Re: pkg() in 4.0.0-rc1 breaks building of some packages

[Carnë Draug]

On 10 March 2015 at 15:47, Olaf Till <address@hidden> wrote:
> On Tue, Mar 10, 2015 at 10:23:11AM -0400, Mike Miller wrote:
>> On Tue, Mar 10, 2015 at 15:19:26 +0100, Juan Pablo Carbajal wrote:
>> > On Tue, Mar 10, 2015 at 2:37 PM, Mike Miller <address@hidden> wrote:
>> > > As the end user, you can call `setenv OMP_NUM_THREADS 1` in the Octave
>> > > shell before doing the pkg install, that should translate to --jobs=1.
>> > > So in this sense it is already an option that the end user has control
>> > > over.
>> > >
>> > This is good to know! but as you see, it is not cristal clear for
>> > everybody. I think pkg() could be more useful if it provided a
>> > solution...maybe just calling the line you just provided.
>>
>> Sure, I don't think there's anything wrong with adding an option to wrap
>> that setting.
>>
>> > I personally think pkg() needs a refactoring and I was involved in it
>> > while it lasted. I will behappy to give it another go if we believe
>> > pkg() will be the "inside octave" tool for the user to handle
>> > packages.
>>
>> Completely agree. I remember you guys working on that project and still
>> hope that it makes its way back to us!
>>
>> I was just pointing out that there are solutions for Olaf's problem and
>> that pkg probably doesn't need to be urgently changed for 4.0.
>
> For the record, I didn't post this because it's "my problem"; the
> Makefile is already changed, and a decent Makefile probably shouldn't
> have this problem. The point is that the _default_ should be "no
> parallel builds", since pkg() can be used to compile any package, not
> only those of OF.

Parallel builds will work unless the Makefile is not correctly written,
in which the problem is on the package, not on Octave.  pkg() can install
any package, even using parallel builds.  Provided the package is
correctly written.  That has nothing to do with OF.

> My suggestion is to just drop the --jobs option unconditionally, this
> change is small, a few lines to be deleted. Most packages are not so
> large that parallel building is a great advantage, so it's barely
> worth to have an additional pkg option for it -- IMHO.

The whole reason why this got implemented was because users complained that
building packages was taking too long [1].  Should we slow it down again,
when the only advantage is to allow installation of packages with
buggy Makefiles?  I really don't think so.  Or is there any other
advantage?

Carnë

[1] https://savannah.gnu.org/bugs/?42915