Re: bug#39324: GNOME Weather doesn't start
I can reproduce. Note that from GNOME, running "gnome-weather" in a shell works. However the icon does not work either. I noticed that the desktop file executes --8<---cut here---start->8--- gapplication launch org.gnome.Weather --8<---cut here---end--->8--- Does any one know how this is supposed to work? If not, as a workaround I suggest we replace this line with "gnome-weather". -- Pierre Neidhardt https://ambrevar.xyz/
Re: bug#39117: GNOME Files: No file thumbnails
I can reproduce. Note that I've managed to get _some_ thumbnails when I first logged in. Now the thumbnail generation seems to be completely gone. Any clue what's happening, anyone? Has this ever worked before? -- Pierre Neidhardt https://ambrevar.xyz/
Unencrypted boot with encrypted root
Hi! I've followed the doc / template to set up an encrypted system on my laptop: --8<---cut here---start->8--- (mapped-devices (list (mapped-device (source (uuid "12345678-1234-1234-1234-123456789abc")) (target "my-root") (type luks-device-mapping (file-systems (append (list (file-system (device (file-system-label "my-root")) (mount-point "/") (type "ext4") (dependencies mapped-devices)) (file-system (device (uuid "1234-ABCD" 'fat)) (mount-point "/boot/efi") (type "vfat"))) %base-file-systems)) --8<---cut here---end--->8--- Problem is, I get prompted for the LUKS password twice: once before GRUB starts and once when booting an OS entry. This is rather annoying (and quite slow by the way, it takes some 10-20 seconds) and probably not too useful. Is it possible to prompt for the password only once? I suppose that one way to do this is to make /boot a separate file system beside /boot/efi. All in all, the configuration would look like this: --8<---cut here---start->8--- (mapped-devices (list (mapped-device (source (uuid "12345678-1234-1234-1234-123456789abc")) (target "my-root") (type luks-device-mapping (file-systems (append (list (file-system (device (file-system-label "my-root")) (mount-point "/") (type "ext4") (dependencies mapped-devices)) (file-system (device (file-system-lavel "boot") (mount-point "/boot") (type "ext4")) (file-system (device (uuid "1234-ABCD" 'fat)) (mount-point "/boot/efi") (type "vfat"))) %base-file-systems)) --8<---cut here---end--->8--- We should probably update the doc and templates to explain this subtlety, since mistakes in the partition design are hard to recover after the fact :) Insights? -- Pierre Neidhardt https://ambrevar.xyz/ signature.asc Description: PGP signature
Re: bug#30435: libreoffice: Fonts don't show up after install
I can reproduce this on guix ce226e9d8d52d2530f057f2000d36c0d55380ade. But libreoffice is not the only victim: it seems that most applications fail to see the fonts installed in the user profile (on Guix System). Emacs is one example. Running --8<---cut here---start->8--- fc-cache -fv --8<---cut here---end--->8--- fixes the issue. Should we run this command in a profile hook? -- Pierre Neidhardt https://ambrevar.xyz/
Re: bug#30435: libreoffice: Fonts don't show up after install
Pierre Neidhardt writes: Hi, I wanted to add that the opposite is also true: removing a font without running fc-cache always makes my emacs die on start, so I agree that running the command automatically makes sense. Are there any unwanted side-effects? Thanks, Nicolò > I can reproduce this on guix ce226e9d8d52d2530f057f2000d36c0d55380ade. > > But libreoffice is not the only victim: it seems that most applications > fail to see the fonts installed in the user profile (on Guix System). > Emacs is one example. > > Running > > --8<---cut here---start->8--- > fc-cache -fv > --8<---cut here---end--->8--- > > fixes the issue. > > Should we run this command in a profile hook? > > -- > Pierre Neidhardt > https://ambrevar.xyz/
Re: bug#30435: libreoffice: Fonts don't show up after install
Hi there, I think this is the same issue as 26877. Can font information be compiled into a fonts.conf? - John
Re: Thank you for your leadership Ludo
Hi Joshua, Joshua Branson skribis: > I have been hanging out in #guix recently, and I have submitted a few > guix bug reports (not many). I just wanted to thank you for the > community that you have built with Guix. I have done some minor > volunteer work with other free software projects, but those other free > software projects did not have Guix's incredible community. It is > incredibly nice to feel like I can help Guix, even when it is sparingly. > > I also want to thank the other /incredibly/ dedicated Guix maintainers. > Thanks for all that you do! Thanks for the kind words. Each one of us contributes to the pleasant atmosphere on the project’s channels, so we’re all to be thanked! > P.S. I hope this is not consider off topic. I just recently read the > first few chapters of "How to Win friends and Influence People". It > recommended giving out lots of compliments. I thought that I would > try it. Now I wonder how you’re going to try to influence me. ;-) Ludo’.
