bug#31109: Retroarch: source tarball has wrong hash

[Ludovic Courtès]

Hello,

I get:

--8<---cut here---start->8---
$ guix build 
/gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv
The following derivation will be built:
   /gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv
@ build-started 
/gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv - 
i686-linux 
/var/log/guix/drvs/cs//v3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv.gz

Starting download of 
/gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz
>From https://github.com/libretro/RetroArch/archive/v1.7.1.tar.gz...
following redirection to 
`https://codeload.github.com/libretro/RetroArch/tar.gz/v1.7.1'...
 v1.7.1.tar.gz  19.6MiB   9.6MiB/s 00:02 [##] 100.0%
sha256 hash mismatch for output path 
`/gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz'
  expected: 0fdribjfc5zz9brzhqcxw6m76kvyg13l67aiigszv4wsjd5j3gpz
  actual:   1wykv0w0kdlh7lh0k1ig0lpk5vh4c7r19jlfa9103jmjlryrq679
@ build-failed 
/gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv - 1 
sha256 hash mismatch for output path 
`/gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz'
  expected: 0fdribjfc5zz9brzhqcxw6m76kvyg13l67aiigszv4wsjd5j3gpz
  actual:   1wykv0w0kdlh7lh0k1ig0lpk5vh4c7r19jlfa9103jmjlryrq679
--8<---cut here---end--->8---

Looking at
, it
seems that this package never built on Hydra.

Rutger, could you take a look and determine if it’s just a mistake or if
the GitHub-generated tarball changed?

TIA!

Ludo’.

bug#31108: Julia: ‘objconv.zip’ modified in place on www.agner.org

[Ludovic Courtès]

The ‘objconv.zip’ file referred to by ‘julia’ was modified in place:

--8<---cut here---start->8---
$ guix build /gnu/store/p9sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv
The following derivation will be built:
   /gnu/store/p9sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv
@ build-started /gnu/store/p9sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv - 
i686-linux 
/var/log/guix/drvs/p9//sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv.gz

Starting download of /gnu/store/2fsjkc8ipqvy2f392m842715h19qdkd7-objconv.zip
>From http://www.agner.org/optimize/objconv.zip...
 objconv.zip  1009KiB 1.3MiB/s 00:01 [##] 100.0%
sha256 hash mismatch for output path 
`/gnu/store/2fsjkc8ipqvy2f392m842715h19qdkd7-objconv.zip'
  expected: 1fi7qa2sd9vb35dvkgripjf0fayzg2qmff215f8agfqfiwd1g8qs
  actual:   195p7dbjqfd6vca9mjj7i0ifpscsrm9njyfmzrpyf9zjdz6w9h7j
@ build-failed /gnu/store/p9sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv - 1 
sha256 hash mismatch for output path 
`/gnu/store/2fsjkc8ipqvy2f392m842715h19qdkd7-objconv.zip'
  expected: 1fi7qa2sd9vb35dvkgripjf0fayzg2qmff215f8agfqfiwd1g8qs
  actual:   195p7dbjqfd6vca9mjj7i0ifpscsrm9njyfmzrpyf9zjdz6w9h7j
guix build: error: build failed: build of 
`/gnu/store/p9sy48pip38gk1c72c5k7wsphppj07vk-objconv.zip.drv' failed
--8<---cut here---end--->8---

Is there a versioned URL for this?

Ludo’.

bug#31109: Retroarch: source tarball has wrong hash

[Rutger Helling]

Hi Ludo, 

The tarball changed. I ran a diff and couldn't find anything malicious
in the changes so I updated the checksum just now. 

On 2018-04-09 10:44, l...@gnu.org wrote:

