make check fails on 182be30fb1a8b847c30492462ec22c08ec7a9849

Done

Details

3 participants

Josselin Poiret
Janneke Nieuwenhuizen
Ludovic Courtès

Owner: unassigned

Submitted by: Janneke Nieuwenhuizen

Severity: normal

Merged with

64598

Janneke Nieuwenhuizen wrote on 21 Jul 2023 08:47

Recipients:(address . bug-guix@gnu.org)

Message-ID:87cz0lu5f3.fsf@gnu.org

Hi,

Make check gives three failures for me on current master

FAIL: tests/packages (package->bag)

FAIL: tests/store-roots (gc-roots, initial)

FAIL: tests/texlive (texlive->guix-package, meta-package)

using this snippet.

Toggle snippet (4 lines)

export SCM_LOG_DRIVER_FLAGS="--select='^(package->bag|gc-roots, initial|texlive->guix-package, meta-package)'"

make check TESTS="tests/packages.scm tests/store-roots.scm tests/texlive.scm"

Find log excerpts attached.

I have bisected the first (package->bag) to be broken by

0dd293b4d9095137c9952e16ca951f887b7e7018

gnu: Add libc-for-target and glibc/hurd.

but as yet fail to why.

Greetings,

Janneke

Attachment: packages.log

Attachment: store-roots.log

Attachment: texlive.log

-- 
Janneke Nieuwenhuizen <janneke@gnu.org>  | GNU LilyPond https://LilyPond.org
Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Josselin Poiret wrote on 21 Jul 2023 12:12

Recipients:

Message-ID:877cqtfu95.fsf@jpoiret.xyz

Hi Janneke,

Janneke Nieuwenhuizen <janneke@gnu.org> writes:

Toggle quote (10 lines)> Hi,
>
> Make check gives three failures for me on current master
>
>     FAIL: tests/packages     (package->bag)
>     FAIL: tests/store-roots  (gc-roots, initial)  
>     FAIL: tests/texlive      (texlive->guix-package, meta-package)
>
> using this snippet.

I have local fixes for all of them, but I am not 100% sure about the

gc-roots one. Ludo, WDYT? Also, guix-shell fails on the fdes test

locally for me, probably because I'm using zsh. All should be fine in

the package though.

Best,

Josselin Poiret

Josselin Poiret wrote on 21 Jul 2023 12:14

[PATCH 1/3] tests: packages: Set system for expected result of package->bag.

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:37df96c0b74c4cc20191eb05a7fcc2ee69d586f0.1689934453.git.dev@jpoiret.xyz

* tests/packages.scm ("package->bag"): Parameterize the expected result by the

system used to lower the package to a bag.

---

tests/packages.scm | 5 +++--

1 file changed, 3 insertions(+), 2 deletions(-)

Toggle diff (20 lines)diff --git a/tests/packages.scm b/tests/packages.scm
index 2b7ab01f7d..2b4f9f8e90 100644
--- a/tests/packages.scm
+++ b/tests/packages.scm
@@ -1277,8 +1277,9 @@ (define compressors '(("gzip"  . "gz")
 ;;                               #:guile guile))))
 
 (test-equal "package->bag"
-  `("foo86-hurd" #f (,(package-source gnu-make))
-    (,(canonical-package glibc)) (,(canonical-package coreutils)))
+  (parameterize ((%current-system "foo86-hurd"))
+    `("foo86-hurd" #f (,(package-source gnu-make))
+      (,(canonical-package glibc)) (,(canonical-package coreutils))))
   (let ((bag (package->bag gnu-make "foo86-hurd")))
     (list (bag-system bag) (bag-target bag)
           (assoc-ref (bag-build-inputs bag) "source")

base-commit: 182be30fb1a8b847c30492462ec22c08ec7a9849
-- 
2.40.1

Josselin Poiret wrote on 21 Jul 2023 12:14

[PATCH 2/3] tests: store-roots: Initial gc-roots should be empty.

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:ad74949be91d909948d5539efad05e01556657b2.1689934453.git.dev@jpoiret.xyz

* tests/store-roots.scm ("gc-roots, initial"): Set expected result to empty.

Also do not error out if /profiles doesn't exist.

