package. make install-lxdm }}} Now that I have found it, it looks ok and is within 'Configuring', so all is good. I didn't know I had to modify the bootscript version in the book. Yeah, some of the editing tasks are black magic ;) I'm sure it is documented somewhere, but probably under a heading Beware of the leopard.¹ Of course, if that doesn't regenerate the bootscripts when the book next renders, then I'll be stuck. Other thing I didn't notice, was the kdm page. Yes, I think it could be started from inittab, instead of a bootscript. I have not ued a display manager on a BLFS system, except when I used gdm to test gnome3 stuff, for many years. Any way of activating it works for me. All being said, I would appreciate you criticizing, or better yet, modifying the page, improving English, etc. Thanks. Now that I'm fully awake, it looks ok. ĸen 1. http://hitchhikerguidetothegalaxy.blogspot.co.uk/2006/04/beware-of-leopard-douglas-adams-quote.html -- Nanny Ogg usually went to bed early. After all, she was an old lady. Sometimes she went to bed as early as 6 a.m. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
On 07-09-2014 15:43, Ken Moffat wrote: On Sun, Sep 07, 2014 at 12:02:21PM -0300, Fernando de Oliveira wrote: On 07-09-2014 11:42, Ken Moffat wrote: On Sun, Sep 07, 2014 at 12:49:38AM -0500, Bruce Dubbs wrote: Right. There is a boot script for ldxm in svn: bootscripts/blfs/init.d/lxdm. Ah, yes. Added this month, but as of yesterday the book still references 20140827. And should we not suggest installing the bootscript on the lxdm page ? ĸen Sorry, ĸen. Only now I noticed the thread. Yes, we have it in the page (perhaps it is misplaced, should go further up, in the page?): {{{ Boot Script I can see it now - I used search on firefox last night, and somehow did not find it. I think I was up way past my bedtime. Install the /etc/rc.d/init.d/lxdm init script from the blfs-bootscripts-20140827 package. make install-lxdm }}} Now that I have found it, it looks ok and is within 'Configuring', so all is good. I didn't know I had to modify the bootscript version in the book. Yeah, some of the editing tasks are black magic ;) I'm sure it is documented somewhere, but probably under a heading Beware of the leopard.¹ Of course, if that doesn't regenerate the bootscripts when the book next renders, then I'll be stuck. Other thing I didn't notice, was the kdm page. Yes, I think it could be started from inittab, instead of a bootscript. I have not ued a display manager on a BLFS system, except when I used gdm to test gnome3 stuff, for many years. Any way of activating it works for me. All being said, I would appreciate you criticizing, or better yet, modifying the page, improving English, etc. Thanks. Now that I'm fully awake, it looks ok. ĸen 1. http://hitchhikerguidetothegalaxy.blogspot.co.uk/2006/04/beware-of-leopard-douglas-adams-quote.html LOL Thanks. Anyway, I just moved it up in the page: revision 14226. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
On Sun, 2014-09-07 at 04:09 +0100, Ken Moffat wrote: I'm looking at what I might test when we get to 7.6. I've repartitioned to get what to me seem enormous root fs's (15GB, with the sources elsewhere), and we have a lot of things I've never tried, so I thought I would try lxde - principally to test lxdm (if I do get round to putting LFS on my netbook, I'll need a graphical login to simplify controlling sleep and the network). For lxdm, the only reference to a bootscript I can find is |You can manually start lxdm, e.g, if the bootscript has been |installed, by running, as root user: |/etc/rc.d/init.d/lxdm start I thought this was a case ofi accidentally forgetting to reference the bootscript (compare e.g. cups) on this page, but looking in the 20140827 bootscripts I cannot see any obvious target for lxdm ? Compparing other DMs, we (still) have a bootscript for gdm, but for kdm we do things differently: |cat /etc/inittab EOF |kd:5:respawn:/opt/kde/bin/kdm |EOF and |sed -i 's#id:3:initdefault:#id:5:initdefault:#' /etc/inittab I know we tend to do things differently to show the possible variations, but should be have a bootscript for lxde, and if not should we do it like kdm ? Confused. ĸen -- Nanny Ogg usually went to bed early. After all, she was an old lady. Sometimes she went to bed as early as 6 a.m. Hello Ken, The bootscript should already be there as both Fernando and Bruce worked on it. The systemd version of lxdm provides a native unit file. Regards, Christopher. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] lxdm : bootscript ?
