On Tue, Nov 06, 2007 at 05:49:11PM +0200, Riku Voipio wrote:
As I see what Neil wrote, nocheck (I prefer this over notest) affects
only testsuite. As for *handling* packages that run compiled code during
build.. well that's the reason why scratchbox was invented :)
I'm confused about how
This is one of the changes sought by Emdebian to support using Debian
on embedded devices where storage space is far from cheap and involves
running counter to the current Debian default of install everything
that works, every time, every package.
nocheck | notest should also be supported whether or not the package is
actually being cross-built. emdebuild currently passes 'nocheck' as
part of the DEB_BUILD_OPTIONS for all Emdebian builds.
Imho this should be either nocheck or notest, supporting both is a bit
pointless. Using notest or
On Tue, Nov 06, 2007 at 02:42:03PM +0100, Bernd Zeimetz wrote:
nocheck | notest should also be supported whether or not the package is
actually being cross-built. emdebuild currently passes 'nocheck' as
part of the DEB_BUILD_OPTIONS for all Emdebian builds.
Imho this should be either
On Tue, 06 Nov 2007 14:42:03 +0100
Bernd Zeimetz [EMAIL PROTECTED] wrote:
nocheck | notest should also be supported whether or not the package is
actually being cross-built. emdebuild currently passes 'nocheck' as
part of the DEB_BUILD_OPTIONS for all Emdebian builds.
Imho this should
2007/11/6, Neil Williams [EMAIL PROTECTED]:
There needs to be some agreement on what nocheck or notest means and
which one to use. For Emdebian needs, whichever name is used, the
imperative is that setting that DEB_BUILD_OPTION *must* completely
prevent the execution of any compiled program
Zitat von Miriam Ruiz [EMAIL PROTECTED]:
2007/11/6, Neil Williams [EMAIL PROTECTED]:
There needs to be some agreement on what nocheck or notest means and
which one to use. For Emdebian needs, whichever name is used, the
imperative is that setting that DEB_BUILD_OPTION *must* completely
prevent
Michael Banck wrote:
On Tue, Nov 06, 2007 at 02:42:03PM +0100, Bernd Zeimetz wrote:
nocheck | notest should also be supported whether or not the package is
actually being cross-built. emdebuild currently passes 'nocheck' as
part of the DEB_BUILD_OPTIONS for all Emdebian builds.
Imho this
On Tue, 6 Nov 2007 15:31:49 +0100
Miriam Ruiz [EMAIL PROTECTED] wrote:
2007/11/6, Neil Williams [EMAIL PROTECTED]:
There needs to be some agreement on what nocheck or notest means and
which one to use. For Emdebian needs, whichever name is used, the
imperative is that setting that
Miriam Ruiz wrote:
2007/11/6, Neil Williams [EMAIL PROTECTED]:
There needs to be some agreement on what nocheck or notest means and
which one to use. For Emdebian needs, whichever name is used, the
imperative is that setting that DEB_BUILD_OPTION *must* completely
prevent the execution of any
On Tue, Nov 06, 2007 at 03:31:49PM +0100, Miriam Ruiz wrote:
2007/11/6, Neil Williams [EMAIL PROTECTED]:
There needs to be some agreement on what nocheck or notest means and
which one to use. For Emdebian needs, whichever name is used, the
imperative is that setting that DEB_BUILD_OPTION
On Tue, 6 Nov 2007 15:53:05 +0100
Michael Banck [EMAIL PROTECTED] wrote:
On Tue, Nov 06, 2007 at 02:42:03PM +0100, Bernd Zeimetz wrote:
nocheck | notest should also be supported whether or not the package is
actually being cross-built. emdebuild currently passes 'nocheck' as
part of the
On Tue, 6 Nov 2007 15:53:05 +0100
Michael Banck [EMAIL PROTECTED] wrote:
I wanted to get around to introduce nocheck in a couple more packages,
but maybe it's better to just settle on it now and propose it for policy
inclusion for lenny.
Actually, Guillem has already filed the bug: 416450
Neil Williams, 2007-11-06 16:08:11 + :
Actually, Guillem has already filed the bug: 416450 [PROPOSAL] New
option in DEB_BUILD_OPTIONS to avoid running test-suites
What needs to happen for that to be mandatory in Lenny?
Get the option widely used, then documented (as a MUST) in policy,
On Tue, Nov 06, 2007 at 04:08:11PM +, Neil Williams wrote:
On Tue, 6 Nov 2007 15:53:05 +0100
Michael Banck [EMAIL PROTECTED] wrote:
I wanted to get around to introduce nocheck in a couple more packages,
but maybe it's better to just settle on it now and propose it for policy
Roland Mas [EMAIL PROTECTED] (06/11/2007):
Get the option widely used, then documented (as a MUST) in policy,
then agreed on as a release goal (or fix the bugs even if they're not
RC :-)
AFAICT, it is sufficient to convince RMs to make it an RC goal. An
explicit MUST in the Policy isn't
On Tue, Nov 06, 2007 at 01:18:25PM +, Neil Williams wrote:
This is one of the changes sought by Emdebian to support using Debian
on embedded devices where storage space is far from cheap and involves
running counter to the current Debian default of install everything
that works, every
Neil Williams wrote:
In this context, I believe package documentation should mean:
All files in the package that are installed beneath /usr/share/doc
which are not mandated by Policy.
