Re: [blfs-support] Suggestions on Desktop Environment
> > (or at least a > 'world-within-a-world'); > ...a society *within* a society; where you have to get up to get down, and where getting it on *usually* means getting them off. The Wicker Rap -- circa 1980 I'm sorry, I just couldn't resist that. Richard -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Wed, Dec 4, 2013 at 7:27 PM, Ken Moffat wrote: > > In my previous reply I included an invalid-unicode symbol, code > U+FFFD displaying as reverse-video question mark : � : if you read > the mail on the LFS-7.4 machine (or copy this paragraph to a file, > scp that file to the LFS machine, and then read it), does the > resulting glyph appear correctly, and if not, does it match the > white squares you had [ you need to revert the LANG and UNICODE > settings ] ? I can see reverse-video question mark on LFS box, but reverse area is square in firefox/gmail and oval in lfs (I assume due to different font) > There are (at least) three things to be done when moving from > non-unicode to unicode: > > (i.) set the environment to use the unicode versions such as > en_US.UTF-8. > > (ii.) use a unicode console font. > > (iii.) adjust environment or other settings for individual > applications, e.g. in the distant past I used to use LESSCHARSET but > that is now redundant. > > I now wonder if you are using a non-unicode console font ? My console file: # Begin /etc/sysconfig/console UNICODE="1" KEYMAP="no-latin1" KEYMAP_CORRECTIONS="euro2" LEGACY_CHARSET="iso-8859-15" FONT="LatArCyrHeb-16 -m 8859-15" # End /etc/sysconfig/console # Begin /etc/profile export LANG=en_US.UTF-8 # End /etc/profile - inputrc exactly as in the book - rc.site has no console parameters set Well, I've decided to look out of the box and changed keyboard from usb to ps/2 type - no difference. However when I tried to boot from exactly the same HDD on another motherboard from different vendor - everything works fine! It doesn't work on recent Atom D2550 motherboard and works on old D510 and on HP desktop with Core2. New Atom board has Intel ICH10R chipset and NUVOTON NTC6776D SuperIO, while old one has Intel ICH9R chipset and Winbond 83627DHG-P SuperIO. Does it mean the problem might be with some kernel drivers? /alexey -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Thu, Dec 05, 2013 at 03:06:01PM +0100, Alexey Orishko wrote: > On Wed, Dec 4, 2013 at 7:27 PM, Ken Moffat wrote: > > > > In my previous reply I included an invalid-unicode symbol, code > > U+FFFD displaying as reverse-video question mark : � : if you read > > the mail on the LFS-7.4 machine (or copy this paragraph to a file, > > scp that file to the LFS machine, and then read it), does the > > resulting glyph appear correctly, and if not, does it match the > > white squares you had [ you need to revert the LANG and UNICODE > > settings ] ? > > I can see reverse-video question mark on LFS box, but reverse area is square > in firefox/gmail and oval in lfs (I assume due to different font) > Good, you are displaying UTF-8. That unfortunately doesn't help explain why you were getting white squares after deleting. Yes, different fonts display things differently (e.g. serif and sans-serif). > > However when I tried to boot from exactly the same HDD on another > motherboard from different vendor - everything works fine! It doesn't > work on recent Atom D2550 motherboard and works on old D510 and > on HP desktop with Core2. New Atom board has Intel ICH10R chipset > and NUVOTON NTC6776D SuperIO, while old one has Intel ICH9R > chipset and Winbond 83627DHG-P SuperIO. > > Does it mean the problem might be with some kernel drivers? > > /alexey This gets weirder and weirder. Most of us don't have a stock of different motherboards to play with, and I've no experience in this area. I think you said that showkey and dumpkeys reported the keyboard was giving the expected values for the BS and DEL keys even though they were giving white squares when you deleted, so in theory the kernel drivers are ok. Perhaps make a list of the values for _every_ key you use (with showkey) with this drive on a "good" motherboard, and then put it back on the bad motherboard and repeat the exercise just in case one or more _other_ keys are somehow corrupted. To be absolutely clear - you haven't installed xorg on the LFS-7.4 system, this is all about a