Re: GSL in octave

[Julien Bect]

Hi Susi,


Le 20/07/2016 à 10:16, Susi Lehtola a écrit :
> On 07/19/2016 01:43 PM, Julien Bect wrote:
> > Most, really.  I haven't pushed (yet) a few minor things:
> >  * your changes to DESCRIPTION -> we'll change that when release 
> > approaches
> >  * your changes to the test files (loading package at the begining) ->
> > shouldn't we turn them into real unit tests?
>
> I have now added some trivial tests that check all of the generated 
> templates, doing which made me catch one final bug.

Ok, I pushed your test script.  But still, turning it into proper unit 
tests would be even better ;-)


> >  * changes to the cleaning rule in Makefile -> I'll look into this
> > later, perhaps should we call src/Makefile clean recursively?
>
> Possibly.

Ok, I have polished the "clean" rules of both Makefiles, see if you like it.


A now few more questions and remarks...


1) When I install the package I have this warning:

warning: doc_cache_create: unusable help text found in file 
'legendre_sphPlm_array'


2) There is a test script name test_ellint.m, I don't what it is 
supposed to do, but it doesn't work any more:

>> test_ellint
file_name = test_sf.c
source_id = -1
source_msg = Aucun fichier ou dossier de ce type
error: feof: invalid stream number = -1
error: called from
    READ_TEST_SF_ellint at line 72 column 3
    test_ellint at line 134 column 1

You removed test_sf.c in a previous changeset, so I assume this file 
should be removed too?


3) In the changeset "It's better to use row vectors instead" (really... 
can't you try to write more explicit commit messages ?) you modify all 
templates in such a way that all functions now return row vectors.  This 
is a little unexpected (and not documented).  For instance,

debye_1 (rand (3, 3))
ans =

   0.91844   0.84803   0.93841   0.88165   0.99141   0.79339 0.85979   
0.84939   0.94766

whereas I would have expected a 3 x 3 output.


4) Just an idea: for functions with several arguments, it would be 
really nice (but I don't how hard) to implement broadcasting:

conicalP_0 ([1 4 3]', [2 3])  # should return a 3 x 2 matrix as a result


5) In the changeset "Autogenerate PKG_ADD...",  you generate PKG_ADD 
from bootstrap, which is good since it eliminates redundancies.

But shouldn't you then remove PKG_ADD from the repo and add a clean rule 
for PKG_ADD in the Makefile?


6) PKG_ADD should perhaps be generated by configure itself, as suggested 
by Sebastian, in which case it would make more sense to put it in src, I 
think.