Re: scilab: not O anymore?
On 02/11/2017 23:28, Julien Puydt wrote:
>
>>> - who's working on packaging 6.0.0?
>> Nobody afaik. It should not be too hard as it is mostly the core which
>> has beeen
>> translated from Fortran to C++
> Strange idea : I thought fortran was still faster.
As always it depends. The parser/lexer/execution was coded in Fortran.
Nobody was able to maintain the core anymore (for a while), preventing
languages changes, bug fixes in the interpreter, etc...
This is why this was redevelopped in C++.
Anyway, most of the time, when you perform computation, most of the time
is spent in the various functions, not really on the language steps. In
some cases,
it will be a bit slower.
The tradeoff was in favor of C++ here.
>>> - is it normal that a package under the Debian Science Team umbrella
>>> still uses svn?
>> I don't see that as a blocker but don't hesitate to migrate it to hg or git.
> That means removing the current git repository and then either finding
> the script used to move the svn repositories to git during the migration
> (keeping the whole history), or using "gbp import-dsc" (or better: gbp
> import-dscs) (keeping a partial history).
I don't think we care enough about the history of the package.
http://snapshot.debian.org/ should be enough.
>
>>> and the final one:
>>>
>>> - how can I help?
>> Trying to upload v6 should be fine.
>>
>> I can help as I wrote the upstream build system and the debian packages!
> I'm currently having a look at how much the existing patches have to be
> changed for 6.0.0. In fact, I'm surprised there are so much of them: why
> aren't they upstream?
Different versions, because the 5 branch is no longer maintained,
because I am lazy, etc.
> The fact that git.scilab.org and bugzilla.scilab.org aren't available
> doesn't help :-/
>
I have a mirror here: https://github.com/opencollab/scilab
Anyway, if you drop the ball on this, please tell me. I will see what I
can do!
S
Reply to: