<mirai>is it possible to retrieve the "latest" patch (or series of patches) for a given issue?
<apteryx>in Emacs with Debbugs: M-x cd -> place your directory to your guix checkout; then put your cursor on the first mail, then C-u N | git am -3 RET, where N is the number of messages/patches you want to apply
<mirai>I'm looking for a more programmatic way of doing it
<lfam>Howdy everyone at the Guix Days. I'm curious if there is a projector / beamer, and if it does HDMI?
<Kabouik> Thanks gabber`, Lumine. I didn't know about Flatseal, I just used it to give access to that other app to my system libraries, but that doesn't seem to be enough. I'll see if any other option might be required.
<lfam>I have a brief presentation / discussion I'd like to give tomorrow
<Kabouik>Do you have any experience using Flatseal to give access to flatpaks to things like openssl Lumine? The options I tried did not change anything. Maybe Flatseal is not of much help with the Guix file hierarchy?
<Fd9a>Hi. I built the debian package of guix for riscv with guix pack --system=riscv64-linux -f deb, after installing in /gnu/store some files have a date other than 1 jan 1970, should I change it manually or should I just leave it as it is?
<gabber`>are these dates reproducible or do they differ with each build?
<Fd9a>It took me a week to build this, so I can hardly check. The date falls in the middle of the week with almost the same time.
<gabber`>why do you want to have the dates be UTC 0? i thought you created a package that you built with the command above and were getting strange artifacts in your store
<Fd9a>gabber`: I'm not sure I want this. Checked it on the main car, it didn't happen there. So far everything seems to be working, so nothing critical.
<gabber`>i think the only reason the timestamp is set to UTC 0 in the store is to guarantee reproducibility. i don't think it's necessary to change anythingif you don't care about that and your software works fine
<Guest46>I would think it is still fine, since it is run only once on the substitute server and it ensures quality.
<futurile>Q: I'm looking at rust-env-logger in Guix, the actual name on crates.io is env_logger: in Guix it's a hyphen, in rust it's an underscore. Is it intentional that we convert underscores to hyphens?
<apteryx>I think I've found my issue with the first guix pack test suddenly using lots of dependencies after my rework: I've introduced some computed-file objects, that don't seem to use the %guile-for-build fluid
<apteryx>the test suite sets that so that the guile-bootstrap variant is used and dependencies kept to a minimum
<mirai>old: imo it should be fine, better to be sure that the package works
<bjc>futurile: yes, it's intentional. in addition, there's no difference between ‘-’ and ‘_’ when it comes to crate names; they map to the same object
<futurile>bjc: ah, OK, interesting - didn't know that about Crates. Not actually that familiar with them/Rust - thanks
<florhizome[m]>I am still getting a Message about a Not supported manifest Format when trying to Install, remove, Upgrade etc
<florhizome[m]><florhizome[m]> "I am still getting a Message..." <- Oh. For some reason i had 2 generations with empty manifests
<seninha>ok, i give up on this ugly service. i'm not that guix/guile literate to change a service or write one from scratc. Since this machine is just for development and testing purposes, im gonna just launch the daemon manually.
<apteryx>old: not everyone uses substitutes, and when testing affected packages, 1 h seems a long time to wait to know the result
<nckx>rekado: I did, when copying residue from 2023-01-31.log to 2023-02-01.log. Bash history says I typed ‘cp’ where I meant ‘cp -a’.
<apteryx>it'd be preferable if a subset of tests could be run in a reasonable time and give a good idea of whether the software works as advertised; some python packages tag their slow test with a marker, like 'slow', and then we can invoke 'pytest -m not slow' to skip them
<apteryx>yay! (guile (%guile-for-build)) in computed-file solved my last 'guix pack' failure
<Fd9a>nckx: directories of packages and dirs inside of them for example /gnu/store/hz4hb8sva8ihm8gynw80kb85hqd7drzc-libunistring-0.9.10/
<apteryx>wait, I think perhaps it works; at least it did for some tests
<mirai>I've had a similar problem when I was using generic-serialize-alist
<seninha>mirai: thanks, but i'll just start the daemon manually, from a script. It's just for developing a custom cryptographic engine and testing it, i do not want to keep the service active for much time all
<seninha>i'll probably hit a point where i'll need to pass more flags and do finer configuration, which will require me to patch the module again
<mirai>This looks redundant: #~(string-append #$@(generic-serialize-alist list mpd-serialize-field value)))
<mirai>You'd think (generic-serialize-alist string-append <lambda> value) would perhaps be better
<mirai>but if lambda returns gexps, string-append won't work with them
<Fd9a>nckx: guix time-machine -C channels.scm --disable-authentication -- pack --system=riscv64-linux -f deb guix I used my channel where I disabled tests for some packages because the cross-compilation fails. https://git.sr.ht/~akagi/guix which is checkout from ae071e21f3a42c0c30b5c83f90a9ca022ebcac4d master
<nate1>That is very handy, is the manual located locally as well?
<vagrantc>efraim: i noticed you changed diffoscope to build with cdrkit-libre ... it should probably also be changed in upstream diffoscope if it is indeed fully compatible ... although guix has both cdrtools and cdrkit-libre ... either could technically work?