> Hello,
> 
> I get:
> 
> --8<---cut here---start->8---
> $ guix build 
> /gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv
> The following derivation will be built:
> /gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv
> @ build-started 
> /gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv - 
> i686-linux 
> /var/log/guix/drvs/cs//v3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv.gz
> 
> Starting download of 
> /gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz
> From https://github.com/libretro/RetroArch/archive/v1.7.1.tar.gz...
> following redirection to 
> `https://codeload.github.com/libretro/RetroArch/tar.gz/v1.7.1'...
> v1.7.1.tar.gz  19.6MiB   9.6MiB/s 00:02 [##] 
> 100.0%
> sha256 hash mismatch for output path 
> `/gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz'
> expected: 0fdribjfc5zz9brzhqcxw6m76kvyg13l67aiigszv4wsjd5j3gpz
> actual:   1wykv0w0kdlh7lh0k1ig0lpk5vh4c7r19jlfa9103jmjlryrq679
> @ build-failed 
> /gnu/store/csv3ibcqacvk1l6np1f6ch5f62gz1yw0-retroarch-1.7.1.tar.gz.drv - 1 
> sha256 hash mismatch for output path 
> `/gnu/store/pddz1nyw4rni60mqha4dp02rk4c9gv36-retroarch-1.7.1.tar.gz'
> expected: 0fdribjfc5zz9brzhqcxw6m76kvyg13l67aiigszv4wsjd5j3gpz
> actual:   1wykv0w0kdlh7lh0k1ig0lpk5vh4c7r19jlfa9103jmjlryrq679
> --8<---cut here---end--->8---
> 
> Looking at
> , it
> seems that this package never built on Hydra.
> 
> Rutger, could you take a look and determine if it's just a mistake or if
> the GitHub-generated tarball changed?
> 
> TIA!
> 
> Ludo'.

bug#31109: Retroarch: source tarball has wrong hash

[Ludovic Courtès]

Hello,

Rutger Helling  skribis:

> The tarball changed. I ran a diff and couldn't find anything malicious
> in the changes so I updated the checksum just now. 

Out of curiosity, could you post the changes as reported by running
‘diff -ur --no-dereference’ on the extracted directories, or
‘diffoscope’?  (The latter takes file metadata into account.)

Thanks,
Ludo’.

bug#31113: mcron ERROR: In procedure mktime:

[George myglc2 Clemmer]

Hello,

somewhere between March ~17 (commit be5ed1421 gnu: Add terraform-docs.)
and April ~5 (commit cfb55cfc4 gnu: Add java-native-access-platform.)
mcron-service stopped working and started throwing errors like ...

#+BEGIN_EXAMPLE
ERROR: In procedure mktime:
Wrong type (expecting exact integer): 45.0
Backtrace:
   4 (primitive-load "/gnu/store/422ngw911hz62510b33wdvrp1a6?")
   In mcron/base.scm:
  103:25  3 (add-job # ?)
   In mcron/job-specifier.scm:
  251:32  2 (_ 1523292689)
   In mcron/vixie-time.scm:
  381:15  1 (_ _)
   In unknown file:
  0 (mktime #(0 45.0 +inf.0 9 3 118 1 98 1 14400 "EDT") #)
ERROR: In procedure mktime:
...
#+END_EXAMPLE

Please see system config attached (my /etc/gen/x/010/sys.scm)

TIA - George

bug#31113: mcron ERROR: In procedure mktime:

[George myglc2 Clemmer]

Oops, forgot to attach the system config (actually my /etc/gen/x/016/sys.scm)



sys.scm
Description: Binary data

bug#31117: eaae07ec2 breaks 'guix pull' for guile 2.0 users

[Alex Vong]

Hello Guix,

Commit eaae07ec2878707744fed76029a28ab9a494fc2e breaks 'guix pull' for
guile 2.0 users. More precisely, it fails with:


alexvong1995@debian:~$ guix pull 
Updating from Git repository at 'https://git.savannah.gnu.org/git/guix.git'...
Building from Git commit 60e1de6d95bd32b4996c199708541781b8f828fd...
Computing Guix derivation for 'x86_64-linux'... /Backtrace:
   7 (primitive-load "/gnu/store/75vknf0m8z0xdqnbbbms93r94j8…")
In ice-9/boot-9.scm:
142:2  6 (dynamic-wind _ _ #)
In ice-9/eval.scm:
155:9  5 (_ _)
   191:35  4 (_ _)
   173:47  3 (_ #(#(# # …)))
In ./guix/self.scm:
588:4  2 (guix-derivation "/gnu/store/v6gspcm5462rf6c4rlkcjsyr4…" …)
   575:24  1 (guile-for-build _)
 80:4  0 (_ _)

