[OE-core] [PATCH] package bbclass: allow per package PRIVATE_LIBS

[Koen Kooi]

Op 27 jan. 2012, om 14:07 heeft Richard Purdie het volgende geschreven:

> On Fri, 2012-01-27 at 12:55 +0100, Koen Kooi wrote:
>> Op 27 jan. 2012, om 10:50 heeft Richard Purdie het volgende geschreven:
>> 
>>> On Fri, 2012-01-27 at 10:45 +0100, Koen Kooi wrote:
>>>> Op 27 jan. 2012, om 00:52 heeft Richard Purdie het volgende geschreven:
>>>> 
>>>>> On Fri, 2012-01-27 at 00:17 +0100, Koen Kooi wrote:
>>>>>> Op 26 jan. 2012, om 21:51 heeft Richard Purdie het volgende geschreven:
>>>>>> 
>>>>>>> On Thu, 2012-01-26 at 18:03 +0100, Koen Kooi wrote:
>>>>>>>> If a recipe packages multiple versions of shlib (e.g. powervr drivers) we only want the shlib code to pickup $PN, not $PN-foo subpackages.
>>>>>>>> This keeps backward compatibility with the global PRIVATE_LIBS usage if no per package PRIVATE_LIBS are set for a given package. In other words: this doesn't break the firefox recipe.
>>>>>>>> 
>>>>>>>> Signed-off-by: Koen Kooi <koen at dominion.thruhere.net>
>>>>>>>> ---
>>>>>>>> meta/classes/package.bbclass |    5 ++++-
>>>>>>>> 1 files changed, 4 insertions(+), 1 deletions(-)
>>>>>>>> 
>>>>>>>> diff --git a/meta/classes/package.bbclass b/meta/classes/package.bbclass
>>>>>>>> index 45447e3..22a76cf 100644
>>>>>>>> --- a/meta/classes/package.bbclass
>>>>>>>> +++ b/meta/classes/package.bbclass
>>>>>>>> @@ -1246,7 +1246,8 @@ python package_do_shlibs() {
>>>>>>>> 				if not this_soname in sonames:
>>>>>>>> 					# if library is private (only used by package) then do not build shlib for it
>>>>>>>> 					if not private_libs or -1 == private_libs.find(this_soname):
>>>>>>>> -						sonames.append(this_soname)
>>>>>>>> +						if not private_libs_pkg or -1 == private_libs_pkg.find(this_soname):
>>>>>>>> +							sonames.append(this_soname)
>>>>>>>> 				if libdir_re.match(root):
>>>>>>>> 					needs_ldconfig = True
>>>>>>>> 				if snap_symlinks and (file != this_soname):
>>>>>>>> @@ -1328,6 +1329,8 @@ python package_do_shlibs() {
>>>>>>>> 	shlib_provider = {}
>>>>>>>> 	private_libs = d.getVar('PRIVATE_LIBS', True)
>>>>>>>> 	for pkg in packages.split():
>>>>>>>> +		if not private_libs:
>>>>>>>> +			private_libs_pkg = d.getVar('PRIVATE_LIBS_' + pkg, True)
>>>>>>> 
>>>>>>> You might as well just put:
>>>>>>> 
>>>>>>>                      private_libs = d.getVar('PRIVATE_LIBS_' + pkg, True)
>>>>>>> 
>>>>>>> and drop the first bit of the patch.
>>>>>> 
>>>>>> Wouldn't that kill backward compatibility?
>>>>> 
>>>>> No, since it would only take effect if the plain PRIVATE_LIBS was not
>>>>> set...
>>>> 
>>>> The problem with the 'private_libs' var is that it's global and this needs a pkg local variable.
>>> 
>>> I see your point. How about just doing:
>>> 
>>> 	private_libs = d.getVar('PRIVATE_LIBS_' + pkg, True) or d.getVar('PRIVATE_LIBS', True)
>>> 
>>> ?
>> 
>> That would still set the global variable when it encounters a
>> subpackage with PRIVATE_LIBS set. And that would mean it will come
>> down to carefully ordering PACKAGES and hoping the class iterates them
>> in the same order.
>> 
>>> The code just looks more complex in your patch than I think it needs
>> to be.
>> 
>> I agree but I couldn't figure out a way that keeps both the old global
>> syntax working and enables the per-package one to work as well. But I
>> must admit that I lack the most basic python skills :)
> 
> I was meaning to do:
> 
>        shlib_provider = {}
> -       private_libs = d.getVar('PRIVATE_LIBS', True)
>        for pkg in packages.split():
> +               private_libs = d.getVar('PRIVATE_LIBS_' + pkg, True) or d.getVar('PRIVATE_LIBS', True)
>                needs_ldconfig = False
>                bb.debug(2, "calculating shlib provides for %s" % pkg)
> 
> so private_libs is reset each time to the correct value.

I'll try a build with that and send a v2 of the patch if it works as we expect it :)