bug#44820: MPD fails after respawning too often
Yeah, owner gets set to root when it should be mpd-user. On 12/3/20 11:05 AM, Ludovic Courtès wrote: So was the ownership of/var/run/mpd/$USER wrong?
bug#44820: MPD fails after respawning too often
I also ran into this problem. Here is the relevnat part of /var/run/mpd//log exception: Failed to create pid file "/var/run/mpd//pid": Permission denied A quick dirty fix is just to chown the /var/run/mpd folder so that mpd can create its pid file. On 11/27/20 4:31 AM, Ludovic Courtès wrote: Hi, Simon skribis: On a recent pull, MPD will not start any more. Herd status reports: Status of mpd: It is stopped. It is disabled. Provides (mpd). Requires (user-processes). Conflicts with (). Will be respawned. Last respawned on Mon Nov 23 10:22:07+0100 2020. Reading my messages, I find: Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Respawning mpd. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been started. Nov 23 10:22:07 localhost shepherd[1]: Service mpd has been disabled. Nov 23 10:22:07 localhost shepherd[1]: (Respawning too fast.) Unfortunately, there are no further messages as to why the service is disabled, or why the daemon is respawning too often. Does /var/log/mpd/mpd.log or similar contain useful info? Could you share your ‘mpd-configuration’? There have been two changes recently, which fixed the mpd system test, but perhaps they introduced other issues: https://git.savannah.gnu.org/cgit/guix.git/log?id=bb124f6e9c0af0a23736f233c2ea2c9c9b4a40a6 Thanks for reporting the issue, Ludo’.
bug#43277: emacs-next is broke, "seq" missing
--debug-init doesn't help and we don't seem to be loading the init file. On 9/8/20 9:01 AM, Michael Rohleder wrote: > Hi Martin, > > Martin Becze writes: >> emacs-next recently broke. It now has this error on start up. >> >> "require: Cannot open load file: No such file or directory, seq" > > Maybe it's possible to find where this is coming from? > (starting emacs with "--debug-init" might help or starting with a > minimal .emacs.el etc) > signature.asc Description: OpenPGP digital signature
bug#43277: emacs-next is broke, "seq" missing
I just tried out emacs27 and what replicated Malte's experience. But even after reloading my environment emacs28 doesn't work. On 9/8/20 7:40 AM, Malte Gerdes wrote: > Hi, > > I had a similar error yesterday, with emacs27. Turned out I had to > reload my environment because some variables still pointed to emacs 26.3 > directories which didn't exist anymore. > > Malte > > On Tue, 8 Sep 2020, 14:25 Martin Becze, <mailto:mjbe...@riseup.net>> wrote: > > emacs-next recently broke. It now has this error on start up. > > "require: Cannot open load file: No such file or directory, seq" > > I think this must have happened relatively recently (with the last 3 > weeks) since it was working fine earlier. > signature.asc Description: OpenPGP digital signature
bug#43277: emacs-next is broke, "seq" missing
emacs-next recently broke. It now has this error on start up. "require: Cannot open load file: No such file or directory, seq" I think this must have happened relatively recently (with the last 3 weeks) since it was working fine earlier. signature.asc Description: OpenPGP digital signature
bug#42947: Chez Scheme libraries are broken
> I see, but now we’re creating a directory that remains empty, and as I > wrote, chez-fmt is actually installed in another directory anyway: Oh yikes! Yeah so I see half of the scheme libraries are installed in lib/csv-site and other are installed in lib/csv-site. It appears that there was a missing make flag for scheme-fmt and several others. Attached is an updated patch that fixes that. On 8/31/20 3:56 PM, Ludovic Courtès wrote: Hi, Martin Becze skribis: This is creating an empty directory, right? Are you sure this helps? Yes, I think it solves the problem. On line 92 of the package definition we have ``` (native-search-paths (list (search-path-specification (variable "CHEZSCHEMELIBDIRS") (files (list (string-append "lib/csv" version "-site")) ``` search-path-specification won't set the environmental variable unless the folder already exists. Note lib/csv does exist at this point but most libraries seem to want to live lib/csv-site. I see, but now we’re creating a directory that remains empty, and as I wrote, chez-fmt is actually installed in another directory anyway: Isn’t the problem that packages like chez-fmt install libraries in ‘csv-site’ instead of ‘csv1.2.3-site’? --8<---cut here---start->8--- $ find $(./pre-inst-env guix build chez-fmt) /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11 /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt-js.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt-c.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11 /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/fmt.css /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/README /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/fmt.html /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/VERSION --8<---cut here---end--->8--- So perhaps the search path should be just lib/csv-site? Is that consistent across Chez packages? Thanks, Ludo’. >From 328ab304b47671dcdc03768f07484aae298ca646 Mon Sep 17 00:00:00 2001 From: Martin Becze Date: Thu, 20 Aug 2020 12:39:50 -0500 Subject: [PATCH v2 1/2] gnu: chez-scheme: Fix libary directory issue. This creates the directory lib/csv-site which the scheme libary packages expect. * gnu/packages/chez.scm (chez-scheme): Create libary directory. * gnu/packages/chez.scm (chez-make-flags): Add chezversion to make flags. --- gnu/packages/chez.scm | 10 +++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/gnu/packages/chez.scm b/gnu/packages/chez.scm index da0b17e799..ac8fce5a1a 100644 --- a/gnu/packages/chez.scm +++ b/gnu/packages/chez.scm @@ -194,6 +194,8 @@ (bin (string-append out "/bin")) (lib (string-append out "/lib")) (name "chez-scheme")) + ;; create the folder for libary installation + (mkdir (string-append lib "/csv" ,version "-site")) (symlink (string-append bin "/scheme") (string-append bin "/" name)) (map (lambda (file) @@ -416,9 +418,11 @@ Chez Scheme.") ;; files. (define (chez-make-flags name version) `(let ((out (assoc-ref %outputs "out"))) - (list (string-append "PREFIX=" out) - (string-append "DOCDIR=" out "/share/doc/" - ,name "-" ,version + (list + (string-append "chezversion=" ,(package-version chez-scheme)) + (string-append "PREFIX=" out) + (string-append "DOCDIR=" out "/share/doc/" + ,name "-" ,version (define-public chez-matchable (package -- 2.28.0
bug#42947: Chez Scheme libraries are broken
> This is creating an empty directory, right? Are you sure this helps? Yes, I think it solves the problem. On line 92 of the package definition we have ``` (native-search-paths (list (search-path-specification (variable "CHEZSCHEMELIBDIRS") (files (list (string-append "lib/csv" version "-site")) ``` search-path-specification won't set the environmental variable unless the folder already exists. Note lib/csv does exist at this point but most libraries seem to want to live lib/csv-site. On 8/28/20 8:22 AM, Ludovic Courtès wrote: Hi Martin, Martin Becze skribis: Chez Scheme libraries are broken. For example `guix environment --ad-hoc chez-scheme chez-fmt`. chez-fmt will not install (for me) because it is dependent on chez-srfi and chez-scheme seem unable to locate it. I think it is because CHEZSCHEMELIBDIRS is not being set. But I'm not sure why that is happening. The chez-scheme package includes the following ``` (native-search-paths (list (search-path-specification (variable "CHEZSCHEMELIBDIRS") (files (list (string-append "lib/csv" version "-site")) ``` Which I think should set it. -Martin Martin Becze skribis: By default chez-scheme seems to be using lib/csv for libaries. This creates the directory lib/csv-site which the scheme libary packages expect. * gnu/packages/chez.scm (chez-scheme): Create libary directory. --- gnu/packages/chez.scm | 2 ++ 1 file changed, 2 insertions(+) diff --git a/gnu/packages/chez.scm b/gnu/packages/chez.scm index b80421d3f7..bb584696ae 100644 --- a/gnu/packages/chez.scm +++ b/gnu/packages/chez.scm @@ -194,6 +194,8 @@ (bin (string-append out "/bin")) (lib (string-append out "/lib")) (name "chez-scheme")) + ;; create the folder for libary installation + (mkdir (string-append lib "/csv" ,version "-site")) This is creating an empty directory, right? Are you sure this helps? Isn’t the problem that packages like chez-fmt install libraries in ‘csv-site’ instead of ‘csv1.2.3-site’? --8<---cut here---start->8--- $ find $(./pre-inst-env guix build chez-fmt) /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11 /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt-js.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt-c.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/lib/csv-site/fmt.so /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11 /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/fmt.css /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/README /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/fmt.html /gnu/store/qr8mgphbivlqbdmn355kps2cxlmi679d-chez-fmt-0.8.11/share/doc/chez-fmt-0.8.11/VERSION --8<---cut here---end--->8--- Thanks, Ludo’.
bug#42990: Nicotine+ GTK segfaults
Does installing a theme like "adwaita-icon-theme" help? On 8/22/20 12:28 PM, Leo Famulari wrote: > I noticed that Nicotine+ segfaults on a foreign distro (Debian) when > clicking the "settings" icon in the upper right: > > -- > (.nicotine-real:9751): GLib-GIO-ERROR **: 12:53:24.031: Settings schema > 'org.gtk.Settings.FileChooser' does not contain a key named 'show-type-column' > -- > > Sounds similar to many other bugs using GTK+ applications from Guix on > other distros, and there's a related report at NixOS too: > > https://discourse.nixos.org/t/gtk-file-dialog-causes-segfaults-glib-gio-error-settings-schema-org-gtk-settings-filechooser-does-not-contain-a-key-named-show-type-column/6853 > > Help wanted! > > > > signature.asc Description: OpenPGP digital signature
bug#42947: [PATCH 2/2] gnu: chez-srfi: Fix tests.