./guix/self.scm:80:4: Throw to key `match-error' with args `("match" "no 
matching pattern" "guile@2.0")'.
Backtrace:
In ice-9/boot-9.scm:
 160: 14 [catch #t # ...]
In unknown file:
   ?: 13 [apply-smob/1 #]
In ice-9/boot-9.scm:
  66: 12 [call-with-prompt prompt0 ...]
In ice-9/eval.scm:
 432: 11 [eval # #]
In ice-9/boot-9.scm:
2404: 10 [save-module-excursion #]
4056: 9 [#]
In guix/ui.scm:
1501: 8 [run-guix-command pull]
In ice-9/boot-9.scm:
 160: 7 [catch srfi-34 # ...]
 160: 6 [catch system-error ...]
 160: 5 [catch git-error ...]
In guix/scripts/pull.scm:
 280: 4 [#]
In guix/store.scm:
1443: 3 [run-with-store # ...]
In guix/scripts/pull.scm:
 178: 2 [# #]
In ice-9/eval.scm:
 386: 1 [eval # <1>) <0>)> (# #t # ...)]
In unknown file:
   ?: 0 [scm-error misc-error #f ...]

ERROR: In procedure scm-error:
ERROR: invalid build result (# 
/gnu/store/75vknf0m8z0xdqnbbbms93r94j8y243f-compute-guix-derivation 
559422aa07d0> "")



The relevant error is:


./guix/self.scm:80:4: Throw to key `match-error' with args `("match" "no 
matching pattern" "guile@2.0")'.


I think the error is caused by the pattern matching code not taking into
account the version suffix. Also, I have used 'guix pull --commit ...'
to make sure it is really eaae07ec2 causing the problem.

Cheers,
Alex


signature.asc
Description: PGP signature

bug#27284: ‘guix pull’ builds using multiple derivations

[Ricardo Wurmus]

Hi Ludo,

> After quite a bit of tweaking I’ve finally pushed this patch series to
> master.  So if you run ‘guix pull’ now, you’ll get the new
> multiple-derivation build.

This is really great!  I’m very happy that this is finally in the master
branch.  Thank you!

As discussed on IRC, there is a problem when running “guix pull” with an
older version of Guix that is still using Guile 2.0, which seems to be
adequately fixed by your patch:

--8<---cut here---start->8---
diff --git a/guix/self.scm b/guix/self.scm
index c9e4a4250..c19b0f6fb 100644
--- a/guix/self.scm
+++ b/guix/self.scm
@@ -88,7 +88,11 @@ GUILE-VERSION (\"2.0\" or \"2.2\"), or #f if none of the 
packages matches."
   ("bzip2"  (ref '(gnu packages compression) 'bzip2))
   ("xz" (ref '(gnu packages compression) 'xz))
   ("guix"   (ref '(gnu packages package-management)
- 'guix-register)
+ 'guix-register))
+  ("guile2.0-json" (ref '(gnu packages guile) 'guile2.0-json))
+  ("guile2.0-ssh"  (ref '(gnu packages ssh) 'guile2.0-ssh))
+  ("guile2.0-git"  (ref '(gnu packages guile) 'guile2.0-git))
+  (_   #f ;no such package

 
;;;
--8<---cut here---end--->8---

A remaining problem for me was that “(guix docker)” could not be built
because “(json)” was supposedly not available.

None of this seems to be a problem when using Guile 2.2.

>   • The “guix-packages” derivation is coarse-grain.  We could perhaps
> automatically split it in clusters of modules by analyzing the (gnu
> packages …) module graph.

I think we should do this anyway and try to split up large modules and
reduce inter-module dependencies.

> This works unlocks some of the stuff in , so
> I’ll try to focus on that now.

Exciting!

--
Ricardo

bug#31109: Retroarch: source tarball has wrong hash

[Ludovic Courtès]

Hi Rutger,

Rutger Helling  skribis:

> Here's what diffoscope gave me (old vs. new).

OK, thanks for sharing.

So it looks like they modified the “1.7.1” tag upstream, leading to a
different tarball.

I’m appending the actual changes below (extracted from what you sent.)
Basically small adjustments.

Case closed, thanks!

Ludo’.

