blockhead | the font "fixed" looks different in my devuan (as compared to a few other distros). i don't even know where to start investigating this one. i'm running devuan daedalus amd64. Can anyone please advise? | 00:19 |
---|---|---|
gnarface | blockhead: by "fixed" you mean the css2 property "font-family: fixed;" ?? if so, that's because it's a generic font family name, not a particular font, so it's just failing over to whatever fixed-width font you have installed. install the same fonts you were using on whatever other install it was and then web pages will find them. | 00:31 |
blockhead | does the "font-family" thing refer to the old-style X bitmap fonts? i ask becuase I only see that in xfontsel, and most of the fonts don't appear in xfontsel. the fixed I was using, and the one I am using now, are both anti-aliased, which i think indicates a newer-type vector font like a ttf or postscript? | 00:33 |
gnarface | blockhead: it would depend on the browser, but i would be surprised if recent builds of firefox supported anything other than true-type fonts | 00:35 |
blockhead | this is not in a browser, this is in some programs within the OS, specifically at least one window manager and in a file manager | 00:36 |
gnarface | are you using a custom xorg.conf ? | 00:36 |
* blockhead couldn't tell the difference between a font problem and bad html on today's web sites :p | 00:36 | |
blockhead | checking | 00:36 |
gnarface | it's not usually the type of thing you'd do and forget about | 00:38 |
blockhead | ok, i'll say "no" then | 00:39 |
gnarface | alright | 00:39 |
gnarface | let's do some sanity checks | 00:40 |
gnarface | try these commands: dpkg -l |grep ttf; dpkg -l |grep fonts | 00:40 |
gnarface | don't paste the result here | 00:40 |
gnarface | just tell me if you see a bunch of font packages | 00:40 |
blockhead | i do | 00:41 |
* blockhead was trying to pipe it to a file and it got weird | 00:41 | |
gnarface | stuff like "fonts-dejavu" and "ttf-bitstream-vera" ? | 00:42 |
blockhead | first time 2>&1 failed me :o | 00:42 |
gnarface | in fact, if you're missing either of those just install them right now. (current Debian default and previous Debian default) | 00:42 |
blockhead | have dejavu (my fav font for console) but not the other. Aha! thank you! | 00:43 |
gnarface | at the very least you should see a bunch of packages matching the glob pattern "fonts-*" and "ttf-*" | 00:43 |
blockhead | that i do, yes | 00:43 |
blockhead | and am installing ttf-bitstream-vera now | 00:44 |
rwp | blockhead, "fixed" in this context means fixed width font, as opposed to a proportionally spaced font like a typeset page uses. | 00:45 |
gnarface | blockhead: make sure you also have packages for libfreetype and fontconfig | 00:45 |
gnarface | blockhead: then run "dpkg-reconfigure fontconfig-config" (the missing anti-aliasing you were talking about could have multiple causes) | 00:45 |
gnarface | ^ as root, you run dpkg-reconfigure as root | 00:46 |
gnarface | just to be clear in case you didn't nkow | 00:46 |
blockhead | fontconfig and libfreetype are already in | 00:46 |
gnarface | know* | 00:46 |
blockhead | ty, that was going to be my next question :) | 00:46 |
gnarface | try the dpkg-reconfigure just in case then restart X | 00:46 |
gnarface | not sure if you have to restart X for that or not, but it won't hurt | 00:47 |
gnarface | you may just have to restart whatever program you were using | 00:47 |
rwp | In my web browser I can right click, select Inspect, pops up the web developer panel. The element should be highlighted on the left. On the far right is a Fonts tab. Clicking it shows me details of the font being used at that point. Something like Noto Sans Mono 12px 400wt and so on. | 00:48 |
blockhead | a moment please, the dkpg operation is asking me some difficult questions | 00:50 |
gnarface | blockhead: for what it's worth, i'm not seeing xfontsel anti-aliasing anything either, and that may not be a modern enough tool to really be useful for this. off the top of my head though, firefox and gimp should both render true-type fonts correctly according to your fontconfig settings, though for firefox it has its own internal defaults setting for every major font class. | 00:51 |
gnarface | (sorry i don't know of anything smaller off the top of my head but i'm sure there are things) | 00:51 |
gnarface | lxappearance maybe for some stuff | 00:52 |
gnarface | gtk2 stuff anyway | 00:52 |
gnarface | unfortunately there's not really any one central place to manage this stuff | 00:53 |
gnarface | every major version of every single toolkit does its own thing, and some programs go off on their own and other programs are using obsolete font format standards | 00:54 |
blockhead | ok, finsihed, exited to the console, typed "startx" | 00:54 |
blockhead | no change. checked in the one window manager and it's still a different fixed. Also in the one program | 00:55 |
gnarface | do you happen to know what the old font was you were using before? | 00:55 |
blockhead | anti-alasing isn't the problem, it's just a different font | 00:55 |
blockhead | yeah, fixed :p | 00:55 |
blockhead | same name as the new one | 00:55 |
gnarface | heh, not specific enough unfortunately | 00:55 |
blockhead | "fixed" | 00:55 |
gnarface | do you at least know which toolkit the window manager in question is using? | 00:56 |
blockhead | this takes me back to windows days when I had to figure out which font JIOKL8jF.ttf was :p | 00:56 |
blockhead | one sec, that i can do | 00:56 |
gnarface | and the name of the window manager itself | 00:57 |
blockhead | dwm | 00:57 |
gnarface | hmm, not one i'm familiar with | 00:57 |
blockhead | it's a tiling window manager i use sometimes | 00:57 |
gnarface | hmm, doing some searches here to try to get some context on the problem... you have to recompile it after changing config.h to change the font??? | 01:00 |
blockhead | yes, you have to recompile it to make any change | 01:00 |
gnarface | fascinating | 01:00 |
blockhead | iirc I left font at defaults and mostly fussed with bar location and colors | 01:00 |
blockhead | it's an oddball program but has its moments | 01:00 |
gnarface | well my first suggestion then would be to identify the old font and just rebuild it with a specific enough font string to force it to use that | 01:00 |
gnarface | well, after making sure that font is installed already, which might be enough without rebuilding | 01:01 |
blockhead | heh, the source has "monospace:size=10" which i guess is not a help | 01:01 |
gnarface | afaik monospace is just another name for fixed | 01:01 |
blockhead | that sounds right | 01:02 |
* blockhead ponders coping the entire font directory over from the other distro, though i expect that will cause new problems | 01:03 | |
gnarface | no, don't do that. it'll probably work but you still won't know which font it's using and then you'll also have a bunch of untracked fonts all over your install | 01:04 |
gnarface | the better approach would be to get the package list from the other distro, and scan it for font packages | 01:04 |
* blockhead was pondering, not doing :p | 01:04 | |
gnarface | was the old distro using the same version of dwm? if the version also changed, maybe the font selection behavior changed too. my search results suggest this is not unprecedented, though 2012 seems like too long ago to be relevant to this particular incident | 01:08 |
blockhead | same version | 01:09 |
gnarface | if it's the same exact version, even with patches, (like if the other install we're talking about here was Debian stable) then it is fairly safe to assume that you can just install all the same font packages and get the same result | 01:11 |
gnarface | but it wouldn't necessarily be safe to assume that you'd get the same font list by default, because some of the dependencies might have slipped a bit while they were replacing key parts of the graphical stack to avoid systemd | 01:12 |
blockhead | the other distros are void lfs and an older slackware, so that shouldn't be a problem | 01:12 |
gnarface | well, the same approach may very well still work, the only complication is the package names are probably not literal matches then | 01:13 |
blockhead | looks like I'll either have to do a long drawn out package name compare (names for the same font package differ between distros) or live with it | 01:13 |
blockhead | yeah, you nailed it | 01:13 |
gnarface | er, yea | 01:13 |
gnarface | well if you want some help with regexp i might be able to assist | 01:14 |
gnarface | maybe it would be easier to compare font file names directly then work backwards to the package names from there | 01:15 |
blockhead | that makes sense | 01:15 |
gnarface | i dunno for sure if it'll help, but it seems logical that if the fonts are coming from the same upstream sources the file names themselves might all be the same even if the installation paths and package names aren't | 01:16 |
* blockhead hopes everyone else has been logical | 01:17 | |
blockhead | gnarface: thanks for the help. i will sort this out over time | 01:23 |
gnarface | blockhead: good luck, let us know what you find otu | 01:25 |
gnarface | find out* | 01:25 |
blockhead | ty. will do | 01:27 |
systemdlete | gnarface: any idea why calls to needrestart are saying it cannot determine the microcode level? | 01:56 |
systemdlete | do I have some misconfigured (most likely possibility I think) | 01:56 |