* gnu/packages/chez.scm (chez-srfi): Added make flags so that chez-scheme
loads the tests correctly.
---
gnu/packages/chez.scm | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/gnu/packages/chez.scm b/gnu/packages/chez.scm
index bb584696ae..79a5ad965d 100644
--- a/gnu/packages/chez.scm
+++ b/gnu/packages/chez.scm
@@ -242,7 +242,9 @@ and 32-bit PowerPC architectures.")
`(("chez-scheme" ,chez-scheme)))
(arguments
`(#:make-flags (let ((out (assoc-ref %outputs "out")))
- (list (string-append "PREFIX=" out)))
+ (list (string-append "PREFIX=" out)
+"CHEZ=chez-scheme --libdirs ./"
+(string-append "chezversion=" ,(package-version
chez-scheme
#:test-target "test"
#:phases (modify-phases %standard-phases
(delete 'configure
--
2.28.0
bug#42947: [PATCH 1/2] gun: chez-scheme: Create libary directory.
By default chez-scheme seems to be using lib/csv for libaries. This creates the directory lib/csv-site which the scheme libary packages expect. * gnu/packages/chez.scm (chez-scheme): Create libary directory. --- gnu/packages/chez.scm | 2 ++ 1 file changed, 2 insertions(+) diff --git a/gnu/packages/chez.scm b/gnu/packages/chez.scm index b80421d3f7..bb584696ae 100644 --- a/gnu/packages/chez.scm +++ b/gnu/packages/chez.scm @@ -194,6 +194,8 @@ (bin (string-append out "/bin")) (lib (string-append out "/lib")) (name "chez-scheme")) + ;; create the folder for libary installation + (mkdir (string-append lib "/csv" ,version "-site")) (symlink (string-append bin "/scheme") (string-append bin "/" name)) (map (lambda (file) -- 2.28.0
bug#42947: Chez Scheme libraries are broken
Chez Scheme libraries are broken. For example `guix environment --ad-hoc chez-scheme chez-fmt`. chez-fmt will not install (for me) because it is dependent on chez-srfi and chez-scheme seem unable to locate it. I think it is because CHEZSCHEMELIBDIRS is not being set. But I'm not sure why that is happening. The chez-scheme package includes the following ``` (native-search-paths (list (search-path-specification (variable "CHEZSCHEMELIBDIRS") (files (list (string-append "lib/csv" version "-site")) ``` Which I think should set it. -Martin signature.asc Description: OpenPGP digital signature
bug#37244: Icecat Audio Issues
I also have this problem with icecat. With ungoogled-chromuim the video is scrambled. I wonder if the two issues are possible related. On 9/1/19 1:15 PM, Tobias Geerinckx-Rice via Bug reports for GNU Guix wrote: > Raghav, > > Raghav Gururajan 写道： >> While streaming/watching movie on some (more than one) websites, audio >> has constant cracking sounds and voice of characters in the video >> becomes robotic. > > If by ‘robotic’ you mean (extremely) low-pitched then, yes, this has > been the case for roughly the past half year. > > I asked around on IRC, even sending URLs to innocent people who were > unable to reproduce the problem, so I then assumed it wasn't in Guix > proper. I almost never watch videos on the Web so I didn't care much, > but you might be happy to learn you're not alone. > > Could you share a (free) example URL? > > Kind regards, > > T G-R signature.asc Description: OpenPGP digital signature
bug#41069: cyclically dependent channels crashes guix pull
Cyclically dependent channels crashes guix pull. Cyclically dependent channels probably shouldn't happen, but guix pull probably shouldn't go into an endless loop and die either. Cheers!, -Martin
bug#38857: X.509 certificate of 'crates.io' could not be verified during a recursive import from crates.io
I have had this problem as well. I noticed that the file descriptors
where not being closed when the connections end. I think this is causing
the issue. Related https://debbugs.gnu.org/cgi/bugreport.cgi?bug=20145
On 1/2/20 2:20 PM, Valentin Ignatev wrote:
I don't think that it's related, but who knows. I only have a
certificate issue when I'm using recursive crates import. I am able to
import packages from crates one by one without an issue as well as
doing other tls-sensitive stuff.
