#archlinux32 | Logs for 2021-06-03
Back
[00:41:35] -!- juanfra has joined #archlinux32
[02:41:42] -!- peterhil has joined #archlinux32
[05:33:04] -!- peterhil has quit [Quit: Must not waste too much time here...]
[05:46:43] -!- peterhil has joined #archlinux32
[06:41:27] -!- volki57 has quit [Quit: ZNC 1.8.2 - https://znc.in]
[06:43:14] -!- volki57 has joined #archlinux32
[07:38:43] -!- drathir_tor has quit [Ping timeout: 252 seconds]
[07:40:37] -!- drathir_tor has joined #archlinux32
[07:55:26] -!- abaumann has joined #archlinux32
[07:55:26] <buildmaster> Hi abaumann!
[07:55:26] <buildmaster> !rq abaumann
[07:55:27] <phrik> buildmaster: <abaumann> I have a theory: Linux decided they want to take over the Windows market, but this is of course only possible by getting the same quality and bugs as Windows has.. ;-)
[08:11:02] <abaumann> /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by /usr/lib/seamonkey/libmozsandbox.so)
[08:11:14] <abaumann> yeah, seamonkey, and it doesn't build upstream either.
[08:11:26] <abaumann> So, everything we don't rebuild now, will end up with glibc issues.
[08:11:56] <abaumann> In the past this seemed to have worked better, I can remember one incompatible ABI change in 10 years (C++ string)
[08:16:19] <abaumann> (everything == everything written in C++, that is)
[09:08:59] <abaumann> oh.. removing gcc10 also makes it impossible to build palemoon.. well..
[09:09:30] <abaumann> otoh software which needs an old compiler and is not adapted will not exist for long maybe anyway..
[09:14:59] <abaumann> mmh. pressing ctrl-c in a chroot doesn't work anymore?
[09:15:10] -!- deep42thought has joined #archlinux32
[09:15:11] <buildmaster> Hi deep42thought!
[09:15:11] <buildmaster> !rq deep42thought
[09:15:12] <phrik> buildmaster: <deep42thought> I have the impression, their main operating area is television via cable
[09:15:16] <abaumann> morning deep42thought
[09:15:18] <deep42thought> I don't see, why it shouldn't
[09:15:20] <deep42thought> hi abaumann!
[09:15:28] <abaumann> well, it doesn't.. :-)
[09:15:32] <abaumann> at least not for me..
[09:15:58] <deep42thought> which command do you try to abort?
[09:16:03] <abaumann> make
[09:16:10] <deep42thought> executing what?
[09:16:22] <abaumann> a GNU make
[09:16:28] <deep42thought> I would rather blame the executed command, than the chroot
[09:16:39] <deep42thought> e.g. if it catches the signal
[09:16:51] <abaumann> we'll yes. might be the funny mozilla build environment's fault..
[09:16:58] <deep42thought> :-D
[09:17:17] <abaumann> it also just simply rebuilds everything.. very handy if you try to find a bug..
[09:17:43] <abaumann> it's almost the defintion of any build system.. it's just utterly awful..
[09:18:07] <deep42thought> you mean, it rebuilds everything *everytime*?
[09:18:12] <abaumann> yes
[09:18:16] <deep42thought> uargh
[09:18:36] <abaumann> I'm running out of maintainble browsers on Archlinux32 :-)
[09:18:41] <abaumann> almost everything is broken now..
[09:18:42] <deep42thought> `make all` does a `git clean` first?
[09:18:51] <deep42thought> :-/
[09:19:39] <abaumann> yeah. everybody complains about make.. writes new even stupider make systems.. and doesn't understand the main role of a build system, incremental builds and linking..
[09:20:00] <abaumann> it's almost an industry standard now :-)
[09:20:05] <deep42thought> I wrote my own make system a while ago, too - but I abandoned it :-D
[09:20:14] <abaumann> ah? cool. :-)
[09:20:56] <abaumann> (cool, not the abandoning part - rather the "writing my own")
[09:21:00] <deep42thought> I can tell you, it was a bit mess
[09:21:04] <deep42thought> *big
[09:21:27] <abaumann> the problem with make is it is tightly coupled to shell
[09:21:28] <deep42thought> it used regexes and elaborate rules to generate targets (and recipes) from sources
[09:21:49] <abaumann> IMHO the compiler/linker itself should handle dependencies and rebuilds
[09:22:15] <abaumann> ninja is actually quite ok already, being minimal and readable.
[09:22:30] <abaumann> all meson, scons, cmake I find utterly unreadable
[09:22:46] <abaumann> a make file is a simple build instruction, not code.
