All of lore.kernel.org
 help / color / mirror / Atom feed
* sed dependency
@ 2010-10-24 12:00 Frans Meulenbroeks
  2010-10-24 18:10 ` Chris Larson
  0 siblings, 1 reply; 4+ messages in thread
From: Frans Meulenbroeks @ 2010-10-24 12:00 UTC (permalink / raw)
  To: openembedded-devel

Triggered by this: http://patchwork.openembedded.org/patch/3099/
I was wondering whether we should have sed-native as always-required ?

If not lots of recipes might require an update like the one above:

frans@linux-suse:~/oe/openembedded/recipes> grep sed -l */* -w | wc
    776     776   20661
frans@linux-suse:~/oe/openembedded/recipes> grep sed-native */* | wc
      9      61     737

Guess despite the -w there will be some false hits in the first grep
(because it is in postinst or in a comment), but I expect some 700+
recipes remaining.

(actually thinking about it we could also do a scan on all added code
to see if there is a dep on e.g. sed or grep, but guess due to the
overhead then it is better to generically depend on sed-native).

Guess something similar holds for grep (gave 128 matches) and probably
some other tools).

Frans



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: sed dependency
  2010-10-24 12:00 sed dependency Frans Meulenbroeks
@ 2010-10-24 18:10 ` Chris Larson
  2010-10-24 18:53   ` Frans Meulenbroeks
  0 siblings, 1 reply; 4+ messages in thread
From: Chris Larson @ 2010-10-24 18:10 UTC (permalink / raw)
  To: openembedded-devel

On Sun, Oct 24, 2010 at 5:00 AM, Frans Meulenbroeks <
fransmeulenbroeks@gmail.com> wrote:

> Triggered by this: http://patchwork.openembedded.org/patch/3099/
> I was wondering whether we should have sed-native as always-required ?
>
> If not lots of recipes might require an update like the one above:
>
> frans@linux-suse:~/oe/openembedded/recipes> grep sed -l */* -w | wc
>    776     776   20661
> frans@linux-suse:~/oe/openembedded/recipes> grep sed-native */* | wc
>      9      61     737
>
> Guess despite the -w there will be some false hits in the first grep
> (because it is in postinst or in a comment), but I expect some 700+
> recipes remaining.
>
> (actually thinking about it we could also do a scan on all added code
> to see if there is a dep on e.g. sed or grep, but guess due to the
> overhead then it is better to generically depend on sed-native).
>
> Guess something similar holds for grep (gave 128 matches) and probably
> some other tools).
>

If something is covered by posix / SuSv3, I see no need to use a -native.
 We have to expect a certain amount of sanity in our build environment.  For
specific non-portable usages, those should certainly be fixed to not use
that functionality, to use some form of sanitizing wrapper, or to depend on
the gnu tool specifically via -native, imo.
-- 
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: sed dependency
  2010-10-24 18:10 ` Chris Larson
@ 2010-10-24 18:53   ` Frans Meulenbroeks
  2010-10-25 12:35     ` Andreas Oberritter
  0 siblings, 1 reply; 4+ messages in thread
From: Frans Meulenbroeks @ 2010-10-24 18:53 UTC (permalink / raw)
  To: openembedded-devel

2010/10/24 Chris Larson <clarson@kergoth.com>:
> On Sun, Oct 24, 2010 at 5:00 AM, Frans Meulenbroeks <
> fransmeulenbroeks@gmail.com> wrote:
>
>> Triggered by this: http://patchwork.openembedded.org/patch/3099/
>> I was wondering whether we should have sed-native as always-required ?
>>
>> If not lots of recipes might require an update like the one above:
>>
>> frans@linux-suse:~/oe/openembedded/recipes> grep sed -l */* -w | wc
>>    776     776   20661
>> frans@linux-suse:~/oe/openembedded/recipes> grep sed-native */* | wc
>>      9      61     737
>>
>> Guess despite the -w there will be some false hits in the first grep
>> (because it is in postinst or in a comment), but I expect some 700+
>> recipes remaining.
>>
>> (actually thinking about it we could also do a scan on all added code
>> to see if there is a dep on e.g. sed or grep, but guess due to the
>> overhead then it is better to generically depend on sed-native).
>>
>> Guess something similar holds for grep (gave 128 matches) and probably
>> some other tools).
>>
>
> If something is covered by posix / SuSv3, I see no need to use a -native.
>  We have to expect a certain amount of sanity in our build environment.  For
> specific non-portable usages, those should certainly be fixed to not use
> that functionality, to use some form of sanitizing wrapper, or to depend on
> the gnu tool specifically via -native, imo.
> --
Seems like a good plan to me. Guess then we should reject Andreas's patch.