problem in the console terminals (the ttys), right ? -- das eine Mal als Tragödie, dieses Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Thu, Dec 5, 2013 at 4:40 PM, Ken Moffat wrote: >> > This gets weirder and weirder. Most of us don't have a stock of > different motherboards to play with, and I've no experience in this > area. I think you said that showkey and dumpkeys reported the > keyboard was giving the expected values for the BS and DEL keys > even though they were giving white squares when you deleted, so in > theory the kernel drivers are ok. > > Perhaps make a list of the values for _every_ key you use (with > showkey) with this drive on a "good" motherboard, and then put it > back on the bad motherboard and repeat the exercise just in case one > or more _other_ keys are somehow corrupted. Well, I only need to type english letter "k" and space to get into problems.. While changing fonts I've noticed that white square boxes have inverted "y" with two dots over it.. Anyway, they appear on empty position of the command line (where no text was typed in). > To be absolutely clear - you haven't installed xorg on the LFS-7.4 > system, this is all about a problem in the console terminals (the > ttys), right ? I don't have any X-related stuff, just plain LFS and I login on local console. On other HDD disk I've added a few BLFS packages with profile.d, dircolors, sudo, mc, etc. to support non-root users. Both plain LFS and one with some BLFS packages works ok on the old motherboard. /alexey -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-support] du on / produces errors from /proc
Does anybody know what causes the following:- du: cannot access '/proc/602/task/602/fd/4': No such file or directory du: cannot access '/proc/602/task/602/fdinfo/4': No such file or directory du: cannot access '/proc/602/fd/4': No such file or directory du: cannot access '/proc/602/fdinfo/4': No such file or directory Can they be ignored? I don't appear to have a problem. Richard -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-support] Troubles Building NSS-3.15.3
This build has failed three times for me. Here is the applicable portion of the build log: rm: cannot remove 'Linux3.8_x86_64_glibc_PTH_64_OPT.OBJ': Is a directory Assembler messages: Fatal error: can't create Linux3.8_x86_64_glibc_PTH_64_OPT.OBJ/pk11cert.o: No such file or directory pk11cert.c: In function 'pk11_fastCert': pk11cert.c:231:14: warning: variable 'status' set but not used [-Wunused-but-set-variable] PRStatus status; ^ pk11cert.c: In function 'PK11_MakeCertFromHandle': pk11cert.c:297:12: warning: variable 'swapNickname' set but not used [-Wunused-but-set-variable] PRBool swapNickname = PR_FALSE; ^ pk11cert.c:296:12: warning: variable 'isFortezzaRootCA' set but not used [-Wunused-but-set-variable] PRBool isFortezzaRootCA = PR_FALSE; ^ pk11cert.c: In function 'PK11_TraverseCertsForNicknameInSlot': pk11cert.c:2007:28: warning: variable 'pk11cb' set but not used [-Wunused-but-set-variable] struct nss3_cert_cbstr pk11cb; ^ pk11cert.c: In fupk11cert.c:2151:21: warning: unused variable 'td' [-Wunused-variable] NSSTrustDomain *td = STAN_GetDefaultTrustDomain(); ^ make[2]: *** [Linux3.8_x86_64_glibc_PTH_64_OPT.OBJ/pk11cert.o] Error 2 make[2]: *** Waiting for unfinished jobs pk11load.c: In function 'SECMOD_UnloadModule': pk11load.c:569:28: warning: unused variable 'status' [-Wunused-variable] PRStatus status = PR_UnloadLibrary(softokenLib); ^ pk11merge.c: In function 'pk11_mergeByObjectIDs': pk11merge.c:852:15: warning: 'error' may be used uninitialized in this function [-Wmaybe-uninitialized] PORT_SetError(error); ^ pk11merge.c:753:9: note: 'error' was declared here int error; ^ make[1]: *** [libs] Error 2 make: *** [libs] Error 2 nction 'PK11_FindCertFromDERCertItem': Each failure involved not being able to remove the directory "Linux3.8_x86_64_glibc_PTH_64_OPT.OBJ" It appears that if the "build" can't remove the directory it thinks it can't create the file. I'm going to stick my nose into the Makefile and see if I can find what's happening. I just wanted to see if anyone else had noticed this and/or if anyone knew any solutions. Thanks, Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-support] Building Thunderbird
