<aterius>
antifuchs: I want to test out the -L/B gcc options but unfortunately still haven't figured out why p11-kit is not building locally, have to rebuild a lot of changes due to the modifications to builder/preConfigure
<antifuchs>
hm, I'm not sure how I got p11-kit built at all, tbqh - might have gotten it from the nixos cache?
<aterius>
Yeah, it works from nix-cache, but I'm not hitting the cache when I use your checkout
<aterius>
* Yeah, it works from hydra cache, but I'm not hitting the cache when I use your checkout
rajivr has joined #nixos-emacs
<antifuchs>
ah, makes sense. hm... I am not sure what's up there - I'm pretty sure my change doesn't touch p11-kit or its deps /:
<antifuchs>
I'm building a version with the -B switches now... ideally, it would also allow us to get rid of the `wrapProgram` thing it does to set $PATH to the linker/assembler
<antifuchs>
also, argh, I've been trying to make a version that supports the macos app bundle, but it's pretty difficult to integrate in the way gccemacs does .eln file relocation in the dumped image
<antifuchs>
probably won't be awake to see the result of this -B build, unfortunately; I'll update you when I wake up tomorrow (:
<antifuchs>
ah! it finished! and it works (:
<antifuchs>
I'll push this when my build on linux finishes (:
cole-h has joined #nixos-emacs
cole-h has quit [Quit: Goodbye]
joesventek has quit [Quit: Quit]
joesventek has joined #nixos-emacs
joesventek has quit [Client Quit]
joesventek has joined #nixos-emacs
<antifuchs>
ok, so that works, but it requires a wrapper program, because if you invoke the emacs binary from $PATH directly, `invocation-directory` is empty and then the pdump fixup paths are all wrong. omg.
neeasade has joined #nixos-emacs
rajivr has quit [Quit: Connection closed for inactivity]
cole-h has joined #nixos-emacs
<aterius>
Sorry I missed this!
<aterius>
I'm assuming you only need the wrapper on macOS?
<antifuchs>
I think so, but am not sure. I am chatting about it on the emacs bts now and apparently this is partly because invocation-directory is not set yet at the time the relocation code runs
<antifuchs>
I’ll have to put in some asserts (ought to build for debug too tbqh)
<aterius>
Nice! Excited for this to land as it fixes the library_path issue as well for linux