[09:22:57] <abaumann> it's the domain for a domain specific language
[09:23:04] <deep42thought> ah, you mean the makefile, not the makesystem's source
[09:23:09] <abaumann> yes
[09:23:14] <abaumann> the descriptive thingy.
[09:23:26] <abaumann> make itself can be written in whatever you like best
[09:23:31] <deep42thought> rust :-D
[09:23:34] <abaumann> that is, you rather stick to C.
[09:23:34] <abaumann> :-)
[09:23:38] <deep42thought> lol
[09:23:38] <abaumann> because of bootstrapping.
[09:23:41] <abaumann> right.
[09:24:13] <abaumann> mmh. seamonkey breaks also on 64-bit, some library is newer (CLEANUP symbol). have to hunt..
[09:24:34] <abaumann> that's another disaster: browsers.
[09:25:03] <abaumann> midori is almost usable on old hardware, but has weird ideas with keyboard shortcuts (backspace=back one page, '/' enter search).
[09:25:15] <abaumann> how on earth am I supposed to enter a '/' in a URL?!
[09:25:28] <deep42thought> escape it?
[09:25:33] <abaumann> epiphany just crawls on the eeepc and is virtually unusable.
[09:25:41] <abaumann> no, hit the programmer and make him think :->
[09:25:50] <abaumann> '/' is a vi shortcut
[09:25:58] <abaumann> vi shortcuts belong into vi
[09:25:58] <deep42thought> also in less and man
[09:26:09] <abaumann> yes. that doesn't make them better..
[09:26:26] <abaumann> at least there should be a preferences page to disable it
[09:26:28] <deep42thought> maybe vi can be used as a browser?
[09:26:31] <abaumann> (couldn't find one)
[09:26:41] <abaumann> sure, and emacs _is_ a browser ;-)
[09:26:54] <deep42thought> !grab abaumann
[09:26:54] <phrik> deep42thought: Tada!
[09:26:57] <abaumann> /build/seamonkey/src/seamonkey-2.53.7.1/obj-x86_64-pc-linux-gnu/dist/include/mozilla/dom/IDBTransaction.h:197:21: error: use of undeclared identifier 'cleanup'
[09:27:09] <abaumann> /usr/include/nss/pk11hpke.h:13:5: note: expanded from macro 'CLEANUP'
[09:27:37] <abaumann> maybe a NSS incompatibility?..
[09:27:54] <abaumann> short.. I _need_ seamonkey back.
[09:28:09] <abaumann> chromium and firefox are not working at all below some gigabytes of memory.
[09:28:20] <abaumann> and I get really fed up building them as they take so long..
[09:28:37] <deep42thought> nss is up-to-date
[09:30:01] <abaumann> 29.5.2021 was an update, so I suppose seamonkey is not adapted to the new version
[09:30:23] <abaumann> netsurf: no proper javascript, try to watch a github page.. :->
[09:31:02] <abaumann> 3.65 to 3.66
[09:31:32] <abaumann> ah, an X-copy-pasting is more not-working than working..
[09:32:05] <abaumann> I don't understand that old libraries have to be maintained and cleaned up, just keep the API stable..
[09:32:17] <abaumann> it makes no sense to clean up old libraries.
[09:34:25] <abaumann> mmh. did pacman 6 already hit testing?
[09:34:31] <deep42thought> wtp pacman
[09:34:34] <deep42thought> buildmaster: wtp pacman
[09:34:35] <buildmaster> deep42thought: i486/pacman: i486/core (5.2.2-4.0), i486/staging (6.0.0-2.0)
[09:34:35] <buildmaster> i686/pacman: i686/core (5.2.2-4.0), i686/staging (6.0.0-2.0)
[09:34:35] <buildmaster> pentium4/pacman: pentium4/core (5.2.2-4.0), pentium4/staging (6.0.0-2.0)
[09:34:39] <deep42thought> no
[09:34:56] <abaumann> ah, right. (robot commands :-) )
[09:39:54] <abaumann> so, joe dropped to the aur?
[09:40:13] <abaumann> sad (though understandable)
[09:56:03] <abaumann> well, this is cool: nss defines a macro CLEANUP in pk11hpke.h which gets expanded in seamonkey-2.53.7.1/obj-x86_64-pc-linux-gnu/dist/include/mozilla/dom/IDBTransaction.h in an enum definition of Mode
[09:56:19] <abaumann> so polution of the global namespace with macros, congratulations..
[09:56:31] <deep42thought> this is a global macro?