Regards,
Valentin
On 1/2/20, Bengt Richter wrote:
Hi Guix,
On +2020-01-02 09:12:43 +0200, Efraim Flashner wrote:
On Thu, Jan 02, 2020 at 01:45:35AM +0300, Valentin Ignatev wrote:
Hi! I'm trying to recursively import a package from crates.io like
this:
guix import crate notify@4.0.14 --recursive
It follows redirections for a while untill at some point throws this:
Backtrace:
12 (primitive-load "/home/vj/.config/guix/current/bin/guix")
In guix/ui.scm:
1806:12 11 (run-guix-command _ . _)
In guix/scripts/import.scm:
116:11 10 (guix-import . _)
In guix/scripts/import/crate.scm:
103:16 9 (guix-import-crate . _)
In guix/import/utils.scm:
425:7 8 (recursive-import _ _ #:repo->guix-package _ #:guix-name
…)
397:31 7 (topological-sort _ #
…)
In srfi/srfi-1.scm:
592:17 6 (map1 ("tempfile"))
In guix/import/utils.scm:
421:36 5 (lookup-node "tempfile")
In guix/import/crate.scm:
222:10 4 (crate->guix-package "tempfile" _)
150:15 3 (make-crate-sexp #:name _ #:version _ #:cargo-inputs _ #
…)
In guix/http-client.scm:
88:25 2 (http-fetch _ #:port _ #:text? _ #:buffered? _ # _ # _ #
…)
In guix/build/download.scm:
419:4 1 (open-connection-for-uri _ #:timeout _ # _)
306:6 0 (tls-wrap # _ # _)
guix/build/download.scm:306:6: In procedure tls-wrap:
X.509 certificate of 'crates.io' could not be verified:
signer-not-found
invalid
I suspect that it happens after the importer hits
"wasm-bindgen-webidl" and starts going circles. Maybe there's some
circullar dependencies going on, but I'm not sure. I'm attaching a
full log for convenience.
For additional info: I'm running Guix on Arch Linux. I've also
installed nss-certs package, exported all neeeded variables
(SSL_CERT_DIR, SSL_CERT_FILE and GIT_SSL_CAINFO) before running guix
import and also made sure nscd.service is running.
Regards,
Valentin Ignatev
I've had it happen to me also sometimes. It's like it forgets that it
just successfully connected 100+ times and then fails.
--
Efraim Flashner אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D 14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received
unencrypted
I don't know if this could be related, but...
I am also running guix on Archlinux and experienced a TLS problem
after doing pacman -Syu.
Mutt got updated and I could no longer get my pop mail.
I reverted the last mutt update:
--8<---cut here---start->8---
[2020-01-01T15:53:13-0800] [ALPM] downgraded mutt (1.13.2-1 -> 1.12.2-1)
--8<---cut here---end--->8---
And am writing this with the reverted verssion.
(So BTW this may be a heads-up not to package 1.13.2-1 until the problem
is resolved, to avoid similar breakage for other Arch users, and perhaps
others?)
BTW2, if you are using pacman on arch, this little snippet is handy to list
what your last pacman {up,down}grade did:
I do listing variants as ls-whatever -- this one is ls-pacupd:
--8<---cut here---start->8---
#!/usr/bin/bash
# ~/bin/ls-pacupd -- list latest pacman Syu upgrades
latest="$(stat -c '%y' /var/log/pacman.log|cut -d ' ' -f1)"
egrep "$latest.* (up|down)graded " /var/log/pacman.log
--8<---cut here---end--->8---
I found that the guix-installed version of mutt worked for getting mail,
and saw that it used the prior version.
However, emacs is mutt's configured editor, and after some longish time
editing
the entire system would freeze and not respond to ANY key input, and I had
to
power down physically (5-sec press of power button).
So I had to go back to the old Arch version.
I am still mystified by this freeze-up. It's possible that I am typing some
fatal
combination of keys on this keyboard or that my migration from a dying
laptop to
an SSD in a USB3 cassette booted with UEFI on a Lenovo Swift did not
entirely succeed.