Christopher Gregory wrote: On Sun, 2014-09-07 at 04:09 +0100, Ken Moffat wrote: I'm looking at what I might test when we get to 7.6. I've repartitioned to get what to me seem enormous root fs's (15GB, with the sources elsewhere), and we have a lot of things I've never tried, so I thought I would try lxde - principally to test lxdm (if I do get round to putting LFS on my netbook, I'll need a graphical login to simplify controlling sleep and the network). For lxdm, the only reference to a bootscript I can find is |You can manually start lxdm, e.g, if the bootscript has been |installed, by running, as root user: |/etc/rc.d/init.d/lxdm start I thought this was a case ofi accidentally forgetting to reference the bootscript (compare e.g. cups) on this page, but looking in the 20140827 bootscripts I cannot see any obvious target for lxdm ? Compparing other DMs, we (still) have a bootscript for gdm, but for kdm we do things differently: |cat /etc/inittab EOF |kd:5:respawn:/opt/kde/bin/kdm |EOF and |sed -i 's#id:3:initdefault:#id:5:initdefault:#' /etc/inittab I know we tend to do things differently to show the possible variations, but should be have a bootscript for lxde, and if not should we do it like kdm ? Confused. ĸen -- Nanny Ogg usually went to bed early. After all, she was an old lady. Sometimes she went to bed as early as 6 a.m. Hello Ken, The bootscript should already be there as both Fernando and Bruce worked on it. The systemd version of lxdm provides a native unit file. Right. There is a boot script for ldxm in svn: bootscripts/blfs/init.d/lxdm. -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 03-09-2014 13:04, Bruce Dubbs wrote: akhiezer wrote: From: Christopher Gregory m...@pc-networking-services.com To: blfs-dev@lists.linuxfromscratch.org Date: Wed, 03 Sep 2014 17:03:46 +1200 Subject: Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d] On Tue, 2014-09-02 at 20:05 -0300, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. -- []s, Fernando Hello Fernando, I have added this to my ~/.bashrc source /etc/profile (Hmmm. Depends of course on how are the startup files organised. Some folks chain: { /etc/profile (+) /etc/profile.d/ } -- { ~/.profile | ~/.bash_profile | ~/.bash_login} The order here is defined by bash and ~/.profile is the last option. Also the syntax might be confused as a pipe. The actual sequence is: if[ -e ~/.bash_profile ]; then source ~/.bash_profile; elsif [ -e ~/.bash_login ]; then source ~/.bash_login; elsif [ -e ~/.profile ]; then source ~/.profile; fi -- { ~/.bashrc } Depends if it's a login shell or not. Of course is can be explicitly sourced from ~/.bash_profile. etc. -- { /etc/bashrc } Only used if explicitly called from another file (e.g. ~/.bashrc). ; and at most perhaps call some parts of /etc/profile.d/ from the bashrc stuff. But yes, I know that you're talking about what worked for you here. The things that need to be set from ~/.bashrc are aliases and PS1. The other unusual possibility is non-exported shell variables. ~/.bashrc is only automatically sourced for interactive subshells. If the profile files are set up properly, then /etc/profile should only be called once at login. The issue is to make sure ldxm or other dm does indeed run it. -- Bruce Bruce, thank you very much for correcting me about the runlevel in the patch. I knew about that, but this was the first time that I needed to be careful about it, and was reckless. akhiezer, thanks for the interest. I am still investigating about how and where /etc/profile is being sourced, but until now, I suspect that the DM does not do it, it is done by the session manager. I am naming like that, because it can be OpenBox, LXDE, etc. And that is the reason lxdm script seem to only deal with LANG, either by i18n or by testing the env variable itself. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 04-09-2014 07:23, Fernando de Oliveira wrote: I am still investigating about how and where /etc/profile is being sourced, but until now, I suspect that the DM does not do it, it is done by the session manager. I am naming like that, because it can be OpenBox, LXDE, etc. And that is the reason lxdm script seem to only deal with LANG, either by i18n or by testing the env variable itself. I was wrong: forgot that I am already using another fix for sourcing /etc/profile, remembered today. This seems the last fix needed for LXDM. Thank you all, very much. Fixed at revision 14189. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 02-09-2014 20:47, Armin K. wrote: On 09/03/2014 01:05 AM, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. pathappend and pathremove are defined in /etc/profile itself. Are you sure that you were sourcing it (. /etc/profile) (it's possible, by mistake, that you were executing it by missing the . (dot))? Also, it's highly unlikely that LXDM could harm any X terminal in any way, simply because any terminal emulator runs bash (well, default user shell I think) and bash itself takes care of sourcing the startup scripts and stuff like that. Thanks for your reply, Armin. Well, I was using source /etc/profile/ Now, just to be sure, I used . /etc/profile no difference. Error appears as terminal opens, but it seems that just running bash again there would produce the same error: bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathappend' bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathremove' bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathprepend' fernando [ ~ ]$ ./sshfsNovoGamer.sh /bin/bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathremove' /bin/bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathappend' /bin/bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathprepend' because the script I ran has #!/bin/bash. Translations: erro de sintaxe: fim prematuro do arquivo === syntax error: unexpected (premature) end of file. erro ao importar a definição da função para `path... === error importing function definition for `path... -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 09/03/2014 12:01 PM, Fernando de Oliveira wrote: On 02-09-2014 20:47, Armin K. wrote: On 09/03/2014 01:05 AM, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. pathappend and pathremove are defined in /etc/profile itself. Are you sure that you were sourcing it (. /etc/profile) (it's possible, by mistake, that you were executing it by missing the . (dot))? Also, it's highly unlikely that LXDM could harm any X terminal in any way, simply because any terminal emulator runs bash (well, default user shell I think) and bash itself takes care of sourcing the startup scripts and stuff like that. Thanks for your reply, Armin. Well, I was using source /etc/profile/ Now, just to be sure, I used . /etc/profile no difference. Error appears as terminal opens, but it seems that just running bash again there would produce the same error: bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathappend' bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathremove' bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathprepend' fernando [ ~ ]$ ./sshfsNovoGamer.sh /bin/bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathremove' /bin/bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathappend' /bin/bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathprepend' because the script I ran has #!/bin/bash. Translations: erro de sintaxe: fim prematuro do arquivo === syntax error: unexpected (premature) end of file. erro ao importar a definição da função para `path... === error importing function definition for `path... Hm. Still makes zero sense to me, but iirc, /etc/profile.d/i18n.sh only exports the language variables, so would that be fine? -- Note: My last name is not Krejzi. signature.asc Description: OpenPGP digital signature -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
From: Christopher Gregory m...@pc-networking-services.com To: blfs-dev@lists.linuxfromscratch.org Date: Wed, 03 Sep 2014 17:03:46 +1200 Subject: Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d] On Tue, 2014-09-02 at 20:05 -0300, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. -- []s, Fernando Hello Fernando, I have added this to my ~/.bashrc source /etc/profile (Hmmm. Depends of course on how are the startup files organised. Some folks chain: { /etc/profile (+) /etc/profile.d/ } -- { ~/.profile | ~/.bash_profile | ~/.bash_login} -- { ~/.bashrc } -- { /etc/bashrc } ; and at most perhaps call some parts of /etc/profile.d/ from the bashrc stuff. But yes, I know that you're talking about what worked for you here. ) I did this so that when I am using a terminal such as in lxde or in gnome-terminal that the title and directory gets correctly displayed when opening a new terminal in a tab or window. Perhaps just a side-note, but: certainly at least for twm, if the locale isn't liked by X then titlebars of windows, and strings in wm menus, c, just