Duplicate package variables

Duplicate package variables

[Tobias Geerinckx-Rice]

[-- Attachment #1.1: Type: text/plain, Size: 750 bytes --]

Guix, John,

On 02/03/17 18:02, John Darrington wrote:
> commit f1021afb41613f5eba3ea4f5225b9a4cd1c15ecb
> Author: John Darrington <jmd@gnu.org>
> Date:   Thu Mar 2 17:57:25 2017 +0100
> 
>     gnu: Merge the two stellaria into one.

Nicely caught.

This made me curious and turned up three similar packages:

  ruby-arel	removed in 6f23e2fb023c921d03be1f3aef0e1f305ae63764
  r-codetools	removed in fe08b1a75953aa7153442dcffdcfd3a6d8c752a1
  r-gtable	removed in 9c53993fb279456f032a843f23fa546e90305910

There may be less obvious ones left.

I'm guessing there's no easy way to detect such shadowing from within
Scheme (and hence the linter), but I'd love to be wrong (and pressured
into adding it).

Kind regards,

T G-R


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 476 bytes --]

Re: Duplicate package variables

[John Darrington]

[-- Attachment #1: Type: text/plain, Size: 1440 bytes --]

On Fri, Mar 03, 2017 at 01:10:04AM +0100, Tobias Geerinckx-Rice wrote:
     Guix, John,
     
     On 02/03/17 18:02, John Darrington wrote:
     > commit f1021afb41613f5eba3ea4f5225b9a4cd1c15ecb
     > Author: John Darrington <jmd@gnu.org>
     > Date:   Thu Mar 2 17:57:25 2017 +0100
     > 
     >     gnu: Merge the two stellaria into one.
     
     Nicely caught.

Danny Milosavljevic made me aware of this.  So credit goes to him.

But yeah, I wondered why I didn't get an error or at least a warning
after I had added a second package with the same name as an existing 
one.  Not even lint complained.
     
     This made me curious and turned up three similar packages:
     
       ruby-arel	removed in 6f23e2fb023c921d03be1f3aef0e1f305ae63764
       r-codetools	removed in fe08b1a75953aa7153442dcffdcfd3a6d8c752a1
       r-gtable	removed in 9c53993fb279456f032a843f23fa546e90305910
     
     There may be less obvious ones left.
     
     I'm guessing there's no easy way to detect such shadowing from within
     Scheme (and hence the linter), but I'd love to be wrong (and pressured
     into adding it).
     
     Kind regards,
     
     T G-R
     




-- 
Avoid eavesdropping.  Send strong encrypted email.
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://sks-keyservers.net or any PGP keyserver for public key.


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: 01/01: gnu: Merge the two stellaria into one.

[Leo Famulari]

On Thu, Mar 02, 2017 at 12:02:40PM -0500, John Darrington wrote:
> jmd pushed a commit to branch master
> in repository guix.
> 
> commit f1021afb41613f5eba3ea4f5225b9a4cd1c15ecb
> Author: John Darrington <jmd@gnu.org>
> Date:   Thu Mar 2 17:57:25 2017 +0100
> 
>     gnu: Merge the two stellaria into one.
>     
>     * gnu/packages/astronomy.scm (stellaruim): Merge details from stellarium-0.14.
>     * gnu/packages/education.scm (stellarium): Remove variable.

This new version of stellarium is failing to build on i686-linux:

https://hydra.gnu.org/build/1853395

This is the failure report in the build log:

cd /tmp/guix-build-stellarium-0.15.1.drv-0/build/src && ./testExtinction
********* Start testing of TestExtinction *********
Config: Using QtTest library 5.6.2, Qt 5.6.2 (i386-little_endian-ilp32 shared (dynamic) release build; by GCC 4.9.4)
PASS   : TestExtinction::initTestCase()
FAIL!  : TestExtinction::testBase() 'mag==2.25' returned FALSE. ()
   Loc: [/tmp/guix-build-stellarium-0.15.1.drv-0/stellarium-0.15.1/src/tests/testExtinction.cpp(44)]
PASS   : TestExtinction::cleanupTestCase()
Totals: 2 passed, 1 failed, 0 skipped, 0 blacklisted
********* Finished testing of TestExtinction *********

Re: Duplicate package variables

[Ludovic Courtès]

John Darrington <john@darrington.wattle.id.au> skribis:

> On Fri, Mar 03, 2017 at 01:10:04AM +0100, Tobias Geerinckx-Rice wrote:
>      Guix, John,
>      
>      On 02/03/17 18:02, John Darrington wrote:
>      > commit f1021afb41613f5eba3ea4f5225b9a4cd1c15ecb
>      > Author: John Darrington <jmd@gnu.org>
>      > Date:   Thu Mar 2 17:57:25 2017 +0100
>      > 
>      >     gnu: Merge the two stellaria into one.
>      
>      Nicely caught.
>
> Danny Milosavljevic made me aware of this.  So credit goes to him.
>
> But yeah, I wondered why I didn't get an error or at least a warning
> after I had added a second package with the same name as an existing 
> one.  Not even lint complained.

The UI (e.g., “guix build stellarium”) emits a warning when referring to
a package that has two instances with the same name and version number.

Ludo’.

Re: Duplicate package variables

[Tobias Geerinckx-Rice]

[-- Attachment #1.1: Type: text/plain, Size: 588 bytes --]

Ludo',

Welcome back! :-)

On 08/03/17 18:11, Ludovic Courtès wrote:
> The UI (e.g., “guix build stellarium”) emits a warning when referring to
> a package that has two instances with the same name and version number.

Hm. Are you sure?

I've seen it do so when there are in fact multiple instances (i.e.
distinct variables), but not when the same variable is re-bound:

  ~/guix $ git checkout 9baab9aeeda394408290613c597bff1d96004dfb
  ~/guix $ ./pre-inst-env guix build ruby-arel r-gtable r-codetools

produces no such warning for me.

Kind regards,

T G-R


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 476 bytes --]

Re: Duplicate package variables

[Ludovic Courtès]

Hey Tobias,

Tobias Geerinckx-Rice <me@tobias.gr> skribis:

> On 08/03/17 18:11, Ludovic Courtès wrote:
>> The UI (e.g., “guix build stellarium”) emits a warning when referring to
>> a package that has two instances with the same name and version number.
>
> Hm. Are you sure?
>
> I've seen it do so when there are in fact multiple instances (i.e.
> distinct variables), but not when the same variable is re-bound:
>
>   ~/guix $ git checkout 9baab9aeeda394408290613c597bff1d96004dfb
>   ~/guix $ ./pre-inst-env guix build ruby-arel r-gtable r-codetools
>
> produces no such warning for me.

Oh right, if there’s just one variable, then there are no warnings.
Sorry for the confusion.

Ludo’.