My context:
I am running on tty1 with guix "disabled" by not setting up its paths etc
in
~/.bash_profile at login, so this is my current boot context here:
┌─┐
│ Booted at 2020-01-02 08:50 -0800 (PST) and logged in as as
bokr@Evo25c2ArchGx4 │
├─┤
│ HW host: Acer Swift SF113-31/ASAHI_AP_S, BIOS V1.08 11/22/2017
│
│ MOUNTPOINT KNAMELABELSIZE FSAVAIL FSUSE%
bug#38709: Recursive crate importer ignores specified version
On 2019-12-22 18:56, Valentin Ignatev wrote: > When I'm trying to recursively import specific crate version, it > imports the latest version instead: > > guix import crate --recursive notify@4.0.14 gives this output: > following redirection to > `https://static.crates.io/crates/notify/notify-5.0.0-pre.1.crate'... > following redirection to > `https://static.crates.io/crates/anymap/anymap-0.12.1.crate'... > following redirection to > `https://static.crates.io/crates/bitflags/bitflags-1.2.1.crate'... > > And so on, so it follows notify-5.0.0 instead of 4.0.14 > > Regards, > Valentin Ignatev Currently the recusive importer doesn't respect version numbers. Progress is being made on this in issue 38408 https://issues.guix.gnu.org/issue/38408 -Martin Becze
bug#37266: gnupg/pinentry not aware of each other
On 2019-09-02 04:56, j...@dorsal.tk wrote: > $ gpg --full-generate-key > > Change (N)ame, (C)omment, (E)mail or (O)kay/(Q)uit? o > We need to generate a lot of random bytes. It is a good idea to perform > some other action (type on the keyboard, move the mouse, utilize the > disks) during the prime generation; this gives the random number > generator a better chance to gain enough entropy. > gpg: agent_genkey failed: No pinentry > Key generation failed: No pinentry > $ which pinentry > /run/current-system/profile/bin/pinentry try added `pinentry-program /home/USER/.guix-profile/bin/pinentry-gtk-2` to `~/.gnupg/gpg-agent.conf` let me know if that works. I also encountered this.
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
ping :P
> ---
>
> gnu/packages/gnome.scm | 19 ++-
> 1 file changed, 10 insertions(+), 9 deletions(-)
>
> diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
> index 1a091de938..ef6d78bb5a 100644
> --- a/gnu/packages/gnome.scm
> +++ b/gnu/packages/gnome.scm
> @@ -8105,19 +8105,20 @@ generic enough to work for everyone.")
> #:phases
> (modify-phases %standard-phases
> (add-after
> - 'install 'wrap-program
> - (lambda* (#:key inputs outputs #:allow-other-keys)
> - (let ((out (assoc-ref outputs "out")))
> - (wrap-program (string-append out "/bin/evolution")
> -`("GSETTINGS_SCHEMA_DIR" = (,(string-append out
> "/share/glib-2.0/schemas/")))
> -`("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
> - #t)))
> + 'install 'wrap-program
> + (lambda* (#:key inputs outputs #:allow-other-keys)
> + (let ((out (assoc-ref outputs "out")))
> + (wrap-program (string-append out "/bin/evolution")
> + `("XDG_DATA_DIRS" prefix (,(getenv "XDG_DATA_DIRS")))
> + `("GSETTINGS_SCHEMA_DIR" =
> + (,(string-append out "/share/glib-2.0/schemas/")
> + #t)))
> #:configure-flags
> (list "-DENABLE_PST_IMPORT=OFF"; libpst is not packaged
> "-DENABLE_LIBCRYPTUI=OFF"))) ; libcryptui hasn't seen a release
> - ; in four years and cannot be
> built.
> +; in four years and cannot be built.
> (native-inputs
> - `(("glib" ,glib "bin") ; glib-mkenums
> + `(("glib" ,glib "bin") ; glib-mkenums
> ("pkg-config" ,pkg-config)
> ("intltool" ,intltool)
> ("itstool" ,itstool)))
> --
> 2.23.0
On 9/6/19 1:49 PM, Martin Becze wrote:
> Yeah I don't know. I just used prefix in the latest patch. Let me know
> if the other method is preferred.
>
> On 9/6/19 12:30 PM, Tobias Geerinckx-Rice wrote:
>> Martin,
>>
>> Martin Becze 写道：
>>> Prefix appends ${MY_VAR:+:}$MY_VAR to the env var, so that seems nice.
>> Sure, that's what I'd expect, but why is it preferred over keeping any
>> user-defined entries first?
>>
>> Kind regards,
>>
>> T G-R
signature.asc
Description: OpenPGP digital signature
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
Yeah I don't know. I just used prefix in the latest patch. Let me know
if the other method is preferred.
On 9/6/19 12:30 PM, Tobias Geerinckx-Rice wrote:
> Martin,
>
> Martin Becze 写道：
>> Prefix appends ${MY_VAR:+:}$MY_VAR to the env var, so that seems nice.
>
> Sure, that's what I'd expect, but why is it preferred over keeping any
> user-defined entries first?
>
> Kind regards,
>
> T G-R
signature.asc
Description: OpenPGP digital signature
bug#36896: [PATCH v2] added prefix to the wrapping of XDG_DATA_DIRS for evolution
---
gnu/packages/gnome.scm | 19 ++-
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
index 1a091de938..ef6d78bb5a 100644
--- a/gnu/packages/gnome.scm
+++ b/gnu/packages/gnome.scm
@@ -8105,19 +8105,20 @@ generic enough to work for everyone.")