---

tests/store-roots.scm | 5 +++--

1 file changed, 3 insertions(+), 2 deletions(-)

Toggle diff (20 lines)diff --git a/tests/store-roots.scm b/tests/store-roots.scm
index 9877987a65..d82a29e313 100644
--- a/tests/store-roots.scm
+++ b/tests/store-roots.scm
@@ -31,10 +31,11 @@ (define %store #f)
 (test-begin "store-roots")
 
 (test-equal "gc-roots, initial"
-  (list (string-append %state-directory "/profiles"))
+  '()
   (begin
     ;; 'gc-roots' should gracefully handle lack of that directory.
-    (delete-file-recursively (string-append %state-directory "/profiles"))
+    (false-if-exception
+     (delete-file-recursively (string-append %state-directory "/profiles")))
     (gc-roots)))
 
 ;; The 'open-connection' call below gets guix-daemon to create
-- 
2.40.1

Josselin Poiret wrote on 21 Jul 2023 12:14

[PATCH 3/3] tests: texlive: Remove texlive-texworks from propagated-inputs.

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:09df89f19b2b360097634520ca4c281aaa4cc852.1689934453.git.dev@jpoiret.xyz

* tests/texlive.scm ("texlive->guix-package, meta-package"): Remove

texlive-texworks from expected propagated-inputs, as it is now ignored by

texlive->guix-package.

---

tests/texlive.scm | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)

Toggle diff (15 lines)diff --git a/tests/texlive.scm b/tests/texlive.scm
index 98461f7e51..b8b1d5c6d6 100644
--- a/tests/texlive.scm
+++ b/tests/texlive.scm
@@ -542,7 +542,7 @@ (define %fake-tlpdb
                ('arguments
                 ('list '#:builder ('gexp ('mkdir ('ungexp 'output)))))
                ('propagated-inputs
-                ('list 'texlive-collection-basic 'texlive-texworks))
+                ('list 'texlive-collection-basic))
                ('home-page "https://www.tug.org/texlive/")
                ('synopsis (? string?))
                ('description (? string?))
-- 
2.40.1

Janneke Nieuwenhuizen wrote on 21 Jul 2023 13:17

Re: bug#64760: make check fails on 182be30fb1a8b847c30492462ec22c08ec7a9849

Recipients:(name . Josselin Poiret)(address . dev@jpoiret.xyz)(address . 64760@debbugs.gnu.org)

Message-ID:874jlxtsx1.fsf@gnu.org

Josselin Poiret writes:

Hi Josselin,

Toggle quote (10 lines)>> Make check gives three failures for me on current master
>>
>>     FAIL: tests/packages     (package->bag)
>>     FAIL: tests/store-roots  (gc-roots, initial)  
>>     FAIL: tests/texlive      (texlive->guix-package, meta-package)
>>
>> using this snippet.
>
> I have local fixes for all of them

Nice! I can confirm they work for me.

Greetings,

Janneke

Janneke Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond https://LilyPond.org

Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Janneke Nieuwenhuizen wrote on 28 Jul 2023 10:52

Re: [PATCH 2/3] tests: store-roots: Initial gc-roots should be empty.

Recipients:(name . Josselin Poiret)(address . dev@jpoiret.xyz)

Message-ID:87mszgcta7.fsf@gnu.org

Josselin Poiret writes:

Hi!

Toggle quote (25 lines)> * tests/store-roots.scm ("gc-roots, initial"): Set expected result to empty.
> Also do not error out if /profiles doesn't exist.
> ---
>  tests/store-roots.scm | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/tests/store-roots.scm b/tests/store-roots.scm
> index 9877987a65..d82a29e313 100644
> --- a/tests/store-roots.scm
> +++ b/tests/store-roots.scm
> @@ -31,10 +31,11 @@ (define %store #f)
>  (test-begin "store-roots")
>  
>  (test-equal "gc-roots, initial"
> -  (list (string-append %state-directory "/profiles"))
> +  '()
>    (begin
>      ;; 'gc-roots' should gracefully handle lack of that directory.
> -    (delete-file-recursively (string-append %state-directory "/profiles"))
> +    (false-if-exception
> +     (delete-file-recursively (string-append %state-directory "/profiles")))
>      (gc-roots)))
>  
>  ;; The 'open-connection' call below gets guix-daemon to create

Just a headsup that this change breaks the store-roots test on the hurd

for me.

Greetings,

Janneke

Janneke Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond https://LilyPond.org

Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Janneke Nieuwenhuizen wrote on 28 Jul 2023 13:58

Recipients:(name . Josselin Poiret)(address . dev@jpoiret.xyz)

Message-ID:87ila4cknl.fsf@gnu.org

Janneke Nieuwenhuizen writes:

Hello,

Toggle quote (6 lines)> Josselin Poiret writes:
>
> Hi!
>
>> * tests/store-roots.scm ("gc-roots, initial"): Set expected result to empty.
>> Also do not error out if /profiles doesn't exist.

