[RFC] error out on .la files pointing to ${WORKDIR}

[RFC] error out on .la files pointing to ${WORKDIR}

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and  
should die. If we want packaged-staging and rm_work to function we  
need to fix them. Since we are all lazy we need some incentive:

#
# old_revision [85e676eb4b64d66cb74acf6220f10c3942e2bd29]
#
# patch "classes/insane.bbclass"
#  from [0712df99743c442ed311e6b4d4be0d28f29d525f]
#    to [fa6b704f402965c8123dfa90527b5a11a8cd3288]
#
============================================================
- --- classes/insane.bbclass      0712df99743c442ed311e6b4d4be0d28f29d525f
+++ classes/insane.bbclass      fa6b704f402965c8123dfa90527b5a11a8cd3288
@@ -178,7 +178,7 @@ def package_qa_make_fatal_error(error_cl

      TODO: Load a whitelist of known errors
      """
- -    return not error_class in [0, 5, 7, 8]
+    return not error_class in [0, 5, 7]

  def package_qa_write_error(error_class, name, path, d):
      """

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFH4PAUMkyGM64RGpERAvkBAJ94969GpNP8493ezmnmV/wkkKd7mgCgqGNG
OKIsbKTFKRQbNI06cnE/32A=
=0rQM
-----END PGP SIGNATURE-----

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Florian Boor]

Hi,

Koen Kooi schrieb:
> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
> should die. If we want packaged-staging and rm_work to function we
> need to fix them. Since we are all lazy we need some incentive:

I'm not really a fan of enforcing things by breaking builds. I agree that we
need to fix this soon, but since some of this bugs are harmless and do not
affect the build at all I would not want to handle this that strict. If these
wrong files really break packaged-staging we get errors that force us to fix the
critical ones anyway.

Greetings

Florian

-- 
The dream of yesterday                  Florian Boor
is the hope of today                    Tel: +49 271-771091-15
and the reality of tomorrow.            Fax: +49 271-771091-19
[Robert Hutchings Goddard, 1904]        florian.boor@kernelconcepts.de

1D78 2D4D 6C53 1CA4 5588  D07B A8E7 940C 25B7 9A76

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Florian Boor schreef:
| Hi,
|
| Koen Kooi schrieb:
|> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
|> should die. If we want packaged-staging and rm_work to function we
|> need to fix them. Since we are all lazy we need some incentive:
|
| I'm not really a fan of enforcing things by breaking builds.

They are already breaking my builds with rm_work enabled.

| I agree that we
| need to fix this soon, but since some of this bugs are harmless and do not
| affect the build at all

There are harmfull and are already breaking builds for everyone using
rm_work.

| I would not want to handle this that strict. If these
| wrong files really break packaged-staging we get errors that force us
to fix the
| critical ones anyway.

No, people went out and bought bigger disks so they could disable rm_work.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFH4Pp5MkyGM64RGpERAvmoAKCOoE+xLwW0Nzl41jiQq9dTRZp/cACeJwLD
MCGaT9yGV8f5DmJUW+KaxRU=
=FP1S
-----END PGP SIGNATURE-----

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Richard Purdie]

On Wed, 2008-03-19 at 12:24 +0100, Florian Boor wrote:
> Koen Kooi schrieb:
> > Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
> > should die. If we want packaged-staging and rm_work to function we
> > need to fix them. Since we are all lazy we need some incentive:
> 
> I'm not really a fan of enforcing things by breaking builds. I agree that we
> need to fix this soon, but since some of this bugs are harmless and do not
> affect the build at all I would not want to handle this that strict. If these
> wrong files really break packaged-staging we get errors that force us to fix the
> critical ones anyway.

If you don't want the build to break just disable the insane checker.
Whilst Angstrom defaults to it being on, you can disable it...

Cheers,

Richard

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Stanislav Brabec]

Koen Kooi wrote:

> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and  
> should die. If we want packaged-staging and rm_work to function we  
> need to fix them. Since we are all lazy we need some incentive:

This policy is applied in OpenSUSE for several years. So if any packages
are broken upstream, you will find a fix in OpenSUSE nearly for sure.

-- 
Best Regards / S pozdravem,

Stanislav Brabec
software developer
---------------------------------------------------------------------
SUSE LINUX, s. r. o.                          e-mail: sbrabec@suse.cz
Lihovarská 1060/12           tel: +420 284 028 966, +49 911 740538747
190 00 Praha 9                                  fax: +420 284 028 951
Czech Republic                                    http://www.suse.cz/

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Stanislav Brabec schreef:
| Koen Kooi wrote:
|
|> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
|> should die. If we want packaged-staging and rm_work to function we
|> need to fix them. Since we are all lazy we need some incentive:
|
| This policy is applied in OpenSUSE for several years. So if any packages
| are broken upstream, you will find a fix in OpenSUSE nearly for sure.

Great! Do you have a link to the patch repository?

regards,

Koen


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFH4Qh5MkyGM64RGpERApkUAKC7p4KC+hkhMY9zbDksowMUBVDrygCgsPHD
iuCKVlnsFJOonlrGTKAFzXE=
=cuBG
-----END PGP SIGNATURE-----

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Stanislav Brabec]

Koen Kooi wrote:
> Stanislav Brabec schreef:
> | Koen Kooi wrote:
> |
> |> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
> |> should die. If we want packaged-staging and rm_work to function we
> |> need to fix them. Since we are all lazy we need some incentive:
> |
> | This policy is applied in OpenSUSE for several years. So if any packages
> | are broken upstream, you will find a fix in OpenSUSE nearly for sure.
> 
> Great! Do you have a link to the patch repository?

