Re: [blfs-support] BLFS 7.10 fonts less clear than in 7.9.
On Sat, Sep 24, 2016 at 04:01:15PM -0400, Michael Shell wrote: > On Sat, 24 Sep 2016 01:41:18 +0100 > Ken Moffatwrote: > > > I've just had the opportunity to compare /etc/fonts/fonts.conf in > > 7.9 and 7.10. Yes, they do differ in size by about 3.5K. The > > reason is that the table of valid blank characters almost at the end > > of the file has been removed. > . > . > > I'm puzzled why the absence of that table would alter things for an > > English speaker running LFS ... > > > Ken, > > Watch out - both those config files load all the links in > /etc/fonts/conf.d: > > > conf.d > > So, we've also got to compare the list of the links in /etc/fonts/conf.d as > well as what is in (or just the sizes) the config files those links are > pointing to /etc/fonts/conf.avail > > This is likely where the difference resides. > > > Cheers, > > Mike > True, a quick diff didn't highlight that. My 7.10 desktop is powered off at the moment, and those symlinks actually point to /usr/share/fontconfig/conf.avail in my case - I looked at my first-stage backups, in /staging on my server, but of course the links are broken when I do that. A (very) quick look yesterday suggested the symlinks on my 7.9 and 7.10 were the same, but I won't swear that they are (more than a screenful using ls -l). But - assuming all the 10-hinting* symlinks are present, the contents of those files DID change between BLFS-7.9 and 7.10. ĸen -- `I shall take my mountains', said Lu-Tze. `The climate will be good for them.' -- Small Gods -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
On Sep 24, 2016 6:26 PM, "Rob"wrote: > > Bruce Dubbs wrote: > Are the service files really that much easier? Look at the LFS scripts. > A typical file is a simple bash script about a page long. Half of that is > comments. Admins are used to reading bash scripts. Reading these are > comfortable and easy to understand. Do you understand what systemd is > doing with their short service files? > > And if it does not do the "right thing", which do you think is easier to fix? > > We have a command > systemctl status > which will tell you exactly why a service failed to run. At least > most of the time. > If you're not a scripter, you can still write service files > because they are basically in the style of ini files from windows. > They are a little more forgiving than shell scripts in that sense. > systemctl enable whatever > will plug in the service to start at boot; and it's a heck of > a lot easier than creating a bunch of symbolic links in > /etc/rc.d > by hand. > Debian has a tool > update-rc.d > which will do this (if you have all your numbers correct in the script > anyway) but it still has to do a lot of work. > I also like the journaling system. You can filter out > various items with command line switches passed to > journalctl > and it's much faster than combing through > /var/log/whatever > to find information. > In short, while systemd does make things a lot more convoluted, > it does have its good points, and, while I'm not completely > happy with it, I'm satisfied enough with it to keep going. > -- I was going to reply with these points, but you beat me to it. Definitely would've been a bit, needed to clear my mind and think rationally first. Thank you, Douglas R. Reno --LFS/BLFS systemd maintainer. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Bruce Dubbswrote: Are the service files really that much easier? Look at the LFS scripts. A typical file is a simple bash script about a page long. Half of that is comments. Admins are used to reading bash scripts. Reading these are comfortable and easy to understand. Do you understand what systemd is doing with their short service files? And if it does not do the "right thing", which do you think is easier to fix? We have a command systemctl status which will tell you exactly why a service failed to run. At least most of the time. If you're not a scripter, you can still write service files because they are basically in the style of ini files from windows. They are a little more forgiving than shell scripts in that sense. systemctl enable whatever will plug in the service to start at boot; and it's a heck of a lot easier than creating a bunch of symbolic links in /etc/rc.d by hand. Debian has a tool update-rc.d which will do this (if you have all your numbers correct in the script anyway) but it still has to do a lot of work. I also like the journaling system. You can filter out various items with command line switches passed to journalctl and it's much faster than combing through /var/log/whatever to find information. In short, while systemd does make things a lot more convoluted, it does have its good points, and, while I'm not completely happy with it, I'm satisfied enough with it to keep going. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Rob wrote: Bruce Dubbswrote: I agree. That's because systemd is stupid. They think everyone wants to run multiple X sessions on a system so they have this 'seats' thing. They have this penchant for making Linux harder to understand and use. Think about it. What does systemd offer you, the user? Well, I can't argue the first point. When I first saw that systemd was being implemented in Arch--my distribution at the time--I glanced over it and thought, wow, this is sure convoluted. However, I like the service files; they're a heck of a lot easier than the old fashioned init scripts. But things still are rather convoluted internally. If we could have the service file usage combined with the simplicity of the old sysv, that would be awesome. Are the service files really that much easier? Look at the LFS scripts. A typical file is a simple bash script about a page long. Half of that is comments. Admins are used to reading bash scripts. Reading these are comfortable and easy to understand. Do you understand what systemd is doing with their short service files? And if it does not do the "right thing", which do you think is easier to fix? The way systemd handles processes seems a little neater, but they introduced a whole lot of new terminology that one has to become accustomed to in order to make full use of the mechanisms. What are these process handling tools? So in that sense, they do indeed make things a little more complicated than they need to be. Ya think? Unfortunately, it is also true that most distributions are moving to systemd, so I finally gave in to the inevitable and started learning about it. This led me to build it from the ground up, here. The distros are going to systemd for the convenience of the distros, not the users. -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Bruce Dubbswrote: > I agree. That's because systemd is stupid. They think everyone wants to > run multiple X sessions on a system so they have this 'seats' thing. They > have this penchant for making Linux harder to understand and use. > Think about it. What does systemd offer you, the user? Well, I can't argue the first point. When I first saw that systemd was being implemented in Arch--my distribution at the time--I glanced over it and thought, wow, this is sure convoluted. However, I like the service files; they're a heck of a lot easier than the old fashioned init scripts. But things still are rather convoluted internally. If we could have the service file usage combined with the simplicity of the old sysv, that would be awesome. The way systemd handles processes seems a little neater, but they introduced a whole lot of new terminology that one has to become accustomed to in order to make full use of the mechanisms. So in that sense, they do indeed make things a little more complicated than they need to be. Unfortunately, it is also true that most distributions are moving to systemd, so I finally gave in to the inevitable and started learning about it. This led me to build it from the ground up, here. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd.service missing
On 09/24/16 23:10, akhiezer wrote: From: spikyDate: Sat, 24 Sep 2016 22:20:20 +0100 Subject: [blfs-support] wicd.service missing blfs-systemd-units-20160602 is missing wicd.service file Maybe related to: http://www.linuxfromscratch.org/blfs/view/stable/basicnet/wicd.html == " Command Explanations sed -e ... -i setup.py: The first expression in this sed command prevents installation of logrotate and systemd configuration files. You may omit it if you use these utilities. The second one fixes building with BLFS." ? akh -- I,m using systemd http://www.linuxfromscratch.org/blfs/view/systemd/basicnet/wicd.html -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/16 23:03, akhiezer wrote: From: spikyDate: Sat, 24 Sep 2016 19:34:39 +0100 Subject: Re: [blfs-support] wicd error . . in the email. Anyways, after running the sed command and before running configure, edit setup.py, line 321, and line up the indentation on that line with the other lines. That fixed it thanks for that. NB of course that the sed includes also the '.../wpath.logrotate\|wpath.systemd/d...' part, and not just the indentation part; so be sure that the former has been applied too. More generally, do you understand what's going on when a single commmand is written/issued across multiple lines, with '\' at the end of each line except for the last. akh -- I was using the systemd version sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ I think your last comment was for init version which are different. sed -e "/wpath.logrotate\|wpath.systemd/d" \ -e "/detection failed/ a\self.init=\'init\/default\/wicd\'" \ And no I,m not very good with sed commands -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd.service missing
