[bug#40643] [PATCH] git-version: Handle invalid arguments gracefully

[bug#40643] [PATCH] git-version: Handle invalid arguments gracefully

[Jakub Kądziołka]

* guix/git-download.scm (git-version): Add a check for commit ID length.
---
If you're curious for the motivation, see [1]. This took a while to
debug, so I'm hoping to ease this for the next person who inevitably
stumbles upon this. Is a change like this okay?

[1]: http://logs.guix.gnu.org/guix/2020-04-14.log#162809

 guix/git-download.scm | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/guix/git-download.scm b/guix/git-download.scm
index 1eae035fc4..40d81c72b9 100644
--- a/guix/git-download.scm
+++ b/guix/git-download.scm
@@ -2,6 +2,7 @@
 ;;; Copyright © 2014, 2015, 2016, 2017, 2018, 2019 Ludovic Courtès <ludo@gnu.org>
 ;;; Copyright © 2017 Mathieu Lirzin <mthl@gnu.org>
 ;;; Copyright © 2017 Christopher Baines <mail@cbaines.net>
+;;; Copyright © 2020 Jakub Kądziołka <kuba@kadziolka.net>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -30,6 +31,7 @@
   #:use-module (ice-9 match)
   #:use-module (ice-9 vlist)
   #:use-module (srfi srfi-1)
+  #:use-module (srfi srfi-35)
   #:export (git-reference
             git-reference?
             git-reference-url
@@ -170,6 +172,13 @@ HASH-ALGO (a symbol).  Use NAME as the file name, or a generic name if #f."
 
 (define (git-version version revision commit)
   "Return the version string for packages using git-download."
+  ;; git-version is almost exclusively executed while modules are being loaded.
+  ;; This makes any errors hide their backtrace. Avoid the mysterious error
+  ;; "Value out of range 0 to N: 7" when the commit ID is too short, which
+  ;; can happen, for example, when the user swapped the revision and commit
+  ;; arguments by mistake.
+  (when (< (string-length commit) 7)
+    (error "git-version: commit ID unexpectedly short"))
   (string-append version "-" revision "." (string-take commit 7)))
 
 (define (git-file-name name version)
-- 
2.26.0

[bug#40643] [PATCH] git-version: Handle invalid arguments gracefully

[Ludovic Courtès]

Hi Jakub,

Jakub Kądziołka <kuba@kadziolka.net> skribis:

> * guix/git-download.scm (git-version): Add a check for commit ID length.
> ---
> If you're curious for the motivation, see [1]. This took a while to
> debug, so I'm hoping to ease this for the next person who inevitably
> stumbles upon this. Is a change like this okay?

Yes, I think so.  The ‘error’ procedure is not great, we would rather
use ‘raise’ with a ‘&message’ condition (which additionally allows for
i18n) but it’s no big deal here.

Thanks,
Ludo’.

[bug#40643] [PATCH] git-version: Handle invalid arguments gracefully

[Jakub Kądziołka]

[-- Attachment #1: Type: text/plain, Size: 937 bytes --]

On Fri, Apr 17, 2020 at 11:16:44PM +0200, Ludovic Courtès wrote:
> Hi Jakub,
> 
> Jakub Kądziołka <kuba@kadziolka.net> skribis:
> 
> > * guix/git-download.scm (git-version): Add a check for commit ID length.
> > ---
> > If you're curious for the motivation, see [1]. This took a while to
> > debug, so I'm hoping to ease this for the next person who inevitably
> > stumbles upon this. Is a change like this okay?
> 
> Yes, I think so.  The ‘error’ procedure is not great, we would rather
> use ‘raise’ with a ‘&message’ condition (which additionally allows for
> i18n) but it’s no big deal here.

I considered using raise instead, but I couldn't get it to work
properly. I was getting a "Wrong type (expecting exact integer)" error
instead:

(raise
  (condition (&message
    (message "git-version: commit ID unexpectedly short"))))

Do you know why that might be?

Regards,
Jakub Kądziołka

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[bug#40643] [PATCH] git-version: Handle invalid arguments gracefully

[Ludovic Courtès]

Hi,

Jakub Kądziołka <kuba@kadziolka.net> skribis:

> On Fri, Apr 17, 2020 at 11:16:44PM +0200, Ludovic Courtès wrote:
>> Hi Jakub,
>> 
>> Jakub Kądziołka <kuba@kadziolka.net> skribis:
>> 
>> > * guix/git-download.scm (git-version): Add a check for commit ID length.
>> > ---
>> > If you're curious for the motivation, see [1]. This took a while to
>> > debug, so I'm hoping to ease this for the next person who inevitably
>> > stumbles upon this. Is a change like this okay?
>> 
>> Yes, I think so.  The ‘error’ procedure is not great, we would rather
>> use ‘raise’ with a ‘&message’ condition (which additionally allows for
>> i18n) but it’s no big deal here.
>
> I considered using raise instead, but I couldn't get it to work
> properly. I was getting a "Wrong type (expecting exact integer)" error
> instead:
>
> (raise
>   (condition (&message
>     (message "git-version: commit ID unexpectedly short"))))
>
> Do you know why that might be?

It must be because you forgot to include (srfi srfi-34): there are two
‘raise’ procedure, and in core Guile ‘raise’ is about signals, not error
conditions.

HTH!

Ludo’.

bug#40643: [PATCH] git-version: Handle invalid arguments gracefully

[Jakub Kądziołka]

[-- Attachment #1: Type: text/plain, Size: 322 bytes --]

On Wed, Apr 22, 2020 at 05:10:18PM +0200, Ludovic Courtès wrote:
> It must be because you forgot to include (srfi srfi-34): there are two
> ‘raise’ procedure, and in core Guile ‘raise’ is about signals, not error
> conditions.

Thanks! That did it! I pushed an updated patch.

Regards,
Jakub Kądziołka

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]