[..]

Toggle quote (3 lines)

> Just a headsup that this change breaks the store-roots test on the hurd

> for me.

aaand as discussed on IRC, the missing output

Toggle snippet (20 lines)

FAIL: tests/store-roots

=======================

test-name: gc-roots, initial

location: /tmp/guix-build-guix-1.4.0-8.5623ab5.drv-0/source/tests/store-roots.scm:33

source:

+ (test-equal

+ "gc-roots, initial"

+ '()

+ (begin

+ (false-if-exception

+ (delete-file-recursively

+ (string-append %state-directory "/profiles")))

+ (gc-roots)))

expected-value: ()

actual-value: ("/tmp/guix-tests/var/19337/profiles")

result: FAIL

-- 
Janneke Nieuwenhuizen <janneke@gnu.org>  | GNU LilyPond https://LilyPond.org
Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Josselin Poiret wrote on 28 Jul 2023 16:00

Recipients:(name . Janneke Nieuwenhuizen)(address . janneke@gnu.org)

Message-ID:873518m8zr.fsf@jpoiret.xyz

Hi everyone,

Janneke Nieuwenhuizen <janneke@gnu.org> writes:

Toggle quote (3 lines)

> Just a headsup that this change breaks the store-roots test on the hurd

> for me.

Thanks for the feedback Janneke.  It seems that there is some left-over
state that can mess with this state's result: I was initially surprised
that the /profiles directory could appear in the gc roots, but that's
because it is symlinked under /gcroots, not because it is itself
searched for gc roots.

In any case, the /gcroots directory along with the /gcroots/profiles
symlink is created when a connection is made to the daemon, which is not
the case here yet.  However, a connection might have been opened before
for the same state dir (which depends on the PID of
build-aux/test-env.in).  It might also depend on whether the clean-up of
the state directory made by `trap` worked and whether PIDs get re-used
quickly on the specific kernel.  I think this is all too unreliable
here (I have one such example of a leftover PID state dir in my tree, so
it might happen more often than not).

In any case, if this test is only here to check if gc-roots doesn't
error out, we could return #t at the end to only fish for errors.  WDYT?

Best,
-- 
Josselin Poiret

Janneke Nieuwenhuizen wrote on 30 Jul 2023 18:19

Re: bug#64760: make check fails on 182be30fb1a8b847c30492462ec22c08ec7a9849

Recipients:(name . Josselin Poiret)(address . dev@jpoiret.xyz)

Message-ID:874jllcqxv.fsf_-_@gnu.org

Josselin Poiret writes:

Hey!

Toggle quote (21 lines)> Janneke Nieuwenhuizen <janneke@gnu.org> writes:
>
>> Just a headsup that this change breaks the store-roots test on the hurd
>> for me.
>
> Thanks for the feedback Janneke.  It seems that there is some left-over
> state that can mess with this state's result: I was initially surprised
> that the /profiles directory could appear in the gc roots, but that's
> because it is symlinked under /gcroots, not because it is itself
> searched for gc roots.
>
> In any case, the /gcroots directory along with the /gcroots/profiles
> symlink is created when a connection is made to the daemon, which is not
> the case here yet.  However, a connection might have been opened before
> for the same state dir (which depends on the PID of
> build-aux/test-env.in).  It might also depend on whether the clean-up of
> the state directory made by `trap` worked and whether PIDs get re-used
> quickly on the specific kernel.  I think this is all too unreliable
> here (I have one such example of a leftover PID state dir in my tree, so
> it might happen more often than not).

Yes, that could be. I've got another data-point for i686-linux (see

also https://issues.guix.gnu.org/64959);the package and texlive tests

now pass, and store-roots is made to fail like so:

Toggle snippet (16 lines)test-name: gc-roots, initial
location: /home/janneke/src/guix/wip-i686/tests/store-roots.scm:33
source:
+ (test-equal
+   "gc-roots, initial"
+   '()
+   (begin
+     (false-if-exception
+       (delete-file-recursively
+         (string-append %state-directory "/profiles")))
+     (gc-roots)))
expected-value: ()
actual-value: ("/home/janneke/src/guix/wip-i686/test-tmp/var/30690/profiles")
result: FAIL

Toggle quote (3 lines)

> In any case, if this test is only here to check if gc-roots doesn't

> error out, we could return #t at the end to only fish for errors. WDYT?