─ RetroArch-1.7.1/frontend/drivers/platform_xdk.c
@@ -115,14 +115,27 @@
  g_defaults.dirs[DEFAULT_DIR_CORE],
  "system",
  sizeof(g_defaults.dirs[DEFAULT_DIR_SYSTEM]));
fill_pathname_join(g_defaults.dirs[DEFAULT_DIR_SCREENSHOT],
  g_defaults.dirs[DEFAULT_DIR_CORE],
  "screenshots",
  sizeof(g_defaults.dirs[DEFAULT_DIR_SCREENSHOT]));
+   fill_pathname_join(g_defaults.dirs[DEFAULT_DIR_OVERLAY],
+ g_defaults.dirs[DEFAULT_DIR_CORE],
+ "overlays",
+ sizeof(g_defaults.dirs[DEFAULT_DIR_OVERLAY]));
+   fill_pathname_join(g_defaults.dirs[DEFAULT_DIR_ASSETS],
+ g_defaults.dirs[DEFAULT_DIR_CORE],
+ "media", sizeof(g_defaults.dirs[DEFAULT_DIR_ASSETS]));
+   fill_pathname_join(g_defaults.dirs[DEFAULT_DIR_CORE_ASSETS],
+ g_defaults.dirs[DEFAULT_DIR_CORE],
+ "downloads", sizeof(g_defaults.dirs[DEFAULT_DIR_CORE_ASSETS]));
+   fill_pathname_join(g_defaults.dirs[DEFAULT_DIR_PLAYLIST],
+ g_defaults.dirs[DEFAULT_DIR_CORE],
+ "playlists", sizeof(g_defaults.dirs[DEFAULT_DIR_PLAYLIST]));
 #elif defined(_XBOX360)
strlcpy(g_defaults.dirs[DEFAULT_DIR_CORE],
  "game:",
  sizeof(g_defaults.dirs[DEFAULT_DIR_CORE]));
strlcpy(g_defaults.path.config,
  "game:\\retroarch.cfg", sizeof(g_defaults.path.config));
strlcpy(g_defaults.dirs[DEFAULT_DIR_SCREENSHOT],
─ RetroArch-1.7.1/gfx/drivers_tracker/video_state_python.c
@@ -124,15 +124,15 @@
if (user > MAX_USERS || user < 1 || key >= RARCH_FIRST_META_KEY)
   return NULL;
 
joypad_info.joy_idx= settings->uints.input_joypad_map[user - 1];
joypad_info.auto_binds = input_autoconf_binds[joypad_info.joy_idx];
 
if (!input_driver_is_libretro_input_blocked())
-  res = current_input->input_state(current_input_data, joypad_info,
+  res = input_driver_input_state(joypad_info,
 py_binds,
 user - 1, RETRO_DEVICE_JOYPAD, 0, key);
return PyBool_FromLong(res);
 }
 
 static PyObject *py_read_analog(PyObject *self, PyObject *args)
 {
@@ -152,15 +152,15 @@
 
if (user > MAX_USERS || user < 1 || index > 1 || id > 1)
   return NULL;
 
joypad_info.joy_idx= settings->uints.input_joypad_map[user - 1];
joypad_info.auto_binds = input_autoconf_binds[joypad_info.joy_idx];
 
-   res = current_input->input_state(current_input_data,
+   res = input_driver_input_state(
  joypad_info, py_binds,
  user - 1, RETRO_DEVICE_ANALOG, index, id);
return PyFloat_FromDouble((double)res / 0x7fff);
 }
 
 static PyMethodDef RarchMethods[] = {
{ "read_wram",py_read_wram,   METH_VARARGS, "Read WRAM from system." },
─ RetroArch-1.7.1/input/input_driver.c
@@ -1031,14 +1031,26 @@
   RETRO_DEVICE_KEYBOARD, 0, ids[i][0]))
 BIT256_SET_PTR(p_new_state, ids[i][1]);
   }
}
 }
 #endif
 
+int16_t input_driver_input_state(
+ rarch_joypad_info_t joypad_info,
+ const struct retro_keybind **retro_keybinds,
+ unsigned port, unsigned device, unsigned index, unsigned id)
+{
+   if (current_input && current_input->input_state)
+  return current_input->input_state(current_input_data, joypad_info,
+retro_keybinds,
+port, device, index, id);
+   return 0;
+}
+
 /**
  * input_keys_pressed:
  *
  * Grab an input sample for this frame.
  *
  * Returns: Input sample containing a mask of all pressed keys.
  */
─ RetroArch-1.7.1/input/input_driver.h
@@ -403,14 +403,19 @@
 
 bool input_driver_init_mapper(void);
 
 bool input_driver_grab_mouse(void);
 
 bool input_driver_ungrab_mouse(void);
 
+int16_t input_driver_input_state(
+ rarch_joypad_info_t joypad_info,
+ const struct retro_keybind **retro_keybinds,
+ unsigned port, unsigned device, unsigned index, unsigned id);
+
 float *input_driver_get_float(enum input_action action);
 
 unsigned *input_driver_get_uint(enum input_action action);
 
 bool input_driver_is_data_ptr_same(void *data);
 