all show 'untitled' (or 'Untitled'). Perhaps you're hitting a similar thing? (Ref e.g.: Warning: locale not supported by Xlib, locale set to C http://www.linuxfromscratch.org/lfs/view/7.5/chapter07/profile.html ; and innumerable reports down the years of the 'untitled' behaviour. ) rgds, akh -- -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 03-09-2014 07:43, Armin K. wrote: On 09/03/2014 12:01 PM, Fernando de Oliveira wrote: On 02-09-2014 20:47, Armin K. wrote: On 09/03/2014 01:05 AM, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. pathappend and pathremove are defined in /etc/profile itself. Are you sure that you were sourcing it (. /etc/profile) (it's possible, by mistake, that you were executing it by missing the . (dot))? Also, it's highly unlikely that LXDM could harm any X terminal in any way, simply because any terminal emulator runs bash (well, default user shell I think) and bash itself takes care of sourcing the startup scripts and stuff like that. Thanks for your reply, Armin. Well, I was using source /etc/profile/ Now, just to be sure, I used . /etc/profile no difference. Error appears as terminal opens, but it seems that just running bash again there would produce the same error: bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathappend' bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathremove' bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo bash: erro ao importar a definição da função para `pathprepend' fernando [ ~ ]$ ./sshfsNovoGamer.sh /bin/bash: pathremove: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathremove' /bin/bash: pathappend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathappend' /bin/bash: pathprepend: line 1: erro de sintaxe: fim prematuro do arquivo /bin/bash: erro ao importar a definição da função para `pathprepend' because the script I ran has #!/bin/bash. Translations: erro de sintaxe: fim prematuro do arquivo === syntax error: unexpected (premature) end of file. erro ao importar a definição da função para `path... === error importing function definition for `path... Hm. Still makes zero sense to me, but iirc, /etc/profile.d/i18n.sh only exports the language variables, so would that be fine? Thanks!! This really works!!! [ -f /etc/profile.d/i18n.sh ] . /etc/profile.d/i18n.sh Actually, it should have come to my mind, after the problem with /etc/profile and even before, because this is the first executable line in lxdm (doh!): [ -f /etc/sysconfig/i18n ] . /etc/sysconfig/i18n For the time being, will just correct in the configuration section of the page, but later, will try to build it with the correction in place, which is more suitable for BLFS. Thanks again. I am just going to commit it now, with some other modifications I was just finishing when received your message. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
akhiezer wrote: From: Christopher Gregory m...@pc-networking-services.com To: blfs-dev@lists.linuxfromscratch.org Date: Wed, 03 Sep 2014 17:03:46 +1200 Subject: Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d] On Tue, 2014-09-02 at 20:05 -0300, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. -- []s, Fernando Hello Fernando, I have added this to my ~/.bashrc source /etc/profile (Hmmm. Depends of course on how are the startup files organised. Some folks chain: { /etc/profile (+) /etc/profile.d/ } -- { ~/.profile | ~/.bash_profile | ~/.bash_login} The order here is defined by bash and ~/.profile is the last option. Also the syntax might be confused as a pipe. The actual sequence is: if[ -e ~/.bash_profile ]; then source ~/.bash_profile; elsif [ -e ~/.bash_login ]; then source ~/.bash_login; elsif [ -e ~/.profile ]; then source ~/.profile; fi -- { ~/.bashrc } Depends if it's a login shell or not. Of course is can be explicitly sourced from ~/.bash_profile. etc. -- { /etc/bashrc } Only used if explicitly called from another file (e.g. ~/.bashrc). ; and at most perhaps call some parts of /etc/profile.d/ from the bashrc stuff. But yes, I know that you're talking about what worked for you here. The things that need to be set from ~/.bashrc are aliases and PS1. The other unusual possibility is non-exported shell variables. ~/.bashrc is only automatically sourced for interactive subshells. If the profile files are set up properly, then /etc/profile should only be called once at login. The issue is to make sure ldxm or other dm does indeed run it. -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