systemdlete | s/some/something/ | 01:56 |
systemdlete | s/$/?/ | 01:56 |
gnarface | systemdlete: no, nothing comes to mind other than the vague guess that maybe you haven't installed the cpu microcode package, or you've installed a backports kernel without getting the corresponding backports microcode package version? | 01:57 |
gnarface | i've never seen that error before | 01:57 |
systemdlete | 6.1.0-15-amd64 and I just installed the amd64-microcode package | 01:59 |
systemdlete | maybe another piece is missing? | 01:59 |
gnarface | maybe they just don't have microcode in there for that cpu. it's a phenom II, you said, right? | 01:59 |
systemdlete | no, this one is a Fx8350... but hold on | 02:00 |
systemdlete | checked. It is the 8 core, so it must be the fx8350 | 02:00 |
gnarface | where are you seeing this error? i'm not sure anything on my system actually calls needrestart, since i don't use the kernel meta-package | 02:01 |
gnarface | i have the same cpu and haven't seen this happening | 02:01 |
systemdlete | when I run needrestart -p (the nagios invocation of it) | 02:02 |
gnarface | ah, hmm | 02:02 |
systemdlete | gnarface, dollars to donuts, this is a systemdlete problem, not linux | 02:02 |
gnarface | well, based on my own data there's a statistically likely possibility it's a nagios problem too | 02:03 |
systemdlete | I "built" this install using debootstrap, and removed/added packages as I felt justified | 02:03 |
systemdlete | ? | 02:03 |
gnarface | i don't have a ton of nagios experience but where i've seen it used for error reporting, most the errors it generated were errors it was causing | 02:04 |
gnarface | so maybe, consider it's just calling needrestart wrong or something like that | 02:05 |
systemdlete | I am actually using naemon, not nagios. But, anyway, needrestart is NOT part of nagios. The "-p" option is just to generate monitoring software compatibility | 02:05 |
gnarface | munin is better, but that's not saying much | 02:05 |
gnarface | (i've had to fix plenty of bugs in munin plugins myself by hand, but at least they're usually easy to fix) | 02:06 |
systemdlete | ok, I'll review my code and see if I can figure this out. | 02:06 |
gnarface | good luck, sorry i don't have better info | 02:06 |
systemdlete | needrestart is not a plugin for any monitoring system. Again, the -p option is just to generate compatible output for monitoring software. | 02:06 |
systemdlete | thanks for at least looking at this for me. | 02:06 |
systemdlete | I'll poke through my code and how I configured the host. | 02:07 |
systemdlete | Most likely operator error (me). | 02:07 |
systemdlete | ok, now. different problem on a different system, this time chimaera. apt update says: | 02:09 |
systemdlete | E: Release file for http://deb.devuan.org/merged/dists/chimaera/InRelease is expired (invalid since 53min 13s). Updates for this repository will not be applied | 02:09 |
systemdlete | I did not alter the apt config on that chimaera system since last time I ran apt update. | 02:10 |
gnarface | and the system clock is right on that one? ...someone was just earlier mentioning there was a gpg key problem on one of the mirrors, maybe the same one | 02:10 |
systemdlete | ah. | 02:10 |
systemdlete | let me check that, thank you. Yes, there have been some clock issues here. | 02:10 |
systemdlete | I will double-check that! | 02:10 |
gnarface | uh, i think this is the stats url? http://veritas.devuan.org/apt-panopticon/results/Report-web.html | 02:11 |
systemdlete | time is syncd. I checked the routers also. | 02:11 |
systemdlete | some other problem then | 02:11 |
gnarface | check the scrollback if you have it, fsmithred and debdog were talking about it this morning, about 11 hours ago | 02:11 |
systemdlete | (but thanks for reminding me--time can cause serious screwups) | 02:12 |
systemdlete | in THIS channel? or one of the other devuan-* channels? | 02:12 |
systemdlete | (I don't see any messages from either of them in this channel) | 02:12 |
gnarface | this channel, almost exactly 11 hours ago to the minute | 02:13 |
gnarface | i inferred one of the mirrors in the round-robin must have fallen out of sync, there wasn't any updates so maybe it's still out of sync | 02:13 |
systemdlete | ah, now I see it | 02:14 |
systemdlete | I just didn't search back far enough. | 02:14 |
systemdlete | Yeah. Maybe a mirror issue | 02:14 |
systemdlete | my apt sources are pointed at deb.devuan.org | 02:16 |
systemdlete | should I be using something like us.devuan.org instead? | 02:16 |
systemdlete | or just leave it alone for a few more hours and let it heal itself as it usually does? | 02:16 |
* gnarface shrugs | 02:17 | |