It is still in TODO of the OpenSUSE team. Now you have two chances:

1) Download the whole source and extract patches:
http://download.opensuse.org/distribution/SL-OSS-factory/inst-source/suse/src/

And then unpacking (for RPM non-native platforms you can use alien or
rpm2cpio).

2) With (free) OpenSUSE Build Service account you can access:
https://build.opensuse.org/project/show?project=openSUSE%3AFactory

-- 
Best Regards / S pozdravem,

Stanislav Brabec
software developer
---------------------------------------------------------------------
SUSE LINUX, s. r. o.                          e-mail: sbrabec@suse.cz
Lihovarská 1060/12           tel: +420 284 028 966, +49 911 740538747
190 00 Praha 9                                  fax: +420 284 028 951
Czech Republic                                    http://www.suse.cz/

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Richard Purdie]

Hi,

On Wed, 2008-03-19 at 11:50 +0100, Koen Kooi wrote:
> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and  
> should die. If we want packaged-staging and rm_work to function we  
> need to fix them. Since we are all lazy we need some incentive:
[...]
> --- classes/insane.bbclass      0712df99743c442ed311e6b4d4be0d28f29d525f
> +++ classes/insane.bbclass      fa6b704f402965c8123dfa90527b5a11a8cd3288
> @@ -178,7 +178,7 @@ def package_qa_make_fatal_error(error_cl
> 
>       TODO: Load a whitelist of known errors
>       """
> - -    return not error_class in [0, 5, 7, 8]
> +    return not error_class in [0, 5, 7]
> 
>   def package_qa_write_error(error_class, name, path, d):
>       """

I'm ok with applying that since we do really need to sort out these
issues.

I did a load of cleanup in this area in poky and found in a lot of cases
this happened since we weren't using oe_libinstall which mangles the .la
files if it notices anything wrong with them. There is at least one fix
in Poky which might help the gstreamer packages from this PoV and I'll
try and merge that.

Cheers,

Richard

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Richard Purdie schreef:
| Hi,
|
| On Wed, 2008-03-19 at 11:50 +0100, Koen Kooi wrote:
|> Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
|> should die. If we want packaged-staging and rm_work to function we
|> need to fix them. Since we are all lazy we need some incentive:
| [...]
|> --- classes/insane.bbclass      0712df99743c442ed311e6b4d4be0d28f29d525f
|> +++ classes/insane.bbclass      fa6b704f402965c8123dfa90527b5a11a8cd3288
|> @@ -178,7 +178,7 @@ def package_qa_make_fatal_error(error_cl
|>
|>       TODO: Load a whitelist of known errors
|>       """
|> - -    return not error_class in [0, 5, 7, 8]
|> +    return not error_class in [0, 5, 7]
|>
|>   def package_qa_write_error(error_class, name, path, d):
|>       """
|
| I'm ok with applying that since we do really need to sort out these
| issues.
|
| I did a load of cleanup in this area in poky and found in a lot of cases
| this happened since we weren't using oe_libinstall which mangles the .la
| files if it notices anything wrong with them.

Richard and I have now finally fixed the native.bbclass staging logic
for autotoolws packages. This means virtually all -native autotooled
recipes now get sanitized automagically.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFH4TB3MkyGM64RGpERAnvLAKC0++o3ntLzOjqv5pmrbcLrFl4pQwCgmRAj
9EL16VVjDtR54mR2axB8dPA=
=FemV
-----END PGP SIGNATURE-----

Re: [RFC] error out on .la files pointing to ${WORKDIR}

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Koen Kooi schreef:
| Hi,
|
| Things in stagin (e.g. .la files) point to ${WORKDIR} are broken and
| should die. If we want packaged-staging and rm_work to function we
| need to fix them. Since we are all lazy we need some incentive:
|
| #
| # old_revision [85e676eb4b64d66cb74acf6220f10c3942e2bd29]
| #
| # patch "classes/insane.bbclass"
| #  from [0712df99743c442ed311e6b4d4be0d28f29d525f]
| #    to [fa6b704f402965c8123dfa90527b5a11a8cd3288]
| #
| ============================================================
| --- classes/insane.bbclass      0712df99743c442ed311e6b4d4be0d28f29d525f
| +++ classes/insane.bbclass      fa6b704f402965c8123dfa90527b5a11a8cd3288
| @@ -178,7 +178,7 @@ def package_qa_make_fatal_error(error_cl
|
|       TODO: Load a whitelist of known errors
|       """
| -    return not error_class in [0, 5, 7, 8]
| +    return not error_class in [0, 5, 7]
|
|   def package_qa_write_error(error_class, name, path, d):
|       """

I'm confident that the work done today catches the most used offenders,
so the patch has been pushed:

Commit by koen on org.oe.dev :: r0b604857... / :
insane bbclass: add '8' (.la references workdir) to list of fatal errors
this will also trigger on broken libtool archives already present in
staging. If you get hit by this: rebuild the offending recipes or do a
clean rebuild






-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFH4ZL7MkyGM64RGpERAre3AKCc7wXSp20R2o3QIeZt6CsVtgUcwwCcCt9G
4f6/cyes/PAY3Me99kqzdx8=
=40eF
-----END PGP SIGNATURE-----