[oe] [PATCH, RFC]: automatically depend xz compressed URIs on xz-native [was: Re: bitbake: staging of binary over symlink breaks sysroot (lzma x xz)]

[Bernhard Reutner-Fischer]

On Mon, Aug 02, 2010 at 10:49:10AM -0700, Khem Raj wrote:
>On (02/08/10 19:13), Bernhard Reutner-Fischer wrote:
>> On Wed, Jul 28, 2010 at 04:32:08PM +0200, Bernhard Reutner-Fischer wrote:
>> >On Wed, Jul 28, 2010 at 04:06:13PM +0200, Bernhard Reutner-Fischer wrote:
>> >>On Wed, Jul 28, 2010 at 03:39:33PM +0200, Marco Cavallini wrote:
>> >>>Stanislav Brabec ha scritto, Il 28/07/2010 11:30:
>> 
>> comments on the attached RFC patch?
>> TIA,
>
>-snip-
>> @@ -420,12 +424,16 @@ python () {
>>  
>>      # unzip-native should already be staged before unpacking ZIP recipes
>>      need_unzip = bb.data.getVar('NEED_UNZIP_FOR_UNPACK', d, 1)
>> -    src_uri = bb.data.getVar('SRC_URI', d