#:phases
(modify-phases %standard-phases
(add-after
- 'install 'wrap-program
- (lambda* (#:key inputs outputs #:allow-other-keys)
- (let ((out (assoc-ref outputs "out")))
- (wrap-program (string-append out "/bin/evolution")
-`("GSETTINGS_SCHEMA_DIR" = (,(string-append out
"/share/glib-2.0/schemas/")))
-`("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
- #t)))
+ 'install 'wrap-program
+ (lambda* (#:key inputs outputs #:allow-other-keys)
+ (let ((out (assoc-ref outputs "out")))
+ (wrap-program (string-append out "/bin/evolution")
+ `("XDG_DATA_DIRS" prefix (,(getenv "XDG_DATA_DIRS")))
+ `("GSETTINGS_SCHEMA_DIR" =
+ (,(string-append out "/share/glib-2.0/schemas/")
+ #t)))
#:configure-flags
(list "-DENABLE_PST_IMPORT=OFF"; libpst is not packaged
"-DENABLE_LIBCRYPTUI=OFF"))) ; libcryptui hasn't seen a release
- ; in four years and cannot be built.
+; in four years and cannot be built.
(native-inputs
- `(("glib" ,glib "bin") ; glib-mkenums
+ `(("glib" ,glib "bin") ; glib-mkenums
("pkg-config" ,pkg-config)
("intltool" ,intltool)
("itstool" ,itstool)))
--
2.23.0
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 9/4/19 2:17 PM, Tobias Geerinckx-Rice wrote:
> Martin,
>
> Martin Becze 写道：
>> hmm maybe i did something wrong when sending with "git-sendemail" it
>> showed up here https://issues.guix.info/issue/36896
>
> Strange… It can't find it in my mailbox either.
>
> Is there a commit message?
>
>> `("GSETTINGS_SCHEMA_DIR" = (,(string-append out
>> "/share/glib-2.0/schemas/")))
>
> Please wrap long lines like these at 80 characters. I'd do so after ‘=’.
>
>> + `("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
>
> I think this is something we want to append (":" suffix), not override
> outright (=), no?
>
> Grepping through Guix, I see that ‘prefix’ is popular too, but I'm not
> sure why. Does that still grant users the same control over the
> result? (This question for extra credit only :-)
>
> Thanks, and sorry for the wait,
>
> T G-R
Prefix appends ${MY_VAR:+:}$MY_VAR to the env var, so that seems nice.
signature.asc
Description: OpenPGP digital signature
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 2019-09-03 11:59, Ricardo Wurmus wrote:
> Hi Martin,
>
>>>> I sent a new patch in (using git sendemail) but it didn't thread quite
>>>> correctly.
>>>
>>> bumping! I have a few more evolution patches but need to get this one in
>>> first
>>
>> bumping!
>
> Sorry for the delay! Where can I find the new patch that you’ve sent?
> I don’t see it in your previous emails.
hmm maybe i did something wrong when sending with "git-sendemail" it
showed up here https://issues.guix.info/issue/36896
Anyways here is the patch!
---
gnu/packages/gnome.scm | 15 ++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
index d5e46e09c1..1a091de938 100644
--- a/gnu/packages/gnome.scm
+++ b/gnu/packages/gnome.scm
@@ -39,6 +39,7 @@
;;; Copyright © 2019 Giacomo Leidi
;;; Copyright © 2019 Jelle Licht
;;; Copyright © 2019 Jonathan Frederickson
+;;; Copyright © 2019 Martin Becze
;;;
;;; This file is part of GNU Guix.
;;;
@@ -8100,7 +8101,18 @@ generic enough to work for everyone.")
"0sdv5lg2vlz5f4raymz9d8a5jq4j18vbqyigaip6508p3bjnfj8l"
(build-system cmake-build-system)
(arguments
- `(#:configure-flags
+ `(
+ #:phases
+ (modify-phases %standard-phases
+ (add-after
+ 'install 'wrap-program
+ (lambda* (#:key inputs outputs #:allow-other-keys)
+ (let ((out (assoc-ref outputs "out")))
+ (wrap-program (string-append out "/bin/evolution")
+`("GSETTINGS_SCHEMA_DIR" = (,(string-append out
"/share/glib-2.0/schemas/")))
+`("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
+ #t)))
+ #:configure-flags
(list "-DENABLE_PST_IMPORT=OFF"; libpst is not packaged
"-DENABLE_LIBCRYPTUI=OFF"))) ; libcryptui hasn't seen a
release
; in four years and cannot be
built.
@@ -8113,6 +8125,7 @@ generic enough to work for everyone.")