Re: [BLOG] On migration to the Hurd
Hi! Tanguy Le Carrour skribis: > Le 04/01, Jan Nieuwenhuizen a écrit : >> We are thrilled to have published a post about migrating to the Hurd: >> >> https://guix.gnu.org/blog/2020/deprecating-support-for-the-linux-kernel/ > > So, I guess I was not the only one to figure out that it was a joke! A > good one, but still… a joke! > > But anyway, I clicked and read the post! And I couldn't help but feel… > mmm… nostalgic for the future!? Is that even possible?! > > The question is now: if not yesterday, when!? > > Thanks to all the people who will help make it a reality! Yup, it can actually become a reality! Yesterday led us to hack some more. The new ‘wip-hurd-vm’ branch is based on ‘core-updates’, which includes Janneke’s Hurd work, notably supporting native builds on GNU/Hurd. On that branch, from a GNU/Linux box, one can run: ./pre-inst-env guix build -f gnu/system/hurd.scm That gives you a QEMU image containing a cross-built GNU/Hurd system, which is pretty cool. Unfortunately, the bootstrap ext2fs.static server currently hangs early on for reasons that haven’t been elucidated yet. For anyone who wants to fiddle with the Hurd, here’s a good hacking opportunity! The branch is almost mergeable, I think. With that in place, it’ll be trivial to provide a “GNU/Hurd VM service” for Guix System running on GNU/Linux, and we can use that to provide GNU/Hurd build VMs on berlin, for instance. Currently, unlike ‘qemu-image’, the new ‘cross-hurd-image’ doesn’t expect a full-blown but simply a bunch of store items. However, we can hopefully use that as a starting point to port Guix System itself to GNU/Hurd. Food for thought! Ludo’. PS: In the meantime, Linux-libre support remains available, fear not!
Re: bug#39117: GNOME Files: No file thumbnails
On Thu, 02 Apr 2020 02:55:34 -0500 Pierre Neidhardt wrote > I can reproduce. > > Note that I've managed to get _some_ thumbnails when I first logged in. > Now the thumbnail generation seems to be completely gone. > > Any clue what's happening, anyone? Has this ever worked before? I've been using GNOME in the Guix System since I installed it (almost a year ago), and this have never worked for me. Also, opening the files with some applications will make the thumbnails work for these particular files. For example, opening images in the GIMP or browsing files with file managers like Thunar and Caja (if I recall correctly).
New Tamil PO file for 'guix' (version 1.1.0-pre1)
Hello, gentle maintainer. This is a message from the Translation Project robot. A revised PO file for textual domain 'guix' has been submitted by the Tamil team of translators. The file is available at: https://translationproject.org/latest/guix/ta.po (We can arrange things so that in the future such files are automatically e-mailed to you when they arrive. Ask at the address below if you want this.) All other PO files for your package are available in: https://translationproject.org/latest/guix/ Please consider including all of these in your next release, whether official or a pretest. Whenever you have a new distribution with a new version number ready, containing a newer POT file, please send the URL of that distribution tarball to the address below. The tarball may be just a pretest or a snapshot, it does not even have to compile. It is just used by the translators when they need some extra translation context. The following HTML page has been updated: https://translationproject.org/domain/guix.html If any question arises, please contact the translation coordinator. Thank you for all your work, The Translation Project robot, in the name of your translation coordinator.
Re: bug#30435: libreoffice: Fonts don't show up after install
Hi, Pierre Neidhardt skribis: > Running > > fc-cache -fv > > fixes the issue. > > Should we run this command in a profile hook? Profile hooks are normal derivations; as such, they don’t have access to anything but their dependencies and their output(s). There’s currently no infrastructure to run arbitrary code upon package installation (which I think is a feature more than a bug :-)). We could make an exception, but it’s kinda ugly. I wonder if, instead, we could have Fontconfig realize that the cache is stale somehow. Alternately, we could generate the cache in a profile hook and have Fontconfig use that cache instead of the one in ~/.cache. However, Fontconfig would need to be able to: 1. Be told which cache to use, not just the one from ~/.guix-profile, so that it works equally well with other profiles. 2. Merge several caches, so it can also account for fonts installed in /run/current-system/profile. We discussed all this several times in the past but I don’t think it went further. Ludo’.
