#archlinux32 | Logs for 2022-10-08

[01:44:15] -!- bill-auger has quit [Ping timeout: 265 seconds]
[13:02:28] -!- bill-auger has joined #archlinux32
[13:08:02] -!- abaumann has joined #archlinux32
[13:08:02] <buildmaster> Hi abaumann!
[13:08:02] <buildmaster> !rq abaumann
[13:08:03] <phrik> buildmaster: <abaumann> there will always be things, we cannot detect.
[13:09:20] <abaumann> ok, starting over with ruby, things are completely broken. Now I try to build an out-of-distro-tree version of ruby-boostrap (no symlinks in /usr/bin). This means, I have to patch every PKGBUILD for building the base ruby packages, but this way I don't get into weird conflicts like wrong shell bangs referencing the wrong interpreter or ruby search package files and shared libraries in the wrong
[13:09:26] <abaumann> place.
[13:44:31] <abaumann> conflicts=ruby is a bad idea, so ruby-bootstrap and ruby get into conflict.
[13:44:44] <abaumann> I'm dropping the symlinks in /usr/bin of the ruby-bootstrap package..
[13:46:25] -!- GNUtoo has quit [Ping timeout: 258 seconds]
[13:46:43] -!- epony has quit [Ping timeout: 252 seconds]
[14:32:07] -!- GNUtoo has joined #archlinux32
[15:20:07] -!- GNUtoo has quit [Remote host closed the connection]
[15:25:19] -!- GNUtoo has joined #archlinux32
[15:40:33] -!- abaumann has quit [Quit: leaving]
[15:55:03] -!- finsternis has quit [Read error: Connection reset by peer]
[16:32:34] -!- abaumann has joined #archlinux32
[16:32:34] <buildmaster> Hi abaumann!
[16:32:34] <buildmaster> !rq abaumann
[16:32:35] <phrik> buildmaster: <abaumann> people tend to rewrite things in new hip languages, just because they are too lazy to read old code
[16:32:54] <abaumann> lot's of swearing and bugs later I have now a rubygems. :-)
[16:33:20] <abaumann> and rake in /usr and /opt/ruby
[16:33:33] <abaumann> now I should be able to properly build ruby packages again..
[16:33:38] <abaumann> ..minus testing, of course.
[16:36:01] <abaumann> /startdir/PKGBUILD: /usr/bin/gem: /opt/ruby-bootstrap-3.0.4/bin/ruby: bad interpreter: No such file or directory
[16:36:08] <abaumann> mmh, well, that's to be expected..
[16:49:42] <abaumann> /usr/lib/ruby/3.0.0/i686-linux/openssl.so: warning: already initialized constant OpenSSL::OPENSSL_VERSION
[16:49:49] <abaumann> LoadError: cannot load such file -- digest.so
[16:49:49] <abaumann> Did you mean? digest
[16:49:49] <abaumann> <internal:/usr/lib/ruby/3.0.0/rubygems/core_ext/kernel_require.rb>:85:in `require'
[16:49:54] <abaumann> this whole thing is annoying..
[17:14:37] -!- epony has joined #archlinux32
[18:12:59] <abaumann> usr/lib/ruby/3.0.0/x86-linux/digest.so
[18:13:03] <abaumann> aeh.. really?
[18:17:56] <abaumann> opt/ruby-bootstrap-3.0.4/lib/ruby/3.0.0/i686-linux/digest.so
[18:34:42] <abaumann> this architecture string must be guessed somehow, I build everything in the same chroots..
[18:36:04] <abaumann> that's in the rebuild of ruby itself: usr/lib/ruby/3.0.0/i486-linux
[18:36:25] <abaumann> ok, so when I am forcing the installation to /usr from /opt, the architecture strings differ.. interesting..
[18:37:49] <abaumann> Should I now fiddle with LD_LIBRARY_PATH, rebuild stuff and then hope, things get magically better..?
[18:38:53] <abaumann> aeh, RUBYLIB..
[18:40:29] <abaumann> ah, and every gem, ruby, rake has to get /usr/bin/ruby /usr/bin/XXX
[18:40:56] <abaumann> rake aborted!
[18:40:56] <abaumann> Errno::ENOENT: No such file or directory - gem
[18:40:58] <abaumann> better, but no..
[18:42:28] <abaumann> rubygems is missing.. aha..
[18:42:51] <abaumann> nope..
[18:43:01] <abaumann> cp tmp/i486-linux/digest/sha2/3.0.4/digest/sha2.so tmp/i486-linux/stage/lib/digest/sha2.so
[18:43:15] <abaumann> but files would now be installed into the right directories..
[18:47:45] <abaumann> oh, bundler is being stupid about calling gem
[18:48:26] <abaumann> #!/opt/ruby-bootstrap-3.0.4/bin/ruby
[18:48:53] <abaumann> yeah, why can't I force the shell bang when installing? Actually, I could rebuild that package and patch it in the end..
[19:10:46] <abaumann> yep, this worked, so let's do that for rake and rake-compiler too..
[19:56:27] <abaumann> usr/lib/ruby/3.0.0/x86-linux/digest.so
[19:56:47] <abaumann> this is so annoying, I don't see a pattern, I cannot pass the build triplet anywhere..
[19:57:57] <abaumann> local _platform="$(gem env platform | cut -d':' -f2)"
[19:58:00] <abaumann> oh.
[19:58:11] <abaumann> *sigh*
[20:13:03] <abaumann> if I overwrite the platform, it fails, if i don't things end up in usr/lib/ruby/3.0.0/x86-linux
[20:20:17] <abaumann> let's rebuild rubygems, maybe gem platform returns then something more sane..
[20:22:03] <abaumann> /usr/bin/rdoc exists in both 'rubygems' and 'ruby-rdoc'
[20:22:03] <abaumann> /usr/bin/ri exists in both 'rubygems' and 'ruby-rdoc'
[20:22:06] <abaumann> seriously!?
[20:25:44] <abaumann> ruby-rdoc is very old, so maybe some files moved in the meantime..
[20:28:03] <abaumann> brilliant, now I cannot rebuild rubygems because it has a conflict with rdoc..
[20:28:23] <abaumann> ..so on every error you have to go back to field one and restart bootstrapping..
[20:28:37] <abaumann> ..unless this process gets automatized, I see little hope..
[20:31:30] <abaumann> I read the mailing list upstream about this ruby-stdlib change, but I don't see, how this helps me..
[20:44:55] <abaumann> brilliant. as soon as one package is wrongly build (which is part of ruby-stdlib), you can stop right there and start all over..
[20:45:00] <abaumann> ..another day.. (or never).
[20:45:02] -!- abaumann has quit [Quit: leaving]