On Aug 29, 2012, at 06:49, Mojca Miklavec wrote:
> On Wed, Aug 29, 2012 at 5:19 AM, Ryan Stonecipher wrote:
>> Some ports have +universal as a default variant.
>> When such ports have "big" dependencies such as qt4-mac, gcc4x, or
>> atlas they take hours to build and install.
>> The buildbot make
On Aug 29, 2012, at 22:40, Ryan Stonecipher wrote:
> This appears to be the issue on my computer as well.
> Virtualbox builds a kernel module for architectures listed by running uname
> -m.
Right. Virtualbox is special in that it builds a kernel module so it needs to
build for the architectur
This appears to be the issue on my computer as well.
Virtualbox builds a kernel module for architectures listed by running uname
-m.
Is this something that should be handled differently in Virtualbox which
calls itself non-universal in its portfile?
Ryan Stonecipher
On Wednesday, August 29, 2012,
On 2012-8-30 05:22 , Bradley Giesbrecht wrote:
>
> I was not suggesting to drop the 25 and 27 subports.
>
> The list thread "handle only subports install files" is where you
> "discouraged" (my words) useless top-level ports:
> On Mar 19, 2012, at 11:31 AM, Joshua Root wrote:
>> Again, why have
On Aug 29, 2012, at 11:33 AM, Joshua Root wrote:
> On 2012-8-30 03:14 , Bradley Giesbrecht wrote:
>> On Aug 29, 2012, at 5:25 AM, MacPorts wrote:
>>
>>> #35124: Python OAuth v1
>>> --+-
>>> Reporter: ctreleaven@… |
On 2012-8-29 23:33 , Rainer Müller wrote:
> Note that universal archives do not indicate which architectures they
> include in the file name and thus, we could not distinguish different
> architectures with the current code.
Actually, they do and we can. :-)
- Josh
___
On 2012-8-30 03:14 , Bradley Giesbrecht wrote:
> On Aug 29, 2012, at 5:25 AM, MacPorts wrote:
>
>> #35124: Python OAuth v1
>> --+-
>> Reporter: ctreleaven@… | Owner: macports-tickets@…
>>
>
On Aug 29, 2012, at 5:25 AM, MacPorts wrote:
> #35124: Python OAuth v1
> --+-
> Reporter: ctreleaven@… | Owner: macports-tickets@…
>
> Type: submission| Status: new
On Aug 28, 2012, at 10:40 PM, Ryan Schmidt wrote:
>
> On Aug 28, 2012, at 22:19, Ryan Stonecipher wrote:
>
>> Some ports have +universal as a default variant.
>> When such ports have "big" dependencies such as qt4-mac, gcc4x, or
>> atlas they take hours to build and install.
>> The buildbot make
Thanks Bradley, I've now updated both the heapCL and torchCL tickets
in trac with new portfiles that address these issues and use built in
linking instead of my own.
--
Ben Smith
Founder / CSA
WBP SYSTEMS
http://www.wbpsystems.com
On Tue, Aug 28, 2012 at 1:56 PM, Bradley Giesbrecht
wrote:
> App
On 2012-08-26 20:48, Jeremy Lavergne wrote:
> There are some ports that cannot build on specific os.major; is there a way
> we can have an accepted error message that the buildbot will not report as an
> error?
>
> Can we make it so `return -code error "unsupported OS X version"` is not an
> er
On 2012-08-29 13:49, Mojca Miklavec wrote:
> Or another buildbot which would try to build all the ports as universal ;)
> (or i386-only when 64-bit cannot be built ;)
The definition of +universal may vary from system to system since the
universal_archs config option can be changed in macports.conf
On Wed, Aug 29, 2012 at 5:19 AM, Ryan Stonecipher wrote:
> Some ports have +universal as a default variant.
> When such ports have "big" dependencies such as qt4-mac, gcc4x, or
> atlas they take hours to build and install.
> The buildbot makes installing default single-architecture binaries painles
On 2012-8-29 13:19 , Ryan Stonecipher wrote:
> Some ports have +universal as a default variant.
> When such ports have "big" dependencies such as qt4-mac, gcc4x, or
> atlas they take hours to build and install.
> The buildbot makes installing default single-architecture binaries painless.
> Could t
14 matches
Mail list logo