Re: bug#30435: libreoffice: Fonts don't show up after install
Hi Nicolò, Nicolò Balzarotti skribis: > Hi, I wanted to add that the opposite is also true: removing a font > without running fc-cache always makes my emacs die on start, so I agree > that running the command automatically makes sense. Actually, if you can capture a stack trace of Emacs, that may tell us what part of Fontconfig to look at if we want to change it to gracefully handle stale cache entries. Ludo’.
Re: bug#30435: libreoffice: Fonts don't show up after install
OK, thanks for the details. So the solution is the same as the one we discussed previously about installing fonts to the non-default profile. Makes sense. I might give it a shot when I find the time ;) -- Pierre Neidhardt https://ambrevar.xyz/ signature.asc Description: PGP signature
Packaging ZynFusion - what to do with ZynAddSubFX
Hi Guix,
I managed to build ZynFusion after some help from its developer (see
attachment for the current working package definitions), but we have a
small problem.
~ Background info ~
ZynFusion uses the exact same source code as ZynAddSubFX but passing a
compilation flag enables a different, more modern-looking UI called
"ZynFusion".
In addition to stand-alone binaries, they both come as LV2 plugins.
Each LV2 plugin is identified by its unique URI, and each plugin UI is
also identified by a separate unique URI (usually in the form of
#UI). Plugins can have multiple UIs, and there is usually 1
UI for each supported platform.
~ End background info ~
The problem appears when when you install both ZynFusion and
ZynAddSubFX and you want to use them as plugins. If you only run their
binary versions (bin/zynaddsubfx) there shouldn't be a problem, but
when you attempt to open them as LV2 plugins, they currently use the
same URI so hosts will ignore one of them since most hosts use the lilv
library for detecting and loading plugins, which skips duplicates (I
believe it is undefined which one gets skipped - I think only the first
one found is used).
Since plugins can have multiple UIs, we could patch zynaddsubfx to
present itself as having 2 X11 UIs, one for zynfusion and one for its
current one, however current hosts don't offer an option to choose a UI
and simply choose whatever's found first. I believe this is the correct
behavior - the users are usually musicians and plugin format stuff
should be hidden as much as possible. Also, it is very rare for a
plugin to have more than 1 UI on the same platform, so hosts usually
just pick up whatever supported UI is found first.
Anyway, there will be problems if we allow users to install both
packages as-is, so here are our options as far as i can tell:
a) Change the URI of the plugin itself so that it's different from
ZynAddSubFX - this IMO is out of the question because presets won't be
portable anymore as they are bound to a plugin's URI.
b) Patch the plugin's turtle code to make it list both UIs as
available. The problem with this is mentioned above - current hosts
don't offer an option to choose a UI (and probably shouldn't) so one of
them (undefined which one - depends on the host's logic) will be
ignored.
c) Deprecate zynaddsubfx - this will mean users will no longer be able
to use the old UI if they prefer it, but it will probably be a good
solution for the majority of users and no changes are required to guix
d) Implement a "conflicts" variable in package definitions where you
can list incompatible packages, so guix won't allow you to install
zynfusion if you have zynaddsubfx installed, and vice versa.
I am personally fine with either c/d, although I believe d) is the
cleanest solution.
Any suggestions welcome.