> From: spiky> Date: Sat, 24 Sep 2016 22:20:20 +0100 > Subject: [blfs-support] wicd.service missing > > blfs-systemd-units-20160602 is missing wicd.service file > Maybe related to: http://www.linuxfromscratch.org/blfs/view/stable/basicnet/wicd.html == " Command Explanations sed -e ... -i setup.py: The first expression in this sed command prevents installation of logrotate and systemd configuration files. You may omit it if you use these utilities. The second one fixes building with BLFS." ? akh -- -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Rob wrote: Bruce Dubbswrote: Make sure the PATH is set properly so xauth is found. On debugging technique is to edit /opt/xorg/bin/startx. At the bottom is the line: xinit "$client" $clientargs -- "$server" $display $serverargs copy that line and on the first copy add echo to the front to see what it is really executing. It may also be useful to redirect the output of that echo to something like /tmp/xinit.debug. If you followed the BLFS instructions, the $display variable should be vt7. I have no such line in my startx file. For ease of use, I elected to install in the regular /usr prefix instead of /opt one. Since this is going on my production system, I'm building for stability. I reinstalled xinit without the seds applied on the page in BLFS. I now have an X session. Apparently, telling it to run on tty7 doesn't work. I agree. That's because systemd is stupid. They think everyone wants to run multiple X sessions on a system so they have this 'seats' thing. They have this penchant for making Linux harder to understand and use. Think about it. What does systemd offer you, the user? -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
> From: spiky> Date: Sat, 24 Sep 2016 19:34:39 +0100 > Subject: Re: [blfs-support] wicd error > . . > > in the email. Anyways, after running the sed command and before > > running configure, edit setup.py, line 321, and line up the > > indentation on that line with the other lines. > That fixed it thanks for that. NB of course that the sed includes also the '.../wpath.logrotate\|wpath.systemd/d...' part, and not just the indentation part; so be sure that the former has been applied too. More generally, do you understand what's going on when a single commmand is written/issued across multiple lines, with '\' at the end of each line except for the last. akh -- -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Bruce Dubbswrote: > Make sure the PATH is set properly so xauth is found. On debugging > technique is to edit /opt/xorg/bin/startx. At the bottom is the line: > xinit "$client" $clientargs -- "$server" $display $serverargs > copy that line and on the first copy add echo to the front to see what it > is really executing. It may also be useful to redirect the output of that > echo to something like /tmp/xinit.debug. If you followed the BLFS > instructions, the $display variable should be vt7. I have no such line in my startx file. For ease of use, I elected to install in the regular /usr prefix instead of /opt one. Since this is going on my production system, I'm building for stability. I reinstalled xinit without the seds applied on the page in BLFS. I now have an X session. Apparently, telling it to run on tty7 doesn't work. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Thanos Baloukaswrote: > Please disregard what I said about systemd setting the tty variable, > it's not true. Does startx work if you specify the vt number? the startx command said that -vt was an unrecognized option. Putting a line in ~/.xserverrc didn't work either; it still tried to open /dev/tty0 and said permission denied. Contents of ~/.xserverrc #!/bin/bash exec /usr/bin/Xorg -nolisten tcp "$@" vt$XDG_VTNR -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-support] wicd.service missing
blfs-systemd-units-20160602 is missing wicd.service file -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] BLFS 7.10 fonts less clear than in 7.9.