`(("enchant" ,enchant)
("evolution-data-server" ,evolution-data-server) ; must be the
same version
("gcr" ,gcr)
+ ("gsettings-desktop-schemas" ,gsettings-desktop-schemas)
("gnome-autoar" ,gnome-autoar)
("gnome-desktop" ,gnome-desktop)
("gtkspell3" ,gtkspell3)
--
2.22.0
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 2019-08-27 04:23, Martin Becze wrote: > On 2019-08-20 09:04, Martin Becze wrote: >> On 2019-08-05 16:30, Tobias Geerinckx-Rice wrote: >>> Martin, Ricardo, >>> >>> I agree with Ricardo (here and in matters of fonts) that propagation >>> is to be avoided at all reasonable costs, so… >>> >>> Martin Becze 写道： >>>> The terminator packagage propagates gsetting-desktop-schema as well >>>> but maybe its also doing the wrong thing? >>> >>> …that was a mistake, in retrospect, fixed in >>> 96681d4be101c771fafd4257aca471685119fedd. >>> >>> You can probably apply that same fix directly to Evolution. There >>> should be no need to propagate anything. >>> >>> Thanks! >>> >>> T G-R >> >> bumping! :P >> I sent a new patch in (using git sendemail) but it didn't thread quite >> correctly. > > bumping! I have a few more evolution patches but need to get this one in > first bumping!
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 2019-08-20 09:04, Martin Becze wrote: > On 2019-08-05 16:30, Tobias Geerinckx-Rice wrote: >> Martin, Ricardo, >> >> I agree with Ricardo (here and in matters of fonts) that propagation >> is to be avoided at all reasonable costs, so… >> >> Martin Becze 写道： >>> The terminator packagage propagates gsetting-desktop-schema as well >>> but maybe its also doing the wrong thing? >> >> …that was a mistake, in retrospect, fixed in >> 96681d4be101c771fafd4257aca471685119fedd. >> >> You can probably apply that same fix directly to Evolution. There >> should be no need to propagate anything. >> >> Thanks! >> >> T G-R > > bumping! :P > I sent a new patch in (using git sendemail) but it didn't thread quite > correctly. bumping! I have a few more evolution patches but need to get this one in first.
bug#36632: ungoogled-chromium needs font-gnu-freefont-ttf
On 2019-08-05 15:36, Ricardo Wurmus wrote: > Hi Martin, > >> ungoogled-chromium will crash until font-gnu-freefont-ttf is installed. > > Thank you for reporting this problem and proposing a patch. > > I’d prefer not to propagate any fonts. Do you happen to know what the > mechanism underlying the fix is? Why does this fix the problem? Why > does it crash in the first place? > > Is this at all related to a stale font cache, perhaps? So i got a chance to dig into this a bit. chromium needs a ttf font and there are no fonts installed yet, so defaintly not a stale cache issue. Chrome is crashing here https://chromium.googlesource.com/chromium/src/+/master/ui/gfx/platform_font_skia.cc#97 Its using fontconfig i think. I have no idea how to get chromium to read fonts from a custom path.
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 2019-08-05 16:30, Tobias Geerinckx-Rice wrote: > Martin, Ricardo, > > I agree with Ricardo (here and in matters of fonts) that propagation > is to be avoided at all reasonable costs, so… > > Martin Becze 写道： >> The terminator packagage propagates gsetting-desktop-schema as well >> but maybe its also doing the wrong thing? > > …that was a mistake, in retrospect, fixed in > 96681d4be101c771fafd4257aca471685119fedd. > > You can probably apply that same fix directly to Evolution. There > should be no need to propagate anything. > > Thanks! > > T G-R bumping! :P I sent a new patch in (using git sendemail) but it didn't thread quite correctly.
bug#36896: [PATCH] wrapped evolution so that it know where its gsettings are
---
gnu/packages/gnome.scm | 15 ++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
index d5e46e09c1..1a091de938 100644
--- a/gnu/packages/gnome.scm
+++ b/gnu/packages/gnome.scm
@@ -39,6 +39,7 @@
;;; Copyright © 2019 Giacomo Leidi
;;; Copyright © 2019 Jelle Licht
;;; Copyright © 2019 Jonathan Frederickson
+;;; Copyright © 2019 Martin Becze
;;;
;;; This file is part of GNU Guix.
;;;
@@ -8100,7 +8101,18 @@ generic enough to work for everyone.")
"0sdv5lg2vlz5f4raymz9d8a5jq4j18vbqyigaip6508p3bjnfj8l"
(build-system cmake-build-system)
(arguments
- `(#:configure-flags
+ `(
+ #:phases
+ (modify-phases %standard-phases
+ (add-after
+ 'install 'wrap-program
+ (lambda* (#:key inputs outputs #:allow-other-keys)
+ (let ((out (assoc-ref outputs "out")))
+ (wrap-program (string-append out "/bin/evolution")
+`("GSETTINGS_SCHEMA_DIR" = (,(string-append out
"/share/glib-2.0/schemas/")))
+`("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
+ #t)))
+ #:configure-flags
(list "-DENABLE_PST_IMPORT=OFF"; libpst is not packaged
"-DENABLE_LIBCRYPTUI=OFF"))) ; libcryptui hasn't seen a release
; in four years and cannot be built.