The first thing is that the build bailed out on me because I didn't have gstreamer installed. That's not listed in any of the dependencies. I just wanted to get a working Thunderbird so that I wouldn't have to keep rebooting to get to my e-mail. I have it, but I want it built with all the system dependencies and right now it's not. My troubles with NSS are documented in another thread. But I'm going to build it again, which leads to my next observation. In preparing for Firefox, the book, BLFS SVN, gives the procedure for linking it agains XULrunner. (And I hope this isn't a mis-print, but that the build time for Firefox drops from 55 to .3 SBU's. Fantastic.). The page for Thunderbird says nothing about XULrunner. I'm just making the assumption, and I would like someone to verify it, that since Thunderbird also employs the gekko-engine it will also link against XULrunner. Is this correct? I want to minimize the number of 50+ SBU builds. :-) Thanks, Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] du on / produces errors from /proc
Le 05/12/2013 18:18, Richard Melville a écrit : > Does anybody know what causes the following:- > > du: cannot access '/proc/602/task/602/fd/4': No such file or directory > du: cannot access '/proc/602/task/602/fdinfo/4': No such file or directory > du: cannot access '/proc/602/fd/4': No such file or directory > du: cannot access '/proc/602/fdinfo/4': No such file or directory > > What were you doing? Building some package? What is the command which generated that output? Regards Pierre -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] Building Thunderbird
On 5.12.2013 18:31, Dan McGhee wrote: > The first thing is that the build bailed out on me because I didn't have > gstreamer installed. That's not listed in any of the dependencies. > > I just wanted to get a working Thunderbird so that I wouldn't have to > keep rebooting to get to my e-mail. I have it, but I want it built with > all the system dependencies and right now it's not. My troubles with > NSS are documented in another thread. > > But I'm going to build it again, which leads to my next observation. > > In preparing for Firefox, the book, BLFS SVN, gives the procedure for > linking it agains XULrunner. (And I hope this isn't a mis-print, but > that the build time for Firefox drops from 55 to .3 SBU's. > Fantastic.). The page for Thunderbird says nothing about XULrunner. > I'm just making the assumption, and I would like someone to verify it, > that since Thunderbird also employs the gekko-engine it will also link > against XULrunner. Is this correct? > > I want to minimize the number of 50+ SBU builds. :-) > > Thanks, > Dan > You can't link thunderbird to an external version of xulrunner. First of all, versions differ and second, I believe thunderbird uses some custom modifications or addition to the engine itself. -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] Building Thunderbird
On 12/05/2013 01:20 PM, Armin K. wrote: > On 5.12.2013 18:31, Dan McGhee wrote: >> The first thing is that the build bailed out on me because I didn't have >> gstreamer installed. That's not listed in any of the dependencies. >> >> I just wanted to get a working Thunderbird so that I wouldn't have to >> keep rebooting to get to my e-mail. I have it, but I want it built with >> all the system dependencies and right now it's not. My troubles with >> NSS are documented in another thread. >> >> But I'm going to build it again, which leads to my next observation. >> >> In preparing for Firefox, the book, BLFS SVN, gives the procedure for >> linking it agains XULrunner. (And I hope this isn't a mis-print, but >> that the build time for Firefox drops from 55 to .3 SBU's. >> Fantastic.). The page for Thunderbird says nothing about XULrunner. >> I'm just making the assumption, and I would like someone to verify it, >> that since Thunderbird also employs the gekko-engine it will also link >> against XULrunner. Is this correct? >> >> I want to minimize the number of 50+ SBU builds. :-) >> >> Thanks, >> Dan >> > You can't link thunderbird to an external version of xulrunner. First of > all, versions differ and second, I believe thunderbird uses some custom > modifications or addition to the engine itself. Thanks, Armin. You saved me 55 SBU's. :-) Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] Building Thunderbird