Date: Wed, 03 Sep 2014 11:04:18 -0500 From: Bruce Dubbs bruce.du...@gmail.com To: BLFS Development List blfs-dev@lists.linuxfromscratch.org Subject: Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d] . . (Hmmm. Depends of course on how are the startup files organised. Some folks chain: { /etc/profile (+) /etc/profile.d/ } -- { ~/.profile | ~/.bash_profile | ~/.bash_login} The order here is defined by bash and ~/.profile is the last option. Also the syntax might be confused as a pipe. The actual sequence is: if[ -e ~/.bash_profile ]; then source ~/.bash_profile; elsif [ -e ~/.bash_login ]; then source ~/.bash_login; elsif [ -e ~/.profile ]; then source ~/.profile; fi Yes, but so what. You're regurgitating bash code: whereas the above just schematically shows what some folks do - the '|' wasn't meaning to imply a partic ordering; and instead that some folks just choose to use only one of the three items. -- { ~/.bashrc } Depends if it's a login shell or not. Yeah, but again so what: you're just regurg bash code/manpage. Of course is can be explicitly sourced from ~/.bash_profile. etc. Yes that's precisely what the 'example of what some folks do' ... er, _says_. -- { /etc/bashrc } Only used if explicitly called from another file (e.g. ~/.bashrc). Ditto. ; and at most perhaps call some parts of /etc/profile.d/ from the bashrc stuff. But yes, I know that you're talking about what worked for you here. The things that need to be set from ~/.bashrc are aliases and PS1. The other unusual possibility is non-exported shell variables. ~/.bashrc is only automatically sourced for interactive subshells. In what way does that relate to the post. Why are you bothering to write it. If the profile files are set up properly, then /etc/profile should only be called once at login. The issue is to make sure ldxm or other dm does indeed run it. Ditto. Bruce, I think you're just yet again completely failing to grasp the simplest info in the likes of the post. You're almost over-thinking it. rgds, akh -- Bruce -- -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 01-09-2014 17:33, Fernando de Oliveira wrote: On 01-09-2014 16:50, Bruce Dubbs wrote: ferna...@higgs.linuxfromscratch.org wrote: --- trunk/bootscripts/Makefile Mon Sep 1 11:26:44 2014 (r14162) +++ trunk/bootscripts/MakefileMon Sep 1 12:34:36 2014 (r14163) @@ -277,9 +277,9 @@ ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc0.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc1.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc2.d/K05lxdm -ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc3.d/K05lxdm +ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc3.d/S95lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc4.d/K05lxdm -ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc5.d/S95lxdm +ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc5.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc6.d/K05lxdm I think you had it right the first time. The standard definition of run level 3 is without a gui. When remembered to test rebooting, after the commit, didn't start. Only if in runlevel 3. No idea why. I don't know much about this, but I think I know where the problem starts and present a first solution. I have now: $ head -n8 /etc/inittab # Begin /etc/inittab id:5:initdefault: si::sysinit:/etc/rc.d/init.d/rc S l0:0:wait:/etc/rc.d/init.d/rc 0 l1:S1:wait:/etc/rc.d/init.d/rc 1 Before, I had: id:3:initdefault: This suggested me that LXDM, which is taking the responsibility of authenticating with PAM, was not allowed, because first, we had to login at runlevel 3. It simply didn't even start (no mention anywhere in the logs). Now, it starts at runlevel 5. As my knowledge is very limited, and I am facing two alternatives, each modifying fundamental rules, what is the best solution? 1. Leave the bootscripts as I modified yesterday to start at runlevel 3. 2. Change in /etc/inittab to have id:5:initdefault. On Mon, 01 Sep 2014 22:55:01 +0200, Armin wrote: On 09/01/2014 10:25 PM, Fernando de Oliveira wrote: Problem: LXDM always start as if LANG=C. Now, I found another solution: sed -i '1a\\nexport LANG=pt_BR.UTF-8' /usr/sbin/lxdm because this is a script calling the true program, lxdm-binary. I haven't built LXDM, but since you say the lxdm is a wrapper script, you can make it source /etc/profile. On systemd systems, systemd-localed is responsible for setting the locale for everything but tty environment, and that is manually handled by /etc/profile.d script. Thank you very much, Armin, localization problem solved with your suggestion of including source /etc/profile in /usr/sbin/lxdm!!! -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