[09:56:33] <deep42thought> great
[09:56:51] <deep42thought> I should include nss in all my projects and use *this* CLEANUP macro, then
[09:57:02] <abaumann> As I said, many people are just trying to maintain C software - without _knowing_ C
[09:57:08] <abaumann> yep.
[09:57:20] <abaumann> and it's quite natural to call something CLEANUP.
[09:57:41] <deep42thought> so we need a fix for nss, then?
[09:57:42] <abaumann> We had rules in the past: every API/library/module has to have prefixes for C global names, macros and functions.
[09:57:44] <deep42thought> or bother upstream
[09:57:57] <abaumann> otherwise you end up in a big mess - and also with that rule you end up in a big mess.
[09:58:00] <deep42thought> yes, this rule sounds sane
[09:58:07] <deep42thought> !grab abaumann
[09:58:07] <phrik> deep42thought: Tada!
[09:58:22] <abaumann> I remember a linearhash module with a prefox lh_ and a lh_insert colliding with another system library :->
[09:58:46] <deep42thought> ...
[09:58:58] <deep42thought> go the haskell-way: put a hash there!
[09:59:30] <abaumann> pk11hpke.h is a public header file, if you define macros there (most likely they are used internally only), for god sake make a private header file and include it only in the library code!
[09:59:45] <abaumann> but I always like NSS, it's a really criplped library..
[09:59:55] <abaumann> is it maintained by mozilla now by any chance? ;-)
[10:00:48] <abaumann> there are so many bugs I could report, I don't even start.. I can hardly do it without a "please learn how to program C"
[10:01:25] <abaumann> now I have a grep I cannot abort with Ctrl-C.
[10:01:32] <abaumann> So, it _is_ the chroot's fault.
[10:01:35] <deep42thought> hmm
[10:01:39] <abaumann> ah. Ctrl-Z is working
[10:01:48] <abaumann> so I can kill -9 %
[10:01:50] <abaumann> handy :-)
[10:01:59] <deep42thought> %?
[10:02:00] <abaumann> something catches SIGINT
[10:02:09] <deep42thought> thought so
[10:02:11] <abaumann> last job number
[10:02:51] <deep42thought> isn't that $! ?
[10:03:45] <abaumann> never heard of that one.
[10:03:51] <abaumann> more correct is kill -9 %1
[10:03:57] <abaumann> give it an explicit job numer
[10:03:59] <abaumann> *number
[10:04:20] <deep42thought> ah, that's a kill ting
[10:04:24] <deep42thought> *thing
[10:04:24] <abaumann> yep
[10:04:28] <deep42thought> $! is a shell thing
[10:04:36] <abaumann> ah, yes. that could be.
[10:04:43] <deep42thought> ctrl+z -> bg <enter> -> kill $!
[10:04:44] <abaumann> it's all quite confusing, actually.
[10:05:06] <deep42thought> kill's man page doesn't mention "%" :-/
[10:05:18] <abaumann> maybe an old csh/tcsh thingy?
[10:05:22] <abaumann> though I have bash
[10:05:29] <abaumann> maybe bash also emulates those?
[10:05:32] <deep42thought> I tested it, it works here, too
[10:05:37] <deep42thought> but echo'ing %1 does not work
[10:06:18] <abaumann> We'll, even if you use shella a lot, you hardly understand 10% of it's features - at least, not features I can easily remember. :-)
[10:08:42] * abaumann just run out of space on a OpenBSD machine, mounting /usr/local now over NFS (who spots the irony?)
[10:35:32] -!- bdju has quit [Ping timeout: 252 seconds]
[10:56:02] <deep42thought> it becomes truely ironic, if you share /usr/local via nfs between *multiple* machinses :-)
[10:57:35] <abaumann> well, as long they have the correct architecture..
[10:57:44] <abaumann> .. /usr/local is the installed ports on my machine
[10:58:38] <abaumann> you just mount OpenBSD_6.x/i686 or whatever accordingly
[10:58:57] <deep42thought> why are the ports installed in /usr/*local* ?
[10:59:47] -!- bdju has joined #archlinux32
[10:59:47] <abaumann> mmh. most likely because it's the default and I didn't bother to change it?
[10:59:50] * abaumann shrugs
[11:00:25] <deep42thought> it sounds funny, because /usr/local on linux is, where things that you build yourself, get installed to (by default)
[11:00:48] <abaumann> it's all just conventions :-)
[11:00:53] <deep42thought> yeah, sure
[11:01:07] <deep42thought> I wouldn't expect windows to install config files into C:\etc\, either
[11:01:09] <abaumann> we had a smart scheme in ancient times with /opt and /opt/package
[11:01:13] <abaumann> nfs automounted
[11:01:40] -!- drathir_tor has quit [Ping timeout: 252 seconds]
[11:01:42] <abaumann> then if you executed /opt/bin/screen (as symlink to /opt/screen-x.y.z/bin/screen) it would automount just that software package.
