Re: Problems with downloading from https

2014-10-30 Thread Alex Kost
Andreas Enge (2014-10-30 01:22 +0300) wrote:

> On Tue, Oct 28, 2014 at 09:03:44AM +0100, Ludovic Courtès wrote:
>> Alex Kost  skribis:
>> > But currently it's not possible to install 2 (or more) packages with the
>> > same name.  So a user can't have guile 2.0 and guile 1.8 in the same
>> > profile.  The same thing with python: there is no ‘python2’ package.
>> > Both python packages have “python” name and can't be installed in the
>> > same profile, as far as I understand.
>> 
>> Good point! (Somehow I thought there was ‘python2’.)
>
> Is it really not possible to install two packages with the same name?
> I thought you could do
>guix package -i python python-2.7.6
> where the first one will chose the latest package? It just requires that
> there are no overlapping paths.

I think such an "evil" case is just not handled currently.  If you have
python-3… installed and you install python-2… in the same profile, then
python-3… would be replaced, but if you install both packages in the
same command, then both would be installed.  It's OK for pythons,
because there are no name collisions in these packages, but if you try
to install both "guile-2.0.11" and "guile-1.8.8" is such a way, then you
will have several collisions.

So I think installing packages with the same name in one transaction
should also be prohibited.

As both python packages can co-exist in one profile, either python-2…
may be renamed into “python2” or python-3… into “python3”.  As python3
is the future, I think it would be better to have “python2” and “python”
(which is python3) packages.  Or maybe they shouldn't be renamed and we
can introduce a little collision instead by adding "…/bin/python"
symlink to python-3… package.



Re: [PATCH 2/2] gnu: Add 'ttf-liberation'.

2014-10-30 Thread Alex Kost
Andreas Enge (2014-10-30 01:16 +0300) wrote:

> On Tue, Oct 28, 2014 at 09:10:30AM +0100, Ludovic Courtès wrote:
>> Alex Kost  skribis:
>> > Not related to this patch: what about renaming ‘freefont-ttf’ package
>> > into ‘ttf-freefont’ to make all TrueType fonts have a name "ttf-…"?
>> I think so.  What do others think?  Andreas?
>
> So far, we have no special rules for font naming. So...
>
>> > Subject: [PATCH 2/2] gnu: Add 'ttf-liberation'.
>> > * gnu/packages/fonts.scm (ttf-liberation): New variable.
>
> ...the package should be called liberation-fonts-ttf.

Why should?  What about “ttf-dejavu”?  Should it be called
“dejavu-fonts-ttf” then?

> We could install a rule similar to that for python modules, that we always
> add ttf in front of true type fonts, or move it in front if it is already
> in the name. But I am not convinced this is a good choice: Being truetype
> is not inherent to a font, it is just a format, and a package may contain
> the same font (or several ones) in different formats. What would we do then?

Then it shouldn't be prefixed with "ttf".  I suggest to use "ttf-…" name
only for packages that provide truetype fonts only.  As for the other
font packages, I think they should have "font"/"fonts" in their names,
no matter would it be in the beginning (font-misc-ethiopic) or in the
end (terminus-font).

> It would be more reasonable to add "font-" in front. But I am also not too
> fond of this.
>
> Andreas



Re: Problems with downloading from https

2014-10-30 Thread Andreas Enge
On Thu, Oct 30, 2014 at 10:27:59AM +0300, Alex Kost wrote:
> I think such an "evil" case is just not handled currently.  If you have
> python-3… installed and you install python-2… in the same profile, then
> python-3… would be replaced, but if you install both packages in the
> same command, then both would be installed.

Interesting, I did not know this, but you are right:

$ guix package -i python
$ guix package -i python-2.7.6
The following package will be upgraded:
   python   3.3.5 → 2.7.6   
/gnu/store/iz5shg4py68mbccv2kkd0siv6ryfl3y1-python-2.7.6

If you do instead
$ guix package -i python python-2.7.6

both packages are installed.

I think the former behaviour is a bug. If I use "-i" and not "-u", a package
should not be "upgraded", but added in every case, independently of its name.

Andreas




Re: Texlive and native-inputs

2014-10-30 Thread John Darrington
On Wed, Oct 29, 2014 at 11:31:18PM +0100, Andreas Enge wrote:
 On Wed, Oct 29, 2014 at 08:55:55PM +0100, John Darrington wrote:
 > What do these scripts do anyway?  Are they fundamental to TeXlive or are 
they for some bells and whistles?
 
 I never use them as far as I know, so they do not seem to be fundamental.
 
In that case, it would seem to me, that perl should not be in input at all.

-- 
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.



signature.asc
Description: Digital signature


Re: [PATCH 2/2] gnu: Add 'ttf-liberation'.

2014-10-30 Thread Andreas Enge
On Thu, Oct 30, 2014 at 10:27:49AM +0300, Alex Kost wrote:
> Why should?  What about “ttf-dejavu”?  Should it be called
> “dejavu-fonts-ttf” then?

Yes, that was a mistake, maybe inspired from the debian package name.
We should rename it.

> Then it shouldn't be prefixed with "ttf".  I suggest to use "ttf-…" name
> only for packages that provide truetype fonts only.  As for the other
> font packages, I think they should have "font"/"fonts" in their names,
> no matter would it be in the beginning (font-misc-ethiopic) or in the
> end (terminus-font).

Then what if it contains other types of fonts? Texlive-data/texfm-dist/fonts
contains the following 20 subdirectories:
afm  cmap  fea  map   ofm   ovf  pfm  sfd tfm   type1
cid  enc   lig  misc  opentype  ovp  pk   source  truetype  vf
(not all of them are fonts, some are just metrics and some I do not know).
Would you suggest to use prefixes type1-, opentype-, pk- also?
Only if the package contains exactly one format? I think modifying our
package name rules for fonts will open a can of worms.

Andreas




Re: Texlive and native-inputs

2014-10-30 Thread Andreas Enge
On Thu, Oct 30, 2014 at 08:53:24AM +0100, John Darrington wrote:
> In that case, it would seem to me, that perl should not be in input at all.

It should. Perl is needed for "make check". But even if it were not, we
always patch-shebang script files in our packages. Otherwise they would not
be usable, since we do not have /usr/bin/perl. That I personally do not
use these texlive scripts is not a reason to make them unusable!

Now if this has not been done for octave, maybe we should rediscuss it and
add the needed script interpreters to its inputs.

Andreas




Re: Texlive and native-inputs

2014-10-30 Thread John Darrington
On Thu, Oct 30, 2014 at 09:02:10AM +0100, Andreas Enge wrote:
 On Thu, Oct 30, 2014 at 08:53:24AM +0100, John Darrington wrote:
 > In that case, it would seem to me, that perl should not be in input at 
all.
 
 It should. Perl is needed for "make check". But even if it were not, we
 always patch-shebang script files in our packages. Otherwise they would not
 be usable, since we do not have /usr/bin/perl. That I personally do not
 use these texlive scripts is not a reason to make them unusable!
 

It could be patched to "perl".  That way, so long as the user has done 
guix package -i perl  it will work for him.
 


-- 
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.



signature.asc
Description: Digital signature