On Sat, 24 Sep 2016 01:41:18 +0100 Ken Moffatwrote: > I've just had the opportunity to compare /etc/fonts/fonts.conf in > 7.9 and 7.10. Yes, they do differ in size by about 3.5K. The > reason is that the table of valid blank characters almost at the end > of the file has been removed. . . > I'm puzzled why the absence of that table would alter things for an > English speaker running LFS ... Ken, Watch out - both those config files load all the links in /etc/fonts/conf.d: conf.d So, we've also got to compare the list of the links in /etc/fonts/conf.d as well as what is in (or just the sizes) the config files those links are pointing to /etc/fonts/conf.avail This is likely where the difference resides. Cheers, Mike -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/16 19:26, Paul Hentschel wrote: On 09/24/2016 02:21 PM, Paul Hentschel wrote: On 09/24/2016 02:15 PM, Paul Hentschel wrote: On 09/24/2016 01:53 PM, spiky wrote: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Try this and tell me if it works: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul Hmm, the extra spaces I added are not in the reply I sent. I added 15 spaces after the a\ in the sed to make it work for a total of 16 spaces. Here it is again: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul It appears the extra spaces I add to that line keep getting stripped in the email. Anyways, after running the sed command and before running configure, edit setup.py, line 321, and line up the indentation on that line with the other lines. That fixed it thanks for that. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/16 19:15, Paul Hentschel wrote: On 09/24/2016 01:53 PM, spiky wrote: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Try this and tell me if it works: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul nope same error wicd-1.7.4$ sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ > -i.orig setup.py && > > rm po/*.po && > > python setup.py configure --no-install-kde \ > --no-install-acpi\ > --no-install-pmutils \ > --no-install-init\ > --no-install-gnome-shell-extensions \ > --docdir=/usr/share/doc/wicd-1.7.4 File "setup.py", line 321 self.init='init/default/wicd' ^ IndentationError: unindent does not match any outer indentation level Even with the 2nd one you sent -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/2016 02:15 PM, Paul Hentschel wrote: On 09/24/2016 01:53 PM, spiky wrote: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Try this and tell me if it works: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul Hmm, the extra spaces I added are not in the reply I sent. I added 15 spaces after the a\ in the sed to make it work for a total of 16 spaces. Here it is again: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/2016 01:53 PM, spiky wrote: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Try this and tell me if it works: sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi\ --no-install-pmutils \ --no-install-init\ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 Paul -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/16 18:19, akhiezer wrote: From blfs-support-boun...@lists.linuxfromscratch.org Sat Sep 24 17:55:25 2016 To: blfs-support@lists.linuxfromscratch.org From: spikyDate: Sat, 24 Sep 2016 18:15:36 +0100 Subject: Re: [blfs-support] wicd error On 09/24/16 17:31, akhiezer wrote: From: spiky Date: Sat, 24 Sep 2016 17:44:55 +0100 Subject: [blfs-support] wicd error Trying to build wicd I get an error, I think with the sed |sed "/detection failed/ a\self.init=\'init\/default\/wicd\'" \> -i.orig Why have you got an '>' in there - it's wrong: Ref: http://www.linuxfromscratch.org/blfs/view/stable/basicnet/wicd.html == sed -e "/wpath.logrotate\|wpath.systemd/d" \ -e "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py (You need to be able to self-detect such things.) setup.py File "setup.py", line 321 self.init='init/default/wicd' ^ IndentationError: unindent does not match any outer indentation level| . . -- Sorry the > was miscopied, it,s the return in the terminal new line -- Issue the sed command properly; the '\' at the end of lines 1 & 2, should be followed immediately by a newline/return - as you'll have encountered many times thru lfs then blfs, if you are now at wicd. akh - so if I copy and paste the command from the book as with all others I have done it fails. sed "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py && rm po/*.po && python setup.py configure --no-install-kde \ --no-install-acpi \ --no-install-pmutils \ --no-install-init \ --no-install-gnome-shell-extensions \ --docdir=/usr/share/doc/wicd-1.7.4 |File "setup.py", line 321 self.init='init/default/wicd' ^ IndentationError: unindent does not match any outer indentation level Sorry if I,m not grasping something, but I have not had issues with other commands. | -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