[11:02:06] <deep42thought> heh, nice
[11:02:08] <abaumann> as I understand, systemd has some automounting capability too..
[11:02:18] <deep42thought> systemd-homed?
[11:02:25] <deep42thought> or do you mean true automount?
[11:02:29] <deep42thought> that's also possible, I think
[11:02:32] <abaumann> I think, it's not just for /home
[11:02:45] <abaumann> but of course very handy for /home/user too (as soon as a user logs in)
[11:03:01] <abaumann> I just don't share the optiomism about roaming profiles..
[11:03:19] <abaumann> software developers have to play along
[11:03:30] <abaumann> it was a mere disaster on Windows.. so..
[11:04:32] <abaumann> ah, systemd can also do samba automount, nice. :-)
[11:05:07] <abaumann> just nowadays you may not want to share software packages anymore (as disks are big enough)..
[11:05:18] <abaumann> ..a complete different story is if you want to share your /home
[11:07:40] <abaumann> ..or some shared data in a group
[11:08:21] -!- drathir_tor has joined #archlinux32
[11:09:03] <abaumann> so, seamonkey seems patched.. I'll throw some bug reports upstream (and upstream/upstream, seamonkey and nss)
[11:17:36] <abaumann> ImportError: libalpm.so.12: cannot open shared object file: No such file or directory
[11:17:40] <abaumann> namcap.
[11:17:56] <deep42thought> buildmaster: wtf libalpm.so
[11:18:00] <buildmaster> deep42thought: pacman [core], pacman [staging]: /usr/lib/libalpm.so
[11:18:21] <deep42thought> buildmaster: wtp namcap
[11:18:22] <buildmaster> deep42thought: any/namcap: i486/extra (3.2.10-2.0), pentium4/extra (3.2.10-2.0), i686/extra (3.2.10-2.0)
[11:18:41] <deep42thought> hmm
[11:19:04] <deep42thought> namcap got updated, but it needs the new pacman?
[11:19:53] <deep42thought> ah, no, libalpm.so.13 is the recent one
[11:19:54] -!- peterhil has quit [Quit: Must not waste too much time here...]
[11:19:58] <deep42thought> so namcap needs a kick
[11:20:08] <deep42thought> (but it's any, so I guess, something else needs the kick)
[11:21:03] -!- peterhil has joined #archlinux32
[11:23:35] <abaumann> XPCOMGlueLoad error for file /usr/lib/seamonkey/libmozsandbox.so:
[11:23:35] <abaumann> /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by /usr/lib/seamonkey/libmozsandbox.so)
[11:23:38] <abaumann> Couldn't load XPCOM.
[11:23:43] <abaumann> cool, I can rebuild it, but it still doesn't work
[11:23:58] <abaumann> seamonkey does some really weird python tricks reading symbols from libxul.so
[11:26:04] <abaumann> ah.
[11:26:14] <abaumann> stupid me. building seamonkey on staging with newer glibc
[11:26:15] <abaumann> :-)
[11:26:18] <abaumann> gcc-libs
[11:26:21] <abaumann> I meant
[11:26:23] <deep42thought> :-D
[11:26:41] <abaumann> which brings us back to the gcc 11 story line..
[11:26:57] <abaumann> I already dropped gcc10 (which might not have been the smartest decision):
[11:27:15] <abaumann> not only cuda needs it (which is anyway blacklisted). some packages in the aur may need it (I know about palemoon)
[11:27:32] <abaumann> Are we risking something, when we push gcc 11 to stable?
[11:28:13] <deep42thought> not much, I think
[11:28:22] <deep42thought> only that said aur packages won't build anymore
[11:28:31] * abaumann shrugs
[11:28:36] <deep42thought> just do it
[11:28:40] <abaumann> okay.
[11:28:46] <abaumann> I'll test on testing first
[11:28:51] <abaumann> (as the name implies :-) )
[11:39:05] <deep42thought> I wouldn't stop gcc's move only because aur packages do not build
[11:39:14] <deep42thought> upstream has similar problems
[11:39:26] <deep42thought> and if they release a gcc10 package, we'll have it, too
[11:39:34] <deep42thought> if they don't, then I don't care, either :-)
[11:43:34] <abaumann> we _had_ it too
[11:43:38] <abaumann> it breaks currently
[11:43:44] <deep42thought> gcc10?