@@ -8113,6 +8125,7 @@ generic enough to work for everyone.")
`(("enchant" ,enchant)
("evolution-data-server" ,evolution-data-server) ; must be the same
version
("gcr" ,gcr)
+ ("gsettings-desktop-schemas" ,gsettings-desktop-schemas)
("gnome-autoar" ,gnome-autoar)
("gnome-desktop" ,gnome-desktop)
("gtkspell3" ,gtkspell3)
--
2.22.0
bug#36896: [PATCH] wrapped evolution so that it know where its gsettings are
---
gnu/packages/gnome.scm | 15 ++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
index d5e46e09c1..1a091de938 100644
--- a/gnu/packages/gnome.scm
+++ b/gnu/packages/gnome.scm
@@ -39,6 +39,7 @@
;;; Copyright © 2019 Giacomo Leidi
;;; Copyright © 2019 Jelle Licht
;;; Copyright © 2019 Jonathan Frederickson
+;;; Copyright © 2019 Martin Becze
;;;
;;; This file is part of GNU Guix.
;;;
@@ -8100,7 +8101,18 @@ generic enough to work for everyone.")
"0sdv5lg2vlz5f4raymz9d8a5jq4j18vbqyigaip6508p3bjnfj8l"
(build-system cmake-build-system)
(arguments
- `(#:configure-flags
+ `(
+ #:phases
+ (modify-phases %standard-phases
+ (add-after
+ 'install 'wrap-program
+ (lambda* (#:key inputs outputs #:allow-other-keys)
+ (let ((out (assoc-ref outputs "out")))
+ (wrap-program (string-append out "/bin/evolution")
+`("GSETTINGS_SCHEMA_DIR" = (,(string-append out
"/share/glib-2.0/schemas/")))
+`("XDG_DATA_DIRS" = (,(getenv "XDG_DATA_DIRS")
+ #t)))
+ #:configure-flags
(list "-DENABLE_PST_IMPORT=OFF"; libpst is not packaged
"-DENABLE_LIBCRYPTUI=OFF"))) ; libcryptui hasn't seen a release
; in four years and cannot be built.
@@ -8113,6 +8125,7 @@ generic enough to work for everyone.")
`(("enchant" ,enchant)
("evolution-data-server" ,evolution-data-server) ; must be the same
version
("gcr" ,gcr)
+ ("gsettings-desktop-schemas" ,gsettings-desktop-schemas)
("gnome-autoar" ,gnome-autoar)
("gnome-desktop" ,gnome-desktop)
("gtkspell3" ,gtkspell3)
--
2.22.0
bug#36896: [PATCH] added gsettings-desktop-schema to progragated inputs
On 2019-08-05 07:40, Ricardo Wurmus wrote:
> Hi,
>
> null writes:
>
>> ---
>> gnu/packages/gnome.scm | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/gnu/packages/gnome.scm b/gnu/packages/gnome.scm
>> index 3f7bbac214..c33653f361 100644
>> --- a/gnu/packages/gnome.scm
>> +++ b/gnu/packages/gnome.scm
>> @@ -8074,6 +8074,8 @@ generic enough to work for everyone.")
>> ("openldap" ,openldap)
>> ("webkitgtk" ,webkitgtk)
>> ("ytnef" ,ytnef)))
>> +(propagated-inputs
>> + `(("gsettings-desktop-schemas" ,gsettings-desktop-schemas)))
>> (home-page "https://gitlab.gnome.org/GNOME/evolution;)
>> (synopsis "Manage your email, contacts and schedule")
>> (description "Evolution is a personal information management application
>
> Thank you for the patch.
>
> I’m not sure this is the right thing to do. I don’t think we ever
> propagate this package. Instead we often wrap executables so that
> relevant environment variables are set. Would this work in the case of
> Evolution as well?
The terminator packagage propagates gsetting-desktop-schema as well
https://github.com/guix-mirror/guix/blob/8126142baf2cadd7b0052ba7b1d0cd248c8cbc40/gnu/packages/gnome.scm#L8159-L8161
but maybe its also doing the wrong thing? I'll look in to how to wrap
the executable and see if that will work.
bug#36632: (no subject)
bump :)
bug#36896: Evolution needs gsettings-desktop-schemas
Evolution will crash immediate if gsettings-desktop-schemas is not also installed
bug#36632: ungoogled-chromium needs font-gnu-freefont-ttf
ungoogled-chromium will crash until font-gnu-freefont-ttf is installed. signature.asc Description: OpenPGP digital signature
bug#36590: its done
its done