Thanks,
Alex
(define-public libuv-static
(package/inherit
libuv
(arguments
'(#:tests? #f
#:make-flags (list "CFLAGS=-fPIC")))
(version "1.9.1")))
(define-public mruby-zest
(package
(name "mruby-zest")
(version "3.0.5-ba39aabd")
(source
(origin
(method git-fetch)
(uri (git-reference
;; this is a meta repo that packs the mruby dependencies
;; as submodules
(url "https://github.com/mruby-zest/mruby-zest-build.git";)
;; ghaction branch - suggested by the developer to avoid
;; automatic downloading of some unneeded and
;; hard-to-package dependencies used only for debugging
(commit "ba39aabd8d4ddc5f14137083b6f9a96c536f5f12")
(recursive? #t)))
(file-name (git-file-name name version))
(sha256
(base32
"1vqzdds30sr982dp7fclg4r19l44rv8pbz6h4a8vcginj494gvjn"
(build-system gnu-build-system)
(arguments
`(#:tests? #f ; no check target
#:make-flags
(list (string-append "CC=gcc"))
#:phases
(modify-phases %standard-phases
(add-after 'unpack 'use-installed-libuv
(lambda* (#:key inputs outputs #:allow-other-keys)
(let* ((libuv (assoc-ref inputs "libuv-static")))
(copy-file (string-append libuv "/lib/libuv.a")
"deps/libuv.a")
(substitute* "Makefile"
(("\\.\\./\\.\\./deps/\\$\\(UV_DIR\\)/include")
(string-append libuv "/include")))
(substitute* "Makefile"
(("\\./deps/\\$\\(UV_DIR\\)/\\.libs/libuv\\.a")
(string-append libuv "/lib/libuv.a"))
(add-after 'unpack 'disable-unused-deps
(lambda _
(substitute* "build_config.rb"
(("conf\\.gem 'deps/mruby-file-stat'")
"#"))
(substitute* "deps/mruby-dir-glob/mrbgem.rake"
(("spec\\.add_dependency 'mruby-file-stat'")
"#"))
Re: Linphone
Hello Guix! At this point for linphoneqt a.k.a linphone-desktop, I am facing following issues. When I build *without* `-DENABLE_DBUS=YES`and run the program, I get: QSocketNotifier: Socket notifiers cannot be enabled or disabled from another thread QMutex: destroying locked mutex When I build *with* `-DENABLE_DBUS=YES` and run the program, I get: Segmentation Fault (Core Dumped) I think the following patch is relevant, but when I use it, doesn't get successfully patched during the build. https://gitlab.linphone.org/BC/public/linphone-desktop/commit/9cf08623e3092fa19366e5c07fbe06898a59f039.diff Any ideas on how to fix this situation? Package definitions are available at https://issues.guix.gnu.org/issue/40264. Latest revision for this program is '14-add-linphoneqt-v3'. Thank you! Regards, RG.
Mistaken authorship on some commits
Hi Guix! I think I am reported as the author of commits 0cc8cdbe1b0018ec37b1de9032c9eca884bedb6e a97b943b5590c6406a85cb0f2f03fa69d7e3b7d8 c26fd5648c2a24dbd71f4c0851f8b5eced75e0f1 I did not author these, though it would have been likely. Thank you so much, maintainers. - John
Adding ‘rottlog-service-type’ to ‘%base-services’
Hello Guix! We discussed recently that we should add a ‘rottlog-service-type’ instance to ‘%base-services’. It’s a trivial change that makes a lot of sense to me. It’s all fine but the problem is that it leads to a build failure of etc.drv for those who were already adding ‘rottlog-service-type’ to their services (because we end up with two instances of that service type, both of which try to add /etc/rottlog.) Perhaps that’s fine, and we can provide a news entry to let people now? Incidentally, I think we should probably stop using GNU rottlog and implement our own stuff: it wouldn’t be much work and would be much more flexible (and we wouldn’t need that /etc/rottlog entry!). Thoughts? Ludo’. diff --git a/gnu/services/base.scm b/gnu/services/base.scm index 8d9a563e2b..a0179c0259 100644 --- a/gnu/services/base.scm +++ b/gnu/services/base.scm @@ -2444,6 +2444,8 @@ to handle." (service guix-service-type) (service nscd-service-type) +(service rottlog-service-type) + ;; The LVM2 rules are needed as soon as LVM2 or the device-mapper is ;; used, so enable them by default. The FUSE and ALSA rules are ;; less critical, but handy.
Re: Mistaken authorship on some commits
John, Thanks for reporting this! I'm CC'ing ngz, who pushed (or at least signed) all 3 commits. John Soo 写道： I think I am reported as the author of commits 0cc8cdbe1b0018ec37b1de9032c9eca884bedb6e a97b943b5590c6406a85cb0f2f03fa69d7e3b7d8 c26fd5648c2a24dbd71f4c0851f8b5eced75e0f1 Indeed. I did not author these, though it would have been likely. Weird. Kind regards, T G-R signature.asc Description: PGP signature
Re: Mistaken authorship on some commits
On Thu, Apr 02, 2020 at 10:03:27AM -0700, John Soo wrote: > Hi Guix! > > I think I am reported as the author of commits > > 0cc8cdbe1b0018ec37b1de9032c9eca884bedb6e > a97b943b5590c6406a85cb0f2f03fa69d7e3b7d8 > c26fd5648c2a24dbd71f4c0851f8b5eced75e0f1 If you do `git show --format=full` on these commits, it shows who actually committed them to the repo. I don't know what happened here, but I can say that sometimes during patch review, the patch files don't apply so I rewrite them by hand and then commit them with `git commit --author="Somebody else"`. I've made similar mistakes with this workflow...