[11:43:47] <abaumann> yepo
[11:44:10] <deep42thought> err and you *removed* it?
[11:44:14] <deep42thought> or it didn't build?
[11:44:19] <abaumann> yes
[11:44:22] <deep42thought> which
[11:44:22] <abaumann> blacklisted
[11:44:28] <deep42thought> :-/
[11:44:35] <abaumann> we can unblacklist and fix the bug :-)
[11:44:42] <deep42thought> it's
[11:44:50] <deep42thought> probably the right thing to do
[11:45:01] <abaumann> ok, then I'll unklacklist
[11:55:25] <abaumann> mmh, so far no damange with gcc-libs 11 as far as I can tell (famous last words)..
[11:55:42] <deep42thought> well, then: move it!
[11:55:50] <abaumann> aeh.. already done. :-)
[11:55:54] <deep42thought> good
[11:56:01] * abaumann has a fast trigger finger
[12:21:01] -!- juanfra has quit [Ping timeout: 244 seconds]
[13:01:19] -!- City-busz has quit [*.net *.split]
[13:01:19] -!- oaken-source has quit [*.net *.split]
[13:01:19] -!- doskoi has quit [*.net *.split]
[13:01:20] -!- wutno has quit [*.net *.split]
[13:01:20] -!- CalimeroTeknik has quit [*.net *.split]
[13:01:20] -!- elibrokeit has quit [*.net *.split]
[13:10:44] -!- abaumann has quit [Quit: leaving]
[13:11:26] -!- City-busz has joined #archlinux32
[13:11:26] -!- oaken-source has joined #archlinux32
[13:11:26] -!- doskoi has joined #archlinux32
[13:11:26] -!- CalimeroTeknik has joined #archlinux32
[13:11:26] -!- wutno has joined #archlinux32
[13:11:26] -!- elibrokeit has joined #archlinux32
[13:22:13] -!- deep42thought has quit [Quit: Leaving.]
[15:05:13] -!- eschwartz has quit [Changing host]
[15:05:13] -!- eschwartz has joined #archlinux32
[15:05:13] -!- elibrokeit has quit [Changing host]
[15:05:13] -!- elibrokeit has joined #archlinux32
[15:05:27] -!- dvzrv has quit [Changing host]
[15:05:27] -!- dvzrv has joined #archlinux32
[15:06:15] -!- Namarrgon has quit [Changing host]
[15:06:15] -!- Namarrgon has joined #archlinux32
[15:36:18] -!- peterhil has quit [Ping timeout: 264 seconds]
[15:52:33] -!- abaumann has joined #archlinux32
[15:52:34] <buildmaster> Hi abaumann!
[15:52:34] <buildmaster> !rq abaumann
[15:52:35] <phrik> buildmaster: <abaumann> I fail to decive what is crappier: rust or LTO in gcc..
[15:58:43] -!- dvzrv has quit [Quit: WeeChat 3.1]
[15:58:50] -!- dvzrv has joined #archlinux32
[16:00:32] -!- peterhil has joined #archlinux32
[16:22:44] -!- wutno has quit [Quit: Kernel Panic]
[16:27:40] -!- juanfra has joined #archlinux32
[17:45:33] -!- doskoi has quit [Quit: kthxbye]
[17:46:28] -!- doskoi has joined #archlinux32
[17:49:07] -!- doskoi has quit [Client Quit]
[17:50:04] -!- doskoi has joined #archlinux32
[18:28:01] -!- slacka has joined #archlinux32
[18:32:12] -!- slacka123_ has joined #archlinux32
[18:35:04] -!- slacka has quit [Ping timeout: 272 seconds]
[19:22:50] -!- abaumann has quit [Quit: leaving]
[19:51:31] -!- abaumann has joined #archlinux32
[19:51:31] <buildmaster> Hi abaumann!
[19:51:32] <buildmaster> !rq abaumann
[19:51:33] <phrik> buildmaster: <abaumann> naa.. not testing never has bad consequences.. ;-)
[19:52:13] -!- abaumann has quit [Client Quit]
[20:46:19] -!- peterhil has quit [Ping timeout: 244 seconds]
[21:28:18] -!- peterhil has joined #archlinux32
[21:38:35] -!- drathir_tor has quit [Remote host closed the connection]
[21:39:36] -!- drathir_tor has joined #archlinux32
[22:43:44] -!- peterhil has quit [Ping timeout: 252 seconds]
[22:50:37] -!- drathir_tor has quit [Ping timeout: 252 seconds]
[23:07:24] -!- drathir_tor has joined #archlinux32