Therefore, copyright and changelogs are excluded as are manpage and
info pages but README, TODO, AUTHORS
Neil Williams wrote:
I propose to file bugs against packages that use inconsistent
DEB_BUILD_OPTIONS or which do not support DEB_BUILD_OPTIONS that would
actually benefit Emdebian.
As with the other mass bug filing from this set, I will tag the reports
'crossbuilt' and file as wishlist.
On Tue, Nov 06, 2007 at 05:34:36PM +0100, Roland Mas wrote:
Neil Williams, 2007-11-06 16:08:11 + :
Actually, Guillem has already filed the bug: 416450 [PROPOSAL] New
option in DEB_BUILD_OPTIONS to avoid running test-suites
What needs to happen for that to be mandatory in Lenny?
Get
On Tue, 6 Nov 2007 14:25:54 -0500
Joey Hess [EMAIL PROTECTED] wrote:
Neil Williams wrote:
I propose to file bugs against packages that use inconsistent
DEB_BUILD_OPTIONS or which do not support DEB_BUILD_OPTIONS that would
actually benefit Emdebian.
I wonder if I should spend some time
Mark Brown wrote:
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
The size of the test suite is completely irrelevant. Executing a single
compiled binary will cause any cross build to break, even if it is a
1kb no-op.
Attempting to execute a binary not supported by the host
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
On Tue, 6 Nov 2007 14:37:56 -0500
Steve Langasek [EMAIL PROTECTED] wrote:
And things shouldn't be must in policy unless they're intended to be RC
bugs. I don't see how this would ever make sense to be a must.
Why? Any test
On Tue, 6 Nov 2007 21:36:31 +
Mark Brown [EMAIL PROTECTED] wrote:
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
The size of the test suite is completely irrelevant. Executing a single
compiled binary will cause any cross build to break, even if it is a
1kb no-op.
On Tue, 6 Nov 2007 22:08:40 +0100
Julien Cristau [EMAIL PROTECTED] wrote:
On Tue, Nov 6, 2007 at 20:54:29 +, Neil Williams wrote:
On Tue, 6 Nov 2007 14:37:56 -0500
Steve Langasek [EMAIL PROTECTED] wrote:
And things shouldn't be must in policy unless they're intended to be RC
On Tue, Nov 6, 2007 at 20:54:29 +, Neil Williams wrote:
On Tue, 6 Nov 2007 14:37:56 -0500
Steve Langasek [EMAIL PROTECTED] wrote:
And things shouldn't be must in policy unless they're intended to be RC
bugs. I don't see how this would ever make sense to be a must.
Why? Any test
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
The size of the test suite is completely irrelevant. Executing a single
compiled binary will cause any cross build to break, even if it is a
1kb no-op.
Attempting to execute a binary not supported by the host will obviously
fail
* Gabor Gombas [Tue, 06 Nov 2007 22:51:33 +0100]:
I wonder if this is the wrong approach. You want to add extra complexity
to _every_ package for the benefit of only a small user base. Instead,
why not patch dpkg-deb -b in Emdebian to interpret -nodoc as leave
out everything under
On Tue, Nov 06, 2007 at 10:48:55PM +0100, Bernd Zeimetz wrote:
If you run a test you usually *want* to have it fail on build time if
the test did not go trough. Otherwise running tests is useless - or do
you check all build logs with every upload?
In the particular case I'm thinking of the
Mark Brown wrote:
On Tue, Nov 06, 2007 at 10:48:55PM +0100, Bernd Zeimetz wrote:
If you run a test you usually *want* to have it fail on build time if
the test did not go trough. Otherwise running tests is useless - or do
you check all build logs with every upload?
In the particular case
On Tue, Nov 06, 2007 at 09:46:52PM +, Neil Williams wrote:
Mark Brown [EMAIL PROTECTED] wrote:
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
The size of the test suite is completely irrelevant. Executing a single
compiled binary will cause any cross build to break,
On Wed, Nov 07, 2007 at 12:19:52AM +0100, Bernd Zeimetz wrote:
So there should be a way to disable them if the tests can't work in the
current build environment.
Don't get me wrong - I think the basic idea is a sensible one, it's just
the making it mandatory bit.
--
You grabbed my hand and
On Tue, Nov 06, 2007 at 11:16:53PM +, Mark Brown wrote:
On Tue, Nov 06, 2007 at 09:46:52PM +, Neil Williams wrote:
Mark Brown [EMAIL PROTECTED] wrote:
On Tue, Nov 06, 2007 at 08:54:29PM +, Neil Williams wrote:
The size of the test suite is completely irrelevant. Executing a
33 matches
Mail list logo