 /**
─ RetroArch-1.7.1/pkg/ctr/Makefile.cores
@@ -1,270 +1,277 @@
+ifeq ($(LIBRETRO), 2048)
+	APP_TITLE= 2048 Libretro
+	APP_PRODUCT_CODE = RARCH-2048
+	APP_UNIQUE_ID= 0xBAC0B
+	APP_ICON = pkg/ctr/assets/2048.png
+	APP_BANNER   = pkg/ctr/assets/2048_banner.png
 
-ifeq ($(LIBRETRO), gambatte)
-	APP_TITLE   = Gambatte Libretro
-	APP_PRODUCT_CODE= RARCH-GAMBATTE
-	APP_UNIQUE_ID   = 0xBAC01
-	APP_ICON= pkg/ctr/assets/gambatte.png
-	APP_BANNER  = pkg/ctr/assets/gambatte_banner.png
+else ifeq ($(LIBRETRO), 81)
+	APP_TITLE= lr-81
+	APP_AUTHOR 

bug#31113: mcron ERROR: In procedure mktime:

[Maxim Cournoyer]

Hi George,

George myglc2 Clemmer  writes:

> Hello,
>
> somewhere between March ~17 (commit be5ed1421 gnu: Add terraform-docs.)
> and April ~5 (commit cfb55cfc4 gnu: Add java-native-access-platform.)
> mcron-service stopped working and started throwing errors like ...
>
> #+BEGIN_EXAMPLE
> ERROR: In procedure mktime:
> Wrong type (expecting exact integer): 45.0
> Backtrace:
>4 (primitive-load "/gnu/store/422ngw911hz62510b33wdvrp1a6?")
>In mcron/base.scm:
>   103:25  3 (add-job # ?)
>In mcron/job-specifier.scm:
>   251:32  2 (_ 1523292689)
>In mcron/vixie-time.scm:
>   381:15  1 (_ _)
>In unknown file:
>   0 (mktime #(0 45.0 +inf.0 9 3 118 1 98 1 14400 "EDT") #)
> ERROR: In procedure mktime:
> ...
> #+END_EXAMPLE
>
> Please see system config attached (my /etc/gen/x/010/sys.scm)
>
> TIA - George

It's of not much help I'm afraid, but FWIW, I'm also having issues since
I reconfigure a fresh guix from yesterday:

--8<---cut here---start->8---
Backtrace:
   8 (apply-smob/1 #)
In mcron/scripts/mcron.scm:
 81:2  7 (main _)
In ice-9/boot-9.scm:
   260:13  6 (for-each # ?)
   2312:4  5 (save-module-excursion _)
  3822:12  4 (_)
In mcron/base.scm:
   103:25  3 (add-job # ?)
In mcron/job-specifier.scm:
   251:32  2 (_ 1523321510)
 91:5  1 (bump-time #(0 51 20 9 3 118 1 98 1 14400 "EDT") _ # ?)
 63:4  0 (%find-best-next _ _)

mcron/job-specifier.scm:63:4: In procedure %find-best-next:
Throw to key `match-error' with args `("match" "no matching pattern"
0)'.
--8<---cut here---end--->8---

Maxim

bug#31095: [aegisub] error: Aegisub requires that boost be built with ICU support.

[Mark H Weaver]

On core-updates I pushed commit 713c975ac0aa01c224b215fcb96274b76920c90b
which hopefully will fix this problem with the aegisub build.  We'll see
when Hydra builds it out.

   Mark

bug#31088: Use '@' as version number separator in guix output

[Chris Marusich]

Hi Pierre and Mathieu,

Thank you for the bug report, Pierre!  Mathieu, can you take a look at
the patch I've attached to this email?  It should fix this bug, but
since you made similar changes in the past, I'd value your feedback.

Pierre Neidhardt  writes:

> Would it be possible to separate the package name from the version
> number with '@' instead?

Pierre's suggestion makes sense to me.  Starting with Mathieu's commit
1b846da8c372bee78851439fd9e72b2499115e5a about one year ago, we use "@"
to separate the package name from the version in commands like "guix
package -i ncdu@1.13" and also in exported scheme procedures like
specification->package (defined in (gnu packages)).  Before that, we
used "-" instead of "@".  When we made the change, it looks like we did
not also update the "dependencies" output from commands like "guix
package --show".