Fernando de Oliveira wrote: On 01-09-2014 17:33, Fernando de Oliveira wrote: On 01-09-2014 16:50, Bruce Dubbs wrote: ferna...@higgs.linuxfromscratch.org wrote: --- trunk/bootscripts/Makefile Mon Sep 1 11:26:44 2014 (r14162) +++ trunk/bootscripts/MakefileMon Sep 1 12:34:36 2014 (r14163) @@ -277,9 +277,9 @@ ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc0.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc1.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc2.d/K05lxdm -ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc3.d/K05lxdm +ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc3.d/S95lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc4.d/K05lxdm -ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc5.d/S95lxdm +ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc5.d/K05lxdm ln -sf ../init.d/lxdm ${EXTDIR}/rc.d/rc6.d/K05lxdm I think you had it right the first time. The standard definition of run level 3 is without a gui. When remembered to test rebooting, after the commit, didn't start. Only if in runlevel 3. No idea why. I don't know much about this, but I think I know where the problem starts and present a first solution. I have now: $ head -n8 /etc/inittab # Begin /etc/inittab id:5:initdefault: si::sysinit:/etc/rc.d/init.d/rc S l0:0:wait:/etc/rc.d/init.d/rc 0 l1:S1:wait:/etc/rc.d/init.d/rc 1 Before, I had: id:3:initdefault: This suggested me that LXDM, which is taking the responsibility of authenticating with PAM, was not allowed, because first, we had to login at runlevel 3. It simply didn't even start (no mention anywhere in the logs). Now, it starts at runlevel 5. As my knowledge is very limited, and I am facing two alternatives, each modifying fundamental rules, what is the best solution? 1. Leave the bootscripts as I modified yesterday to start at runlevel 3. 2. Change in /etc/inittab to have id:5:initdefault. Yes, have the user change initdefault, if desired. The 'init 5' command also should work from the command line. -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LXDM: localization fixed but runlevel two different fixes [Was Re: ... bootscripts: . blfs/init.d]
On 09/03/2014 01:05 AM, Fernando de Oliveira wrote: Fixed at revision 14177. Thanks Armin and Bruce. Armin, I could not use your suggestion (my preferred one), because each time an x-terminal emulator started, several error messages about not finding patappend and pathremove appeared in the newly open terminal.. I recall that they used to be unset, and is not anymore, thought it was /etc/profile that unset it, but was not. If I could find how to avoid this problem, I would replace the fix I put in the book, and don't like, by yours. pathappend and pathremove are defined in /etc/profile itself. Are you sure that you were sourcing it (. /etc/profile) (it's possible, by mistake, that you were executing it by missing the . (dot))? Also, it's highly unlikely that LXDM could harm any X terminal in any way, simply because any terminal emulator runs bash (well, default user shell I think) and bash itself takes care of sourcing the startup scripts and stuff like that. -- Note: My last name is not Krejzi. signature.asc Description: OpenPGP digital signature -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] LXDM Localization
This is an old problem I faced, when used LXDM (stopped with 0.4.1, I think). Today, I don't normally use a DM. Problem: LXDM always start as if LANG=C. I change in the corresponding place in the screen, after reboot, C again. In the past I solved it, by modifying the start section of the boot script like this: start) env LANG=pt_BR.utf8 /usr/sbin/lxdm -d ;; It is not a good general solution, worse yet for the book. Now, I found another solution: sed -i '1a\\nexport LANG=pt_BR.UTF-8' /usr/sbin/lxdm because this is a script calling the true program, lxdm-binary. If someone (Christopher?) has a better solution, I would much appreciate to learn, please. If not, I think I need to introduce one para (as optional), with an explanation/para and then screenuserinputsed -i '1a\\nexport your_LANG' /usr/sbin/lxdm/userinput/screen Any help would be much appreciated. Thanks. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page