systemdlete | hey gnarface. | 02:18 |
systemdlete | I am using apt-cacher-ng, remember? | 02:18 |
gnarface | yea, i remember | 02:18 |
systemdlete | so... | 02:18 |
systemdlete | I did reboot that system today (the one that runs the cacher) | 02:19 |
gnarface | doesn't really change my answer... i don't know if us.devuan.org or us.deb.devuan.org or something like that is up yet, i remember there was talk of it... i think debdog basically took that approach, i would take the latter approach typically and just wait, but probably not if it stopped solving the problem one day | 02:20 |
gnarface | if you think apt-cacher-ng is a factor here you can just temporarily remove your apt config file for it and run "apt-get update" again to see for sure, but since debdog already ran into it this morning i was fairly sure you're seeing the same thing | 02:22 |
systemdlete | I tend to agree. | 02:23 |
systemdlete | I was just thinking about that, is all. I wouldn't expect it to suddenly go south, so to speak. Esp when others are having similar experience. | 02:23 |
gnarface | it's a minor issue but apt-cacher-ng might slightly delay your ability to see the mirror get repaired otherwise, but i don't think by more than a few minutes and even then only if you had tried to access it within that amount of time before the repo actually got fixed | 02:24 |
gnarface | manually flushing the apt-cacher-ng's cache without actually disabling it is also an optino | 02:24 |
gnarface | option* | 02:25 |
onefang | The stinprinza mirror that debdog was talking about before has been removed from the deb.devuan.org DNS RR, so it's unlikely to be causing that problem now. | 02:25 |
gnarface | i've done it once or twice while pressed for time, but i haven't seen any permanent failures so far based on corrupted cache | 02:26 |
gnarface | it always seems to clean itself up without help | 02:26 |
systemdlete | :) | 02:26 |
systemdlete | I am running maintenance. I checked the box for forced downloading of index files, just to be extra sure. | 02:32 |
systemdlete | during the tagging/removing, it says that devrep/merged/dists/chimaera/InRelease ERROR no such file or directory | 02:33 |
rrq | "devrep"? | 02:37 |
systemdlete | I am only seeing errors for chimaera, not daedalus or beowulf. So it must be just that release that's out of sync | 02:37 |
systemdlete | devrep == devuan repository. It's an alias defined in the cacher-ng configs. | 02:37 |
systemdlete | (I just followed the directions. I don't really know every detail.) | 02:38 |
* systemdlete needs to reboot; back shortly | 02:42 | |
plasma41 | systemdlete: I'm also seeing the error about expired InRelease for chimaera (and beowulf). | 02:45 |
rrq | yes it's from amprolla... not sure why | 02:59 |
rrq | (file created Dec 10 03:22) | 03:06 |
cousin_luigi | rwp: Thanks, I've included your changes and I'll upload my updated version in a short while. Just a question: does debian/changelog support multi-line entries? Being accustomed to rpm, I thought it did. | 06:01 |
cousin_luigi | Perhaps the style is more concise on debian? Should I omit all the files I deleted, or is the upstream changelog supposed to be kept separate? | 06:02 |
cousin_luigi | Note, this is my WiP debian/changelog http://sprunge.us/LxD6U3 which is causing errors: I feel like I should document my changes, but perhaps there's a stylistic aspect that evades me. How does one do it on debian? | 06:17 |
rwp | cousin_luigi, Where are you planning on uploading that package? I thought you just wanted it for your own local use. The upstream part of it does not comply with Policy, installs files into /usr/local/sbin, that type of thing. It would take some effort to get it into a shape suitable for general installation. | 06:46 |
rwp | cousin_luigi, Format of the changelog file: | 06:47 |
cousin_luigi | rwp: I thought of putting it on OBS for the time being, but I'm a stickler for form. | 06:47 |
rwp | Format of the changelog file.https://manpages.debian.org/testing/dpkg-dev/deb-changelog.5.en.html | 06:47 |
rwp | OBS? Open Broadcast Software? That's what I know OBS to be. | 06:48 |
cousin_luigi | rwp: openSUSE Build Service, which despite the name is not only for openSUSE or rpm in general. | 06:49 |
cousin_luigi | (and yes, many petitioners for the other OBS enter the wrong channel at times) | 06:49 |
rwp | It was building before. If it is not building now then it was a problem between point A and point B. By comparing the difference you should be able to figure it out. | 06:50 |
cousin_luigi | rwp: Oh, I'm not too worried about that, I just wish to make sure I'm going in the right direction. | 06:57 |