Re: Mistaken authorship on some commits
Tobias Geerinckx-Rice 写道： Thanks for reporting this! I'm CC'ing ngz, who pushed (or at least signed) all 3 commits. They answered me on IRC: Ah yes. I made a mistake on two commits. Not a disaster but would be nice to find out what happened and how to avoid it. OK. (3?) It's simple. I had to apply manually all their patches. So I changed the "=A" flag in Magit, and apparently, I forgot to clear it. 3? I'm only aware of "emacs-spacemacs-theme" and "emacs-all-the-icons". Oh, right, there is also "emacs-graphviz-dot-mode". As a side note, I have not seen any new mail from guix-devel since January, 8th. oO; All right 🙂 Magit explains a lot. Next time do send a quick heads-up to guix-devel@ (and in this case the ‘author’) when you discover it yourself. Kind regards, T G-R signature.asc Description: PGP signature
Re: Mistaken authorship on some commits
Leo, Leo Famulari 写道： If you do `git show --format=full` on these commits, it shows who actually committed them to the repo. Are you sure this isn't trivially spoofable? Not that it would be a big deal, but ‘actually committed’ might give someone wrong ideas. I don't know what happened here, but I can say that sometimes during patch review, the patch files don't apply so I rewrite them by hand and then commit them with `git commit --author="Somebody else"`. I've made similar mistakes with this workflow... See my other reply. Kind regards, T G-R signature.asc Description: PGP signature
Re: Mistaken authorship on some commits
Tobias Geerinckx-Rice 写道： If you do `git show --format=full` on these commits, it shows who actually committed them to the repo. Are you sure this isn't trivially spoofable? Not that it would be a big deal, but ‘actually committed’ might give someone wrong ideas. Forgot to add: I'd recommend --show-signature for that reason. Kind regards, T G-R signature.asc Description: PGP signature
Re: Mistaken authorship on some commits
On Thu, Apr 02, 2020 at 07:43:13PM +0200, Tobias Geerinckx-Rice wrote: > Tobias Geerinckx-Rice 写道： > > > If you do `git show --format=full` on these commits, it shows who > > > actually committed them to the repo. > > > > Are you sure this isn't trivially spoofable? > > > > Not that it would be a big deal, but ‘actually committed’ might give > > someone wrong ideas. > > Forgot to add: I'd recommend --show-signature for that reason. Sure, that's why we require signed commits.
Re: Adding ‘rottlog-service-type’ to ‘%base-services’
Ludo', A good idea! Ludovic Courtès 写道： Perhaps that’s fine, and we can provide a news entry to let people now? I agree that's fine. Kind regards, T G-R signature.asc Description: PGP signature
Re: Linphone
Hello Guix! > At this point for linphoneqt a.k.a linphone-desktop, I am facing following > issues. > > When I build *without* `-DENABLE_DBUS=YES`and run the program, I get: > > QSocketNotifier: Socket notifiers cannot be enabled or disabled from another > thread > QMutex: destroying locked mutex > > When I build *with* `-DENABLE_DBUS=YES` and run the program, I get: > > Segmentation Fault (Core Dumped) > > I think the following patch is relevant, but when I use it, doesn't get > successfully patched during > the build. > > https://gitlab.linphone.org/BC/public/linphone-desktop/commit/9cf08623e3092fa19366e5c07fbe06898a59f0 > 9.diff > > Any ideas on how to fix this situation? Package definitions are available at > https://issues.guix.gnu.org/issue/40264. Latest revision for this program is > '14-add-linphoneqt-v3'. So I fixed "QSocketNotifier" error in version 4 patch (14-add-linphoneqt-v4). I still don't know how to fix "QMutex" error. Regards, RG.
sorting Rust crates [was Re: rav1e AV1 encoder]
On Fri, Feb 21, 2020 at 04:15:43AM -0500, Martin Becze wrote: > sort2.scm will sort a files exported packages alphanumerically. I'm working on packaging rav1e for Guix again and so I'm using sort2.scm on the big 'gnu/packages/crates-io.scm' module. One problem I noticed with sort2.scm is that inherited packages need to be after the package they inherit from, but this means they will not be alphanumerically sorted. For example, sort2.scm will put rust-bytes-0.3 before rust-bytes-0.4, but then the build fails, because the former inherits from the latter but can no longer find it.
Re: Linphone
Hello Guix! LINPHONE WORKS! \o/ Now we have a working client. It is a Legacy GTK version (linphone/liblinphone). I am still working on Qt version (linphoneqt/linphone-desktop). Please test the client and let me know if there is any issue. The latest revision patch is "13-add-linphone-v3" at https://issues.guix.gnu.org/issue/40264. Thank you! Regards, RG.
