bug#68378: Gawk 5.3.0 test suite fails non-deterministically

[Ludovic Courtès]

Hello,

Evaluation  of ‘core-updates’
(commit ec1ce76a3a604efdc329b482fb46fa03cd11e184) failed due to test
failures in /gnu/store/pjdg38adv5d8lad0kk7cga4dd57cpvb4-gawk-5.3.0.drv:

--8<---cut here---start->8---
--- time.ok 2023-06-14 22:13:40.0 +
+++ _time   2024-01-11 00:47:05.999386040 +
@@ -1,4 +1,4 @@
 gettimeofday - systime = 0
 sleep(1.3) = 0
-gettimeofday - systime = 0
+gettimeofday - systime = 1
  <>
make[3]: *** [Makefile:5696: diffout] Error 1
make[3]: Leaving directory '/tmp/guix-build-gawk-5.3.0.drv-0/gawk-5.3.0/test'
make[2]: *** [Makefile:2189: check] Error 1
--8<---cut here---end--->8---

However, the same derivation builds fine when I retry.

It would be nice to address that.

Ludo’.

bug#68333: Time bomb in icedtea/openjdk

[Rostislav Svoboda]

Hi,

> What do you think of the attached patch?  The difference is that it
> patches code at its root (in the actual source tarball rather than after
> it’s been copied), it does so for IcedTea 7 and 8, and there are patch
> files that are slightly clearer than a substitution pattern.

That's clearly a much better solution than that substitution pattern of mine.

Besides, I missed the path difference:
  jdk-3.19.0-2d5d2c77faa3:
/make/src/classes/build/tools/generatecurrencydata/GenerateCurrencyData.java
  jdk-2.6.13-68b6bb380175:
/make/tools/src/build/tools/generatecurrencydata/GenerateCurrencyData.java

Speaking of which, I'd go for e.g.:
  jdk-3.19.0-currency-time-bomb.patch
  jdk-2.6.13-currency-time-bomb.patch
instead of:
  jdk-currency-time-bomb.patch
  jdk-currency-time-bomb2.patch

Yeah naming is hard ;-)

Also, changing the error string from:
  time is more than 10 years from present
to:
  time is more than 10 years from \"present\"

i.e. adding double quotes will slightly complicate googling for that
error message in the future.

> Pushed as 5c0f77f4241c9beac0c82deae946bfdc70b49ff0.

Thanks

> Let’s hope there’s no similar time bomb elsewhere in the Java stack.

I'm looking nd... I haven't found anything so far.

Cheers Bost

bug#39703: MPV chooses the slow AV1 decoder

[Leo Famulari]

After 4 years, many upstream releases of mpv and FFmpeg, and a new
laptop, I'm closing this bug as "not actionable".

bug#68333: Time bomb in icedtea/openjdk

[Ludovic Courtès]

Ludovic Courtès  skribis:

>>From 92561a776ad88eb73034948beedbe9e6be4077b4 Mon Sep 17 00:00:00 2001
> Message-ID: 
> <92561a776ad88eb73034948beedbe9e6be4077b4.1704897216.git.l...@gnu.org>
> From: =?UTF-8?q?Ludovic=20Court=C3=A8s?= 
> Date: Wed, 10 Jan 2024 15:27:47 +0100
> Subject: [PATCH] =?UTF-8?q?gnu:=20icedtea:=20Fix=20time=20bomb=20in=20?=
>  =?UTF-8?q?=E2=80=98GenerateCurrencyData.java=E2=80=99.?=
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
>
> Fixes .
>
> * gnu/packages/java.scm (icedtea-7)[drop]: Add optional ‘patches’
> parameter and honor it.
> [native-inputs]: Add patch for “jdk-drop”.
> (icedtea-8): Likewise.
> * gnu/packages/patches/jdk-currency-time-bomb.patch,
> gnu/packages/patches/jdk-currency-time-bomb2.patch: New files.
> * gnu/local.mk (dist_patch_DATA): Add them.
>
> Change-Id: I3b4cf562ce4bedb87ce0e7c289e431fbb0dbb6f8

Pushed as 5c0f77f4241c9beac0c82deae946bfdc70b49ff0.

Let’s hope there’s no similar time bomb elsewhere in the Java stack.

Thanks again,
Ludo’.

bug#39885: Bioconductor tarballs are not archived

[Simon Tournier]

Hi,

On Wed, 10 Jan 2024 at 13:42, Ricardo Wurmus  wrote:

> Do we have a list of desired tarballs?

I made one. :-)  I will share the manifest once a bit more polish. ;-)

Roughly, until 3.11 -- for earlier, I am still fighting with
time-machine -- and only the versions right before a Bioconductor
upgrade (and sadly not the version after such upgrade that, for many
have disappeared from Bioconductor).  If you have them, cool!  Well,
for now, I am speaking about all the Bioconductor packages except
annotations for these Guix revisions:

--8<---cut here---start->8---
git log --format="%P %s" --after=2019-04-30 \
| grep -i -E 'bioconductor' | grep -i -E '(update|upgrade)' \
| cut -f1 -d' ' \
| head -8
--8<---cut here---end--->8---

And for instance, I have this kind of sources.json.

--8<---cut here---start->8---
{
  "sources": [
{
  "type": "url",
  "urls": [

"https://bordeaux.guix.gnu.org/file/progeny_1.22.0.tar.gz/sha256/047x6by3xa15gvi3kny5pkqxaq8d2kzcfi55ic5j7a351715l6l7;,

"https://ci.guix.gnu.org/file/progeny_1.22.0.tar.gz/sha256/047x6by3xa15gvi3kny5pkqxaq8d2kzcfi55ic5j7a351715l6l7;,

"https://bioconductor.org/packages/3.17/bioc/src/contrib/progeny_1.22.0.tar.gz;,

"https://tarballs.nixos.org/sha256/047x6by3xa15gvi3kny5pkqxaq8d2kzcfi55ic5j7a351715l6l7;
  ],
  "integrity": "sha256-hxpawgllqCMLi6VEx/4UDWHV8bzF2znifiWoPvwy/RA=",
  "outputHashAlgo": "sha256",
  "outputHashMode": "flat"
},
{
  "type": "url",
  "urls": [

"https://bordeaux.guix.gnu.org/file/AWFisher_1.14.0.tar.gz/sha256/1c6rr1z1rhvn8w1kb3nnjlfacfr22vwm1rsa1xqm2hmghs01bq4x;,

"https://ci.guix.gnu.org/file/AWFisher_1.14.0.tar.gz/sha256/1c6rr1z1rhvn8w1kb3nnjlfacfr22vwm1rsa1xqm2hmghs01bq4x;,

"https://bioconductor.org/packages/3.17/bioc/src/contrib/AWFisher_1.14.0.tar.gz;,

"https://tarballs.nixos.org/sha256/1c6rr1z1rhvn8w1kb3nnjlfacfr22vwm1rsa1xqm2hmghs01bq4x;
  ],
  "integrity": "sha256-neAVgIavQlFxD0rnUPkWIjumHJXWjjUDR3bDHH7I2bA=",
  "outputHashAlgo": "sha256",
  "outputHashMode": "flat"
},
--8<---cut here---end--->8---

Where the URL for ci.guix is incorrect, I guess.

Cheers,
simon

bug#68333: Time bomb in icedtea/openjdk

[Ludovic Courtès]

Hi,

Rostislav Svoboda  skribis:

> From 1e86e32825a5025b4ef439e7f678143416f622b5 Mon Sep 17 00:00:00 2001
> Message-ID: 
> <1e86e32825a5025b4ef439e7f678143416f622b5.1704815654.git.rostislav.svob...@gmail.com>
> From: Rostislav Svoboda 
> Date: Tue, 9 Jan 2024 14:08:02 +0100
> Subject: [PATCH] gnu: Fix time bomb preventing build of icedtea.
>
> Fixes .
>
> Fixes time bomb preventing build of icedtea by patching the java source code
> so that the java-RuntimeException is not thrown when 'more than 10 years ago'
> condition is true.
>
> * gnu/packages/java.scm (icedtea): Patch java source code so that the
> offending line is commented out.
>
> Change-Id: I4861ee2ffc5aaffb6a9244110d222ae64bfbdf94

Thanks for the quick investigation and patch!  I just realized that this
affects current ‘master’ so we’d rather fix it soon.

What do you think of the attached patch?  The difference is that it
patches code at its root (in the actual source tarball rather than after
it’s been copied), it does so for IcedTea 7 and 8, and there are patch
files that are slightly clearer than a substitution pattern.

It’s still building on my machine but if it works for everyone, I’d like
to push as soon as I got it built.

Ludo’.

>From 92561a776ad88eb73034948beedbe9e6be4077b4 Mon Sep 17 00:00:00 2001
Message-ID: <92561a776ad88eb73034948beedbe9e6be4077b4.1704897216.git.l...@gnu.org>
From: =?UTF-8?q?Ludovic=20Court=C3=A8s?= 
Date: Wed, 10 Jan 2024 15:27:47 +0100
Subject: [PATCH] =?UTF-8?q?gnu:=20icedtea:=20Fix=20time=20bomb=20in=20?=
 =?UTF-8?q?=E2=80=98GenerateCurrencyData.java=E2=80=99.?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fixes .