I've attached a patch that fixes the issue.  It applies cleanly to
df154c05dc65025eac2d14ab45699b34c7a1c7c8 (the tip of the master branch
at the time I drafted this email), and "make check" succeeds.  With the
change, the output now looks like this:

  $ ./pre-inst-env guix package --show=ncdu | recsel -p dependencies
  dependencies: ncurses@6.0-20170930

I was not able to run "make check-system" successfully, but there is no
reason to suspect that the attached patch causes any problem, since the
"check-system" target was already failing in the same way at commit
df154c05dc65025eac2d14ab45699b34c7a1c7c8.  I'll open up a separate bug
report for that separate issue.

Because I changed the package-full-name procedure, this also fixes the
package names in other places, such as graphs produced by "guix graph".
I'm not sure how many package rebuilds (if any?) this will cause, and
I'd like to hear what Mathieu has to say about my patch, so I'll wait to
push it until after I get feedback from him or another interested
person.  We should probably also regenerate the images in
guix/doc/images (so that the delimiter used in the images is "@" instead
of "-", although I'm not sure yet how to regenerate the source dot
files.  Mathieu, please let me know what you think!

In the meantime, Pierre, you can work around the issue by writing a
Guile script that gets the package version using the Guix API like so:

  scheme@(guile-user)> ,use (gnu packages ncdu)
  scheme@(guile-user)> ,use (guix packages)
  scheme@(guile-user)> (package-version ncdu)
  $1 = "1.13"

If you aren't sure how to get started writing Guile scripts, please
refer to the section "Running Guile Scripts" in the Guile manual.

-- 
Chris

From 96160c457c852d88e9a584cb16eb7703b02d8de5 Mon Sep 17 00:00:00 2001
From: Chris Marusich 
Date: Sun, 8 Apr 2018 16:51:42 -0700
Subject: [PATCH] guix: Separate the package name and version with "@", not
 "-".

* guix/packages.scm (package-full-name): By default, use "@" to separate
  the package name and package version.  Add #:delimiter keyword
  argument so that there is still a way to explicitly use a different
  delimiter.
* gnu/packages/commencement.scm (gcc-boot0) : Adjust
  accordingly.
* tests/graph.scm: Adjust accordingly.
* tests/profiles.scm: Adjust accordingly.
* NEWS: Mention the change.

Fixes: .
Reported by Pierre Neidhardt .
---
 NEWS  |  1 +
 gnu/packages/commencement.scm |  2 +-
 guix/packages.go.134WZR   |  0
 guix/packages.scm | 13 +
 tests/graph.scm   |  4 ++--
 tests/profiles.scm| 11 ++-
 6 files changed, 19 insertions(+), 12 deletions(-)
 create mode 100644 guix/packages.go.134WZR

diff --git a/NEWS b/NEWS
index 2c898e65f..9769bfc87 100644
--- a/NEWS
+++ b/NEWS
@@ -777,6 +777,7 @@ zynaddsubfx@3.0.2
 
 ** Programming interfaces
 
+*** package-full-name (guix packages) now uses "@" instead of "-" as its delimiter.
 *** New build systems: ‘font’, ‘meson’, ‘minify’, ‘scons’, ‘texlive’
 *** ‘cmake-build-system’ now supports cross-compilation
 *** Various improvements to ‘asdf-build-system’, ‘emacs-build-system’,
diff --git a/gnu/packages/commencement.scm b/gnu/packages/commencement.scm
index fe9fbebcc..9506fbb2b 100644
--- a/gnu/packages/commencement.scm
+++ b/gnu/packages/commencement.scm
@@ -282,7 +282,7 @@
 ;; Drop trailing letters, as gmp-6.0.0a unpacks
 ;; into gmp-6.0.0.
 `(symlink ,(string-trim-right
-(package-full-name lib)
+(package-full-name lib #:delimiter "-")
 char-set:letter)
   ,(package-name lib)))
   (list gmp-6.0 mpfr mpc
diff --git a/guix/packages.go.134WZR b/guix/packages.go.134WZR
new file mode 100644
index 0..e69de29bb
diff --git a/guix/packages.scm b/guix/packages.scm
index b5c0b6044..38e6344bf 100644
--- a/guix/packages.scm
+++ b/guix/packages.scm
@@ -388,