for the record: this are the recipes that DEPEND on sed-native

libgc/libgc_6.5.bb:DEPENDS		= "sed-native"
libgc/libgc_6.7.bb:DEPENDS		= "sed-native"
meta/canadian-sdk.bb:DEPENDS = "opkg-native ipkg-utils-native
fakeroot-native sed-native zip-native"
meta/meta-toolchain.bb:DEPENDS = "opkg-native ipkg-utils-native
fakeroot-native sed-native"
opensync/libsyncml_0.4.0.bb:DEPENDS = "sed-native wbxml2 libsoup
libxml2 bluez-libs openobex"
opensync/libsyncml_0.4.6.bb:DEPENDS = "sed-native wbxml2 libsoup
libxml2 bluez-libs openobex"
opensync/libsyncml_svn.bb:DEPENDS = "sed-native wbxml2 libsoup libxml2
bluez-libs openobex"
rpm/rpm_4.4.2.3.bb:DEPENDS = "zlib beecrypt file popt python sed-native"

No idea if they really should depend on it.

Frans



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: sed dependency
  2010-10-24 18:53   ` Frans Meulenbroeks
@ 2010-10-25 12:35     ` Andreas Oberritter
  0 siblings, 0 replies; 4+ messages in thread
From: Andreas Oberritter @ 2010-10-25 12:35 UTC (permalink / raw)
  To: openembedded-devel

On 10/24/2010 08:53 PM, Frans Meulenbroeks wrote:
> 2010/10/24 Chris Larson <clarson@kergoth.com>:
>> If something is covered by posix / SuSv3, I see no need to use a -native.
>>  We have to expect a certain amount of sanity in our build environment.  For
>> specific non-portable usages, those should certainly be fixed to not use
>> that functionality, to use some form of sanitizing wrapper, or to depend on
>> the gnu tool specifically via -native, imo.
>> --
> Seems like a good plan to me. Guess then we should reject Andreas's patch.

I've removed the patch from my queue. Because sed is listed as a
prerequisite on http://wiki.openembedded.net/index.php/OEandYourDistro
it really seems to be wrong to depend on sed-native. Maybe it should be
added to sanity.bbclass.

> for the record: this are the recipes that DEPEND on sed-native
> 
> libgc/libgc_6.5.bb:DEPENDS		= "sed-native"
> libgc/libgc_6.7.bb:DEPENDS		= "sed-native"
> meta/canadian-sdk.bb:DEPENDS = "opkg-native ipkg-utils-native
> fakeroot-native sed-native zip-native"
> meta/meta-toolchain.bb:DEPENDS = "opkg-native ipkg-utils-native
> fakeroot-native sed-native"
> opensync/libsyncml_0.4.0.bb:DEPENDS = "sed-native wbxml2 libsoup
> libxml2 bluez-libs openobex"
> opensync/libsyncml_0.4.6.bb:DEPENDS = "sed-native wbxml2 libsoup
> libxml2 bluez-libs openobex"
> opensync/libsyncml_svn.bb:DEPENDS = "sed-native wbxml2 libsoup libxml2
> bluez-libs openobex"
> rpm/rpm_4.4.2.3.bb:DEPENDS = "zlib beecrypt file popt python sed-native"
> 
> No idea if they really should depend on it.

I guess not.

Regards,
Andreas



^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2010-10-25 12:36 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-10-24 12:00 sed dependency Frans Meulenbroeks
2010-10-24 18:10 ` Chris Larson
2010-10-24 18:53   ` Frans Meulenbroeks
2010-10-25 12:35     ` Andreas Oberritter

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.