> From blfs-support-boun...@lists.linuxfromscratch.org Sat Sep 24 17:55:25 2016 > To: blfs-support@lists.linuxfromscratch.org > From: spiky> Date: Sat, 24 Sep 2016 18:15:36 +0100 > Subject: Re: [blfs-support] wicd error > > > > On 09/24/16 17:31, akhiezer wrote: > >> From: spiky > >> Date: Sat, 24 Sep 2016 17:44:55 +0100 > >> Subject: [blfs-support] wicd error > >> > >> Trying to build wicd I get an error, I think with the sed > >> > >> > >> |sed "/detection failed/ a\self.init=\'init\/default\/wicd\'" \> -i.orig > > > > Why have you got an '>' in there - it's wrong: > > > > Ref: http://www.linuxfromscratch.org/blfs/view/stable/basicnet/wicd.html > > == > > sed -e "/wpath.logrotate\|wpath.systemd/d" \ > > -e "/detection failed/ a\ > > self.init=\'init\/default\/wicd\'" \ > > -i.orig setup.py > > > > > > > > (You need to be able to self-detect such things.) > > > > > >> setup.py File "setup.py", line 321 self.init='init/default/wicd' ^ > >> IndentationError: unindent does not match any outer indentation level| > >> . . > > -- > Sorry the > was miscopied, it,s the return in the terminal new line > -- Issue the sed command properly; the '\' at the end of lines 1 & 2, should be followed immediately by a newline/return - as you'll have encountered many times thru lfs then blfs, if you are now at wicd. akh -- -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] wicd error
On 09/24/16 17:31, akhiezer wrote: From: spikyDate: Sat, 24 Sep 2016 17:44:55 +0100 Subject: [blfs-support] wicd error Trying to build wicd I get an error, I think with the sed |sed "/detection failed/ a\self.init=\'init\/default\/wicd\'" \> -i.orig Why have you got an '>' in there - it's wrong: Ref: http://www.linuxfromscratch.org/blfs/view/stable/basicnet/wicd.html == sed -e "/wpath.logrotate\|wpath.systemd/d" \ -e "/detection failed/ a\ self.init=\'init\/default\/wicd\'" \ -i.orig setup.py (You need to be able to self-detect such things.) setup.py File "setup.py", line 321 self.init='init/default/wicd' ^ IndentationError: unindent does not match any outer indentation level| akh -- Sorry the > was miscopied, it,s the return in the terminal new line -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Douglas R. Reno wrote: On Sat, Sep 24, 2016 at 2:34 AM, Bruce Dubbswrote: Thanos Baloukas wrote: On 23/09/2016 09:21 μμ, Rob wrote: BLFS-7.10 Systemd We're now in the section of Chapter 24 called Testing Xorg When I tried running startx I got a lot of messages, but it basically said at the end: Fatal server error: (EE) parse_vt_settings: Cannot open /dev/tty0 (Permission denied) I thought adding my normal user to the video group would have fixed that. Was I wrong? Thank you all for the help. That possibly happens because Xorg wants to start on tty0 while you login on another tty. If systemd is compiled with pam it sets the XDG_VTNR var to the tty number the login occurs, otherwise I think sets tty var to /dev/ttyn To force Xorg to start on the tty you are on try startx -- -vt To make that permanent put in ~/.xserverrc #!/bin/sh exec /usr/bin/Xorg -nolisten tcp "$@" vt<$XDG_VTNR or ${tty#/dev/tty} or a tty number> and run plain startx In xinit, we make a change so that Xorg starts up on the first unused terminal, usually vt7. See the seds on that page. I do not know if systemd modifies the startx script command or not, but I doubt it since Xorg is installed after systemd. I have never had to modify ~/.xserverrc. That file does not even exist on my systems. Make sure the PATH is set properly so xauth is found. On debugging technique is to edit /opt/xorg/bin/startx. At the bottom is the line: xinit "$client" $clientargs -- "$server" $display $serverargs copy that line and on the first copy add echo to the front to see what it is really executing. It may