On 12/05/2013 02:22 PM, Dan McGhee wrote: > On 12/05/2013 01:20 PM, Armin K. wrote: >> On 5.12.2013 18:31, Dan McGhee wrote: >>> >>> In preparing for Firefox, the book, BLFS SVN, gives the procedure for >>> linking it agains XULrunner. (And I hope this isn't a mis-print, but >>> that the build time for Firefox drops from 55 to .3 SBU's. >>> Fantastic.). The page for Thunderbird says nothing about XULrunner. >>> I'm just making the assumption, and I would like someone to verify it, >>> that since Thunderbird also employs the gekko-engine it will also link >>> against XULrunner. Is this correct? >>> >>> I want to minimize the number of 50+ SBU builds. :-) >>> >>> Thanks, >>> Dan >>> >> You can't link thunderbird to an external version of xulrunner. First of >> all, versions differ and second, I believe thunderbird uses some custom >> modifications or addition to the engine itself. > Thanks, Armin. You saved me 55 SBU's. :-) > > Dan > Not sure what you mean by that, if you're compiling both Firefox and Thunderbird you'd still be building xulrunner twice regardless. -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Thu, Dec 05, 2013 at 05:43:28PM +0100, Alexey Orishko wrote: > On Thu, Dec 5, 2013 at 4:40 PM, Ken Moffat wrote: > >> > > This gets weirder and weirder. Most of us don't have a stock of > > different motherboards to play with, and I've no experience in this > > area. I think you said that showkey and dumpkeys reported the > > keyboard was giving the expected values for the BS and DEL keys > > even though they were giving white squares when you deleted, so in > > theory the kernel drivers are ok. > > > > Perhaps make a list of the values for _every_ key you use (with > > showkey) with this drive on a "good" motherboard, and then put it > > back on the bad motherboard and repeat the exercise just in case one > > or more _other_ keys are somehow corrupted. > > Well, I only need to type english letter "k" and space to get into problems.. > While changing fonts I've noticed that white square boxes have inverted "y" > with two dots over it.. Anyway, they appear on empty position of the > command line (where no text was typed in). > From memory, you said that k was involved when you specified that you were NOT using unicode in the setup, so I'm inclined to ignore that. I thought your problem was that *deleting* characters on the commandline caused spurious garbage to appear after it ? If that is wrong, ignore the rest of this reply and instead please restate the problem. I still think my suggestion of checking every key will identify if any of the keys are being garbled in the kernel, and therefore if this is a kernel _input_ problem. But here is an alternative train of thought : When you trigger the garbage, is it really there, or only in the screen output ? Start a command with '#' so that it is treated as a comment, type, do whatever you need to so that the spurious garbage appears, then press enter. And make a written note of exactly what you typed, and a guide to what appeared and where it was on the line, so that you can repeat the test. Then key enter again : does the garbage scroll up, so that the current line is empty apart from the prompt ? Then key the up-arrow once to see the last line of history - is the garbage there in the history ? If the garbage is only on the screen then I guess it _is_ a kernel problem, but in the _video_ driver (i915 I suppose, but maybe you compiled in vga). Conversely, if the garbage appears in the history then ignore the rest of this mail. Your LatArCyrHeb font has two, possibly three, characters which match a "white square". If you manage to prove that it is a kernel video problem, can you please "loadkeys LatGrkCyr-8x16" and repeat the test. I'm guessing you will get a capital T with a comma below it (either U+021A or, in that font, U+0162), or a capital U with an acute accent (U+00DA), or else a capital U with a breve (U+016C). Knowing what appears might be useful if this is a bug in a video driver. I'll attach the text listing of that font - I'm guessing it is one of character positions 210, 212, or 216. I base that paragraph on what 'showconsolefont' shows me. If you get a different character, or more than one, please attempt to identify them against that listing. You mentioned a different character appeared when you tried other fonts, but there is no rule about where any particular character is stored _within_ a console font. If it is a video-driver problem, please