* gnu/packages/java.scm (icedtea-7)[drop]: Add optional ‘patches’
parameter and honor it.
[native-inputs]: Add patch for “jdk-drop”.
(icedtea-8): Likewise.
* gnu/packages/patches/jdk-currency-time-bomb.patch,
gnu/packages/patches/jdk-currency-time-bomb2.patch: New files.
* gnu/local.mk (dist_patch_DATA): Add them.

Change-Id: I3b4cf562ce4bedb87ce0e7c289e431fbb0dbb6f8
---
 gnu/local.mk |  2 ++
 gnu/packages/java.scm| 16 ++--
 .../patches/jdk-currency-time-bomb.patch | 13 +
 .../patches/jdk-currency-time-bomb2.patch| 13 +
 4 files changed, 38 insertions(+), 6 deletions(-)
 create mode 100644 gnu/packages/patches/jdk-currency-time-bomb.patch
 create mode 100644 gnu/packages/patches/jdk-currency-time-bomb2.patch

diff --git a/gnu/local.mk b/gnu/local.mk
index 16a34065c6..22970932e9 100644
--- a/gnu/local.mk
+++ b/gnu/local.mk
@@ -1478,6 +1478,8 @@ dist_patch_DATA =		\
   %D%/packages/patches/java-xerces-build_dont_unzip.patch	\
   %D%/packages/patches/java-xerces-xjavac_taskdef.patch	\
   %D%/packages/patches/jbr-17-xcursor-no-dynamic.patch	\
+  %D%/packages/patches/jdk-currency-time-bomb.patch		\
+  %D%/packages/patches/jdk-currency-time-bomb2.patch		\
   %D%/packages/patches/jfsutils-add-sysmacros.patch		\
   %D%/packages/patches/jfsutils-gcc-compat.patch		\
   %D%/packages/patches/jfsutils-include-systypes.patch		\
diff --git a/gnu/packages/java.scm b/gnu/packages/java.scm
index e2c2a041a9..ded73bc461 100644
--- a/gnu/packages/java.scm
+++ b/gnu/packages/java.scm
@@ -100,13 +100,14 @@ (define-module (gnu packages java)
 
 (define-public icedtea-7
   (let* ((version "2.6.13")
- (drop (lambda (name hash)
+ (drop (lambda* (name hash #:optional (patches '()))
  (origin
(method url-fetch)
(uri (string-append
  "http://icedtea.classpath.org/download/drops;
  "/icedtea7/" version "/" name ".tar.bz2"))
-   (sha256 (base32 hash))
+   (sha256 (base32 hash))
+   (patches patches)
 (package
   (name "icedtea")
   (version version)
@@ -614,7 +615,8 @@ (define-public icedtea-7
  "110j7jlz47x2gg6f7653x12mssan5kvj9l9h1m1c8c92drfxbqyk"))
  ("jdk-drop"
   ,(drop "jdk"
- "0d1mca38ksxvdskp9im3pp7fdijhj1n3lwq9w13r9s4v3qyskgdd"))
+ "0d1mca38ksxvdskp9im3pp7fdijhj1n3lwq9w13r9s4v3qyskgdd"
+ (search-patches "jdk-currency-time-bomb.patch")))
  ("langtools-drop"
   ,(drop "langtools"
  "0nq5236fzxn3p6x8cgncl56mzcmsj07q9gymysnws4c8byc6n0qj"))
@@ -687,13 +689,14 @@ (define-public icedtea-7
 
 (define-public icedtea-8
   (let* ((version "3.19.0")
- (drop (lambda (name hash)
+ (drop (lambda* (name hash #:optional (patches '()))
  (origin
(method url-fetch)
(uri (string-append
  "http://icedtea.classpath.org/download/drops;
  "/icedtea8/" version "/" name ".tar.xz"))
-  

bug#39885: Bioconductor tarballs are not archived

[Ricardo Wurmus]

Ludovic Courtès  writes:

> I my previous message, I wrote:
>
>> As for past tarballs, #swh-devel comrades say we could send them a list
>> of URLs and they’d create “Save Code Now” requests on our behalf (we
>> cannot do it ourselves since the site doesn’t accept plain tarballs.)
>
> Were you able to retrieve some of these?  What are the chances of
> success?

Do we have a list of desired tarballs?  I still have an archived
/gnu/store from before we moved the shared cluster installation at the
MDC to different storage.  It might contain old tarballs that we no
longer have elsewhere.

-- 
Ricardo