Re: [sage-devel] sage.misc.package considered harmful

2017-04-18 Thread Nicolas M. Thiery
On Mon, Apr 03, 2017 at 05:48:30PM +1200, François Bissey wrote: > I am starting this debate because of discussion I had earlier > in #22670. It was pointed out to me that there was no policy > of avoiding `sage.misc.package` and I would very much want one. >From following the many discussions on

Re: [sage-devel] sage.misc.package considered harmful

2017-04-11 Thread Isuru Fernando
> I welcome the work on the Conda port because it will allow just that in the medium term. You could replace the whole sage packaging with Conda and a few scripts to set things up. Once you do that sage.misc.package should die (and please do not replace it by calls to Conda). FYI, I have a branch

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Francois Bissey
> On 5/04/2017, at 11:25, Nils Bruin wrote: > > On Tuesday, April 4, 2017 at 4:01:50 PM UTC-7, François wrote: > > With the current system you could install and then remove > some essential files manually and the doctesting framework > would still try to use it. It is installed according to t

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Nils Bruin
On Tuesday, April 4, 2017 at 4:01:50 PM UTC-7, François wrote: > > > With the current system you could install and then remove > some essential files manually and the doctesting framework > would still try to use it. It is installed according to the > packaging system after all. runtime testing

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Francois Bissey
> On 5/04/2017, at 10:41, Nils Bruin wrote: > > On Tuesday, April 4, 2017 at 2:24:52 PM UTC-7, François wrote: > Let’s be clear, I could ship a list of possible > optional packages supported in sage-on-gentoo > but any checking of package availability would > have to go through the the distri

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Nils Bruin
On Tuesday, April 4, 2017 at 2:24:52 PM UTC-7, François wrote: > > Let’s be clear, I could ship a list of possible > optional packages supported in sage-on-gentoo > but any checking of package availability would > have to go through the the distribution package > manager. > > Or through the sa

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Francois Bissey
> On 5/04/2017, at 08:42, Nils Bruin wrote: > > On Tuesday, April 4, 2017 at 1:00:31 PM UTC-7, François wrote: > (2) while being just one use, is probably not replaceable. > Not in this form at the very least. > > How can one get the appropriate information in sage-on-gentoo? If we compare >

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Nils Bruin
On Tuesday, April 4, 2017 at 1:00:31 PM UTC-7, François wrote: > > (2) while being just one use, is probably not replaceable. > Not in this form at the very least. > How can one get the appropriate information in sage-on-gentoo? If we compare the mechanisms that work in the different scenarios

Re: [sage-devel] sage.misc.package considered harmful

2017-04-04 Thread Francois Bissey
> On 5/04/2017, at 01:12, kcrisman wrote: > > > I am starting this debate because of discussion I had earlier > in #22670. It was pointed out to me that there was no policy > of avoiding `sage.misc.package` and I would very much want one. > > > This seems very reasonable, given how much wo

[sage-devel] sage.misc.package considered harmful

2017-04-02 Thread François Bissey
Hi all, This is a rare rant cross-posted to sage-packaging about one of the annoyance faced by the people packaging sage and not using sage's packaging system. The functions found in `sage.misc.package` are a result of the fact that sagemath and its development is closely linked to its own packa