attach the relevant part of your kernel config ("Graphics support", "I2C encoder or helper chips" (in mine that is actually Direct Rendering Manager, possibly because the I2C stuff is disabled and DRM doesn't have a heading), "Frame buffer hardware drivers" and "Console display driver support" - cut at "CONFIG_SOUND" which is the start of the sound drivers - they don't have a neat heading comment. And, which kernel is this ? At this point (for a video problem) you should have a reliable and simple testcase to show if the system has a problem. Either I will suggest that you change the .config (it is unfortunately very easy to create a config which doesn't work properly) or else to try building older and newer released kernels. ĸen -- das eine Mal als Tragödie, dieses Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Thu, Dec 05, 2013 at 10:01:00PM +, Ken Moffat wrote: > > Your LatArCyrHeb font has two, possibly three, characters which > match a "white square". If you manage to prove that it is a kernel > video problem, can you please "loadkeys LatGrkCyr-8x16" and repeat > the test. I'm guessing you will get a capital T with a comma below > it (either U+021A or, in that font, U+0162), or a capital U with an > acute accent (U+00DA), or else a capital U with a breve (U+016C). > Knowing what appears might be useful if this is a bug in a video > driver. I'll attach the text listing of that font - I'm guessing it > is one of character positions 210, 212, or 216. > Forgot to attach it the first time, and when I did it bounced (too big, whoops!). So here's the third attempt, using xz to compress it from 177K to < 7K. ĸen -- das eine Mal als Tragödie, dieses Mal als Farce LatGrkCyr-8x16.psfu.txt.xz Description: Binary data -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] problem using BS or DEL key
On Fri, Dec 6, 2013 at 1:04 AM, Ken Moffat wrote: > Forgot to attach it the first time, and when I did it bounced (too > big, whoops!). So here's the third attempt, using xz to compress it > from 177K to < 7K. Thanks and sorry for confusion! I had no Unicode initially, but added it after initial comments. However adding Unicode font didn't help. I'll try your suggestions tomorrow. Anyway I've already found troubling news re. my Intel Atom GPU: Intel D2550 Cedarview (GMA3600) has PowerVR 545 graphics core. And its problem is described here: https://gist.github.com/Aissen/2925633 Now I understand why I can't get grub and kernel to get a higher resolution... At least it explains a difference in behaviour between 2 different Atom motherboards (old one has D510). Found some suggestion online: - Upstream kernel from 3.2 should have cedarview support with 'gma500' driver maintained by Alan Cox. It provides basic KMS support for LVDS/VGA/HDMI but without graphics acceleration. - make sure CONFIG_DRM_GMA500 and CONFIG_DRM_GMA3600 is on in kernel config. I use LFS-7.4 with kernel 3.10.10 and I've tried 3.10.22 but apparently with wrong driver. /alexey -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] Building Thunderbird
On 12/5/2013 12:31 PM, Dan McGhee wrote: > The first thing is that the build bailed out on me because I didn't have > gstreamer installed. That's not listed in any of the dependencies. > > I just wanted to get a working Thunderbird so that I wouldn't have to > keep rebooting to get to my e-mail. I have it, but I want it built with > all the system dependencies and right now it's not. My troubles with > NSS are documented in another thread. > > But I'm going to build it again, which leads to my next observation. > > In preparing for Firefox, the book, BLFS SVN, gives the procedure for > linking it agains XULrunner. (And I hope this isn't a mis-print, but > that the build time for Firefox drops from 55 to .3 SBU's. > Fantastic.). The page for Thunderbird says nothing about XULrunner. > I'm just making the assumption, and I would like someone to verify it, > that since Thunderbird also employs the gekko-engine it will also link > against XULrunner. Is this correct? > > I want to minimize the number of 50+ SBU builds. :-) By way of encouragement, between Tuesday and Wednesday I got Xulrunner, Firefox and Thunderbird installed, using the BLFS development branch. No problems, but I was surprised at the number of dependencies. Alan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page