(good question.../me passes insights here to "others" ;)

Greetings,

Janneke

Janneke Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond https://LilyPond.org

Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Ludovic Courtès wrote on 9 Aug 2023 23:41

Recipients:(address . 64760@debbugs.gnu.org)

Message-ID:874jl7opvj.fsf@gnu.org

Hello!

Janneke Nieuwenhuizen <janneke@gnu.org> skribis:

Toggle quote (21 lines)> test-name: gc-roots, initial
> location: /home/janneke/src/guix/master/tests/store-roots.scm:33
> source:
> + (test-equal
> +   "gc-roots, initial"
> +   (list (string-append %state-directory "/profiles"))
> +   (begin
> +     (delete-file-recursively
> +       (string-append %state-directory "/profiles"))
> +     (gc-roots)))
> expected-value: ("/home/janneke/src/guix/master/test-tmp/var/9734/profiles")
> actual-value: #f
> actual-error:
> + (system-error
> +   "lstat"
> +   "~A: ~S"
> +   ("No such file or directory"
> +    "/home/janneke/src/guix/master/test-tmp/var/9734/profiles")
> +   (2))
> result: FAIL

Hmm, this test passes for me.

Could you get a backtrace so we see where the ‘lstat’ exception comes

from? (To get the backtrace, you can raise (begin …) out of

‘test-equal’ and then “make check TESTS=tests/store-root.scm”.)

Thanks,

Ludo’.

Janneke Nieuwenhuizen wrote on 11 Aug 2023 14:31

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:871qg9sqtg.fsf@gnu.org

Ludovic Courtès writes:

Hey!

Toggle quote (25 lines)> Janneke Nieuwenhuizen <janneke@gnu.org> skribis:
>
>> test-name: gc-roots, initial
>> location: /home/janneke/src/guix/master/tests/store-roots.scm:33
>> source:
>> + (test-equal
>> +   "gc-roots, initial"
>> +   (list (string-append %state-directory "/profiles"))
>> +   (begin
>> +     (delete-file-recursively
>> +       (string-append %state-directory "/profiles"))
>> +     (gc-roots)))
>> expected-value: ("/home/janneke/src/guix/master/test-tmp/var/9734/profiles")
>> actual-value: #f
>> actual-error:
>> + (system-error
>> +   "lstat"
>> +   "~A: ~S"
>> +   ("No such file or directory"
>> +    "/home/janneke/src/guix/master/test-tmp/var/9734/profiles")
>> +   (2))
>> result: FAIL
>
> Hmm, this test passes for me.

Well, what would you know, it passes for me too...

Toggle quote (4 lines)

> Could you get a backtrace so we see where the ‘lstat’ exception comes

> from? (To get the backtrace, you can raise (begin …) out of

> ‘test-equal’ and then “make check TESTS=tests/store-root.scm”.)

...as I found out after trying this.  I had my source tree configured
using --disable-daemon --disable-nls (a nifty feature that I encountered
only recently)...and then this (patched) test fails.  After
reconfiguring, all is well.

Greetings,
Janneke

-- 
Janneke Nieuwenhuizen <janneke@gnu.org>  | GNU LilyPond https://LilyPond.org
Freelance IT https://www.JoyOfSource.com| Avatar® https://AvatarAcademy.com

Ludovic Courtès wrote on 18 Aug 2023 16:11

control message for bug #64598

Recipients:(address . control@debbugs.gnu.org)

Message-ID:87wmxsqw2r.fsf@gnu.org

merge 64598 64760

quit

Ludovic Courtès wrote on 18 Aug 2023 16:13

Re: bug#64760: make check fails on 182be30fb1a8b847c30492462ec22c08ec7a9849

Recipients:(name . Josselin Poiret)(address . dev@jpoiret.xyz)

Message-ID:87sf8gqw0e.fsf_-_@gnu.org

Hello!

Pushed as 4b1fdd44ed87dc07d23cc9313a8b6b14d0646923!

Ludo’.

Closed

Your comment

This issue is archived.

To comment on this conversation send an email to 64760@debbugs.gnu.org

is:open	open issues
is:done	closed issues
submitter:<who>	search issue submitter
author:<who>	search by message author
date:yesterday..now	search by issue date
mdate:3m..2d	search by message date