octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: very small packages - merge into general/miscelleneous or move into


From: c.
Subject: Re: very small packages - merge into general/miscelleneous or move into core
Date: Fri, 24 Jan 2014 08:11:56 +0100

On 23 Jan 2014, at 20:41, Thomas Weber <address@hidden> wrote:

> On Thu, Jan 23, 2014 at 04:46:45PM +0100, c. wrote:
>> 
>> On 23 Jan 2014, at 13:34, Thomas Weber <address@hidden> wrote:
>> 
>>>> That's why I whish Debian packagers would let pkg.m work 
>>>> so Octave users could just decide for themselves.
>>> 
>>> We share that wish. Have you a fix for 
>>> http://savannah.gnu.org/bugs/?32049 ?
>> 
>> I'm not sure this is exactly the same, but is this problem related to the 
>> one discussed here?
>> 
>> http://octave.1599824.n4.nabble.com/Ubuntu-12-04-pkg-install-does-not-work-tp4657080p4657093.html
>> 
>> If I compile Octave myself on ubuntu without applying that change pkg.m 
>> works as expected.
> 
> Yes, the issue is shortly described. Once you have set the prefix in
> pkg, you cannot change it anymore as it is a persistent variable in
> pkg.m (and this overrides the -local switch). We need to set it however
> once to tell Octave where the system packages are. 

If what you need is a way to set the default path for globally installed
packages as a configure option I think I can produce an easy fix.

> If my memory serves correctly, Octave used to include its own version
> number in the packages' path, which was the main problem - an update
> from say 3.6.3 to 3.6.4 would mean that Octave would not find the
> installed packages anymore with the default settings. 

I think the current policy is to try and not break ABI with 
whith a change of the third digit in the version number, so 
probably the name should contain only two digits?

> I don't have time right now, but maybe with the refactored pkg.m we
> might be able to overcome this by patching pkg.m directly (which we have
> avoided in the past). 

I'm preparing a patch to make this approach easier.

>       Thomas

c.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]