[bug#40143] [PATCH] gnu: git: Return #t in install-man-pages phase.

[bug#40143] [PATCH] gnu: git: Return #t in install-man-pages phase.

[Brice Waegeneire]

* gnu/packages/version-control.scm (git)[arguments]: Return #t in
install-man-pages phase.
---
 gnu/packages/version-control.scm | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/gnu/packages/version-control.scm b/gnu/packages/version-control.scm
index d25df4c63d..7eb85cc749 100644
--- a/gnu/packages/version-control.scm
+++ b/gnu/packages/version-control.scm
@@ -441,7 +441,8 @@ as well as the classic centralized workflow.")
                    (manpages (assoc-ref inputs "git-manpages")))
               (mkdir-p man)
               (with-directory-excursion man
-                (invoke "tar" "xvf" manpages))))))))
+                (invoke "tar" "xvf" manpages)))
+            #t)))))
 
    (native-search-paths
     ;; For HTTPS access, Git needs a single-file certificate bundle, specified

base-commit: 837644b48a5107a28acbdb6d172aff0040a8247d
prerequisite-patch-id: b9e1cb7e09e15cac4d607e985e2a3b5e301828e5
prerequisite-patch-id: e22ea3f99fc3d9f3e11ddfa31ba60212f2ed4d62
-- 
2.25.1

[bug#40143] [PATCH] gnu: git: Return #t in install-man-pages phase.

[Danny Milosavljevic]

[-- Attachment #1: Type: text/plain, Size: 103 bytes --]

invoke already returns #t, so there's no technical bug.

That said, maybe for clarity?  Not sure...

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[bug#40143] [PATCH] gnu: git: Return #t in install-man-pages phase.

[Maxim Cournoyer]

Danny Milosavljevic <dannym@scratchpost.org> writes:

> invoke already returns #t, so there's no technical bug.
>
> That said, maybe for clarity?  Not sure...

I don't see any technical nor clarity issue here.  The vision laid out
by Mark more than two years ago (!) was to get rid of all the boolean
return codes and move to an exception based system entirely, which I
find much more elegant.  The transition plan was outlined in this
message:
https://lists.gnu.org/archive/html/guix-devel/2017-12/msg00278.html.

I think we should go back to work on finishing the missing bits :-).

Maxim

bug#40143: [PATCH] gnu: git: Return #t in install-man-pages phase.

[Mathieu Othacehe]

Hello Brice,

> Danny Milosavljevic <dannym@scratchpost.org> writes:
>
>> invoke already returns #t, so there's no technical bug.
>>
>> That said, maybe for clarity?  Not sure...

I agree with Danny and Maxim here, we already return a boolean, or
raise an exception in case of error, so I think we are fine :)

Closing this bug,

Thanks,

Mathieu

[bug#40143] [PATCH] gnu: git: Return #t in install-man-pages phase.

[Brice Waegeneire]

On 2020-03-21 14:29, Mathieu Othacehe wrote:
> Hello Brice,
> 
>> Danny Milosavljevic <dannym@scratchpost.org> writes:
>> 
>>> invoke already returns #t, so there's no technical bug.
>>> 
>>> That said, maybe for clarity?  Not sure...
> 
> I agree with Danny and Maxim here, we already return a boolean, or
> raise an exception in case of error, so I think we are fine :)
> 
> Closing this bug,
> 
> Thanks,
> 
> Mathieu

Looks like a forgot to reply to all, I replied to Danny that it
should be closed. I missed the RTFM sign... I didn't knew about
Maxim's link and would like to know where are we in transitioning
away from getting rid of boolean return code in phases.