on current linux distros, what are potential candidates for ASSUME_PROVIDED?

on current linux distros, what are potential candidates for ASSUME_PROVIDED?

[Robert P. J. Day]

  for the sake of reducing build time in the classroom, what are some
of the potential (and relatively safe) candidates to add to
ASSUME_PROVIDED for a build from scratch?

  i do recall that, once one starts adding more native packages to
that variable, QA becomes an issue but, given a modern distro such as
fedora 21, surely there are quite a number of packages that are safe
to add.

  the first thing i always add is "subversion-native" (since i don't
think subversion is even *used* in a core-image-minimal build, but i
could be wrong). i'm looking through all the -native builds and can
see other possible candidates i'm going to try -- open to suggestions
as to what should be safe, and i'll test it out.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: on current linux distros, what are potential candidates for ASSUME_PROVIDED?

[Richard Purdie]

On Wed, 2015-02-25 at 03:25 -0500, Robert P. J. Day wrote:
>   for the sake of reducing build time in the classroom, what are some
> of the potential (and relatively safe) candidates to add to
> ASSUME_PROVIDED for a build from scratch?

This path is fraught with danger to be honest. There are some things
which are "safe" like subversion and git but they don't make that much
difference to a build and there are not as many as you'd think.

The biggest difference you can make is an sstate cache you share amongst
the pupils. The time is spent:

a) building gcc
b) building libc
c) building gettext
d) building glib

each of these is a bottle neck which then opens up a new set of targets
as none of them are ASSUME_PROVIDED material.

Cheers,

Richard

Re: on current linux distros, what are potential candidates for ASSUME_PROVIDED?

[Robert P. J. Day]

On Wed, 25 Feb 2015, Richard Purdie wrote:

> On Wed, 2015-02-25 at 03:25 -0500, Robert P. J. Day wrote:
> >   for the sake of reducing build time in the classroom, what are some
> > of the potential (and relatively safe) candidates to add to
> > ASSUME_PROVIDED for a build from scratch?
>
> This path is fraught with danger to be honest. There are some things
> which are "safe" like subversion and git but they don't make that much
> difference to a build and there are not as many as you'd think.
>
> The biggest difference you can make is an sstate cache you share amongst
> the pupils. The time is spent:
>
> a) building gcc
> b) building libc
> c) building gettext
> d) building glib
>
> each of these is a bottle neck which then opens up a new set of targets
> as none of them are ASSUME_PROVIDED material.

  ok, fair enough.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: on current linux distros, what are potential candidates for ASSUME_PROVIDED?

[Stephen Arnold]

[-- Attachment #1: Type: text/plain, Size: 2345 bytes --]

I would agree that for classroom purposes, some of the best time savers are
things like pre-fetching downloads to a local machine, pre-cloning kernel
repos, and shared cache.

Also, IIRC several of the course lab examples (in the LF Yocto course)
didn't quite work as expected (mostly kernel variables for some reason).  I
would probably recommend keeping things simple, as in teaching the config
fragment approach (as is done in the kernel lab training thing).  Can't
remember what the all the issues were exactly as I normally build off
master which gives me recurring deja vu, but keeping course materials
current is always an ongoing issue for me...

Just my $.02 ...

Steve

On Wed, Feb 25, 2015 at 12:41 AM, Robert P. J. Day <rpjday@crashcourse.ca>
wrote:

> On Wed, 25 Feb 2015, Richard Purdie wrote:
>
> > On Wed, 2015-02-25 at 03:25 -0500, Robert P. J. Day wrote:
> > >   for the sake of reducing build time in the classroom, what are some
> > > of the potential (and relatively safe) candidates to add to
> > > ASSUME_PROVIDED for a build from scratch?
> >
> > This path is fraught with danger to be honest. There are some things
> > which are "safe" like subversion and git but they don't make that much
> > difference to a build and there are not as many as you'd think.
> >
> > The biggest difference you can make is an sstate cache you share amongst
> > the pupils. The time is spent:
> >
> > a) building gcc
> > b) building libc
> > c) building gettext
> > d) building glib
> >
> > each of these is a bottle neck which then opens up a new set of targets
> > as none of them are ASSUME_PROVIDED material.
>
>   ok, fair enough.
>
> rday
>
> --
>
> ========================================================================
> Robert P. J. Day                                 Ottawa, Ontario, CANADA
>                         http://crashcourse.ca
>
> Twitter:                                       http://twitter.com/rpjday
> LinkedIn:                               http://ca.linkedin.com/in/rpjday
> ========================================================================
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
>

[-- Attachment #2: Type: text/html, Size: 3351 bytes --]