also be useful to redirect the output of that echo to something like /tmp/xinit.debug. If you followed the BLFS instructions, the $display variable should be vt7. The whole line will probably be something like: xinit xterm -- /opt/xorg/bin/X vt7 Note that the screen normally comes up black in twm with some xterms. You have to click the mouse sometimes to see the cursor. Also, make sure you run ldconfig so libraries can be found after installation. Has anyone actually had success with those seds? They cause me nothing but problems. I've never once gotten X to launch with it in there. I've always had to remove it from my build scripts, otherwise I get a similar problem. We normally have users rebuild systemd with logind support as part of X server. I wonder if that sed causes an incompatibility there. With logind enabled, systemd controls who owns what DRI devices and what TTY's they are available on. That being said, I wonder if X doesn't have privileges to claim the TTY device. The sed only changes three characters. Ther original script has: if [ "$have_vtarg" = "no" ]; then serverargs="$serverargs $vtarg" fi The sed inserts ': "' (colon-space-hash) at the beginning of the serverargs line (on my system, line 137). This may well be a problem for systemd as there is a comment in the startx script: # When starting the defaultserver start X on the current tty to avoid # the startx session being seen as inactive: # "https://bugzilla.redhat.com/show_bug.cgi?id=806491; It looks like the sed should be marked sysv only. Douglas, if you do that, please fix the note that refers to /usr/bin/startx. It should be $XORG_PREFIX/bin/startx. -- Bruce -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-support] wicd error
Trying to build wicd I get an error, I think with the sed |sed "/detection failed/ a\self.init=\'init\/default\/wicd\'" \> -i.orig setup.py File "setup.py", line 321 self.init='init/default/wicd' ^ IndentationError: unindent does not match any outer indentation level| -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] BLFS 7.10 fonts less clear than in 7.9.
On 09/23/2016 07:41 PM, Ken Moffat wrote: On Thu, Sep 22, 2016 at 02:26:02PM -0500, rhubarbpie...@gmail.com wrote: I thought fontconfig creates /etc/fonts/fonts.conf and the files differ between 7.9 and 7.10. In fact they differ significantly in size. It seems compiling fontconfig is pretty straightforward but I guess there could have been an error I missed. Replacing my BadFonts_7.10 /etc/fonts with my 7.9 version does fix the problem. I've just had the opportunity to compare /etc/fonts/fonts.conf in 7.9 and 7.10. Yes, they do differ in size by about 3.5K. The reason is that the table of valid blank characters almost at the end of the file has been removed. Previously I came across a reference to this table in the context (some years and versions ago) of one of the Source Sans fonts being treated by fontconfig as not usable for English (in that case, I think it was missing both the caret and the back-tick, which I interpreted to mean they were ostensibly present, but blank). I'm puzzled why the absence of that table would alter things for an English speaker running LFS - unless you are trying to use some uncommon font (e.g. in your browser's preferences, if the browser is where you noticed this) which used to be ignored because some other characters were blank, but is now in use. But diagnosing what happens in fontconfig seems to be painful. The link I found was, I think, in https://eev.ee/blog/2015/05/20/i-stared-into-the-fontconfig-and-the-fontconfig-stared-back-at-me/ which also talks of using fc-match and pango-view to see which font will be used for a particular codepoint, based on the font name you pass and what is installed. Unfortunately, I think the examples used a smiley or other graphic character. But perhaps the process may be useful if you have both the 7.9 and 7.10 systems available. ĸen Thank you, your confirmation is helpful. I probably should have been specific in the change in size (7.9/5464 / 7.10/2306) but assumed it was simply due to an error on my part. I have nothing but serif, sans-serif, or monospace in my browser preferences. I first noticed the font difference in my Fluxbox menu. That's what I've used as a reference when testing. I still assume the problem is something I've done wrong in compiling X but just don't see it. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
Thanos Baloukaswrote: > That possibly happens because Xorg wants to start on tty0 > while you login on another tty. If systemd is compiled with > pam it sets the XDG_VTNR var to the tty number the login occurs, > otherwise I think sets tty var to /dev/ttyn Something else I also notice Whenever I switch to a new console, I see a message that says console not in use for a few seconds. And then the regular login prompt appears. Could this have something to do with it? -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
On Sat, Sep 24, 2016 at 2:34 AM, Bruce Dubbswrote: > Thanos Baloukas wrote: > >> On 23/09/2016 09:21 μμ, Rob wrote: >> >>> BLFS-7.10 Systemd >>> We're now in the section of Chapter 24 called >>> Testing Xorg >>> When I tried running >>> startx >>> I got a lot of messages, but it basically said at the end: >>> Fatal server error: >>> (EE) parse_vt_settings: Cannot open /dev/tty0 (Permission denied) >>> I thought adding my normal user to the >>> video >>> group would have fixed that. Was I wrong? >>> Thank you all for the help. >>> >>> That possibly happens because Xorg wants to start on tty0 >> while you login on another tty. If systemd is compiled with >> pam it sets the XDG_VTNR var to the tty number the login occurs, >> otherwise I think sets tty var to /dev/ttyn >> >> To force Xorg to start on the tty you are on try >> >> startx -- -vt >> >> To make that permanent put in ~/.xserverrc >> >> #!/bin/sh >> >> exec /usr/bin/Xorg -nolisten tcp "$@" vt<$XDG_VTNR or ${tty#/dev/tty} or a >> tty number> >> >> and run plain startx >> > > In xinit, we make a change so that Xorg starts up on the first unused > terminal, usually vt7. See the seds on that page. > > I do not know if systemd modifies the startx script command or not, but I > doubt it since Xorg is installed after systemd. > > I have never had to modify ~/.xserverrc. That file does not even exist on > my systems. > > Make sure the PATH is set properly so xauth is found. On debugging > technique is to edit /opt/xorg/bin/startx. At the bottom is the line: > > xinit "$client" $clientargs -- "$server" $display $serverargs > > copy that line and on the first copy add echo to the front to see what it > is really executing. It may also be useful to redirect the output of that > echo to something like /tmp/xinit.debug. If you followed the BLFS > instructions, the $display variable should be vt7. > > The whole line will probably be something like: > > xinit xterm -- /opt/xorg/bin/X vt7 > > Note that the screen normally comes up black in twm with some xterms. You > have to click the mouse sometimes to see the cursor. Also, make sure you > run ldconfig so libraries can be found after installation. > > -- Bruce > > Has anyone actually had success with those seds? They cause me nothing but problems. I've never once gotten X to launch with it in there. I've always had to remove it from my build scripts, otherwise I get a similar problem. We normally have users rebuild systemd with logind support as part of X server. I wonder if that sed causes an incompatibility there. With logind enabled, systemd controls who owns what DRI devices and what TTY's they are available on. That being said, I wonder if X doesn't have privileges to claim the TTY device. -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-support] startx failes to work
On 23/09/2016 09:21 μμ, Rob wrote: BLFS-7.10 Systemd We're now in the section of Chapter 24 called Testing Xorg When I tried running startx I got a lot of messages, but it basically said at the end: Fatal server error: (EE) parse_vt_settings: Cannot open /dev/tty0 (Permission denied) I thought adding my normal user to the video group would have fixed that. Was I wrong? Thank you all for the help. That possibly happens because Xorg wants to start on tty0 while you login on another tty. If systemd is compiled with pam it sets the XDG_VTNR var to the tty number the login occurs, otherwise I think sets tty var to /dev/ttyn To force Xorg to start on the tty you are on try startx -- -vt To make that permanent put in ~/.xserverrc #!/bin/sh exec /usr/bin/Xorg -nolisten tcp "$@" vt<$XDG_VTNR or ${tty#/dev/tty} or a tty number> and run plain startx -- Thanos -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page