cousin_luigi | dpkg-source: error: can't build with source format '3.0 (quilt)': non-native package version does not contain a revision <- do I actually need to use quilt instead of native? I don't plan on using patches anytime soon | 06:57 |
cousin_luigi | rwp: https://filebin.net/99korun6qphvwin4 <- so, should I move /usr/local/sbin to /usr/sbin ? Any other recommendations? | 07:11 |
rwp | That's a tar.xz file, and I would need to download it to browse it. And I am busy with other things at the moment. Sorry. | 07:17 |
rwp | But yes, you should review the guidelines and policy about packaging. The /usr/local is left to the local admin. | 07:18 |
cousin_luigi | rwp: You're right. I should use gitlab and the like for such things. Thanks for all your help. | 07:19 |
rwp | I know from looking previously that it creates /lib/voyage-utils as an empty directory. It shouldn't do that. And it should be /usr/lib now. | 07:19 |
cousin_luigi | Right. | 07:20 |
rwp | It installs a long list of scripts with a .sh extension. Those extensions need to be removed. Those names are way to generic and should be made more namespace unique. | 07:20 |
cousin_luigi | rwp: I removed most of those | 07:20 |
cousin_luigi | Now there's only three left | 07:20 |
cousin_luigi | hmm | 07:21 |
rwp | You can't just drop commands like console, debug, fastreboot, nat, remountro, remountrw, in PATH. That's just not acceptable for general purpose distribution. | 07:21 |
rwp | autoinstall, copyfiles, really all of that entire set should be renamed to have a namespace prefix or something. | 07:22 |
rwp | For the lurkers who might peek, here is the installed file list: https://paste.debian.net/plain/1301352 | 07:22 |
rwp | Regarding native packaging, if you want to convert it back to a native packaging that is your choice. But since I wanted to hack on it and not change the original package I had to change it to a non-native packaging. And 3.0 (quilt) is the current standard of the day. You don't have to actually use quilt though. | 07:25 |
cousin_luigi | rwp: That's no longer the case. https://paste.debian.net/hidden/e802b1fd/ this is the current filelist | 07:25 |
cousin_luigi | I actually use quilt when working on rpm spec files. Perhaps I will come to also appreciate it to handle debian packages. | 07:27 |
rwp | That certainly looks a lot better. Of that list what does remountro and remountrw do? I think they should be namespaced. And voyage-rwfile.sh needs the .sh suffix removed. | 07:32 |
rwp | And those need to move from /usr/local/sbin to /usr/sbin too of course. | 07:33 |
* rwp is afk for a while | 07:33 | |
cousin_luigi | rwp: They remount the root filesystem as ro and rw, respectively. voyage-rwfile.sh should be used to repair links, I'm not 100% sure it works as it should. Full tree here https://gitlab.com/lbaldoni/voyage-util | 08:21 |
cousin_luigi | Does one even need a Makefile to install scripts? | 08:22 |
cousin_luigi | https://paste.debian.net/hidden/d4b623f6/ <- a list of my current troubles (built on devuan this time) The syntax error in changelog is the one I don't understand. | 09:17 |
fsmithred | cousin_luigi, you need to format the changelog correctly. Look at some others for examples. Even an extra or missing space between elements can cause the build to fail. | 09:25 |
fsmithred | https://www.debian.org/doc/manuals/maint-guide/dreq.en.html | 09:30 |
cousin_luigi | fsmithred: Yeah, missing leading space, I found it by trial and error. What a fool I am. | 10:10 |
fsmithred | It's not easy. There are lots of details to get right. | 10:15 |
cousin_luigi | https://paste.debian.net/hidden/174a9bce/ changelog is now ok. How do I filter systemd warnings? | 10:22 |
cousin_luigi | On opensuse I would add a rpmlintrc file | 10:22 |
fsmithred | I don't know about filtering systemd warnings. I assume you need to look at what generates the warnings. | 10:25 |
fsmithred | dpkg-gencontrol: warning: Depends field of package voyage-util: substitution variable ${shlibs:Depends} used, but is not defined | 10:27 |
fsmithred | ^^^ and fix the control file | 10:27 |
fsmithred | or maybe not. did it build ok? | 10:28 |
cousin_luigi | It builds and I'm all for fixing problems at the source, but "W: voyage-util: missing-systemd-service-for-init.d-script voyage-util [etc/init.d/voyage-util] | 10:30 |
cousin_luigi | cannot be fixed on devuan, can it | 10:30 |
cousin_luigi | ? | 10:30 |
fsmithred | is init.d-script something that turns a service file into an init script? | 10:30 |
fsmithred | you might get better answers in #devuan-dev | 10:32 |
cousin_luigi | fsmithred: No, it's ancient, probably dates back from etch | 11:23 |
cousin_luigi | lintian-explain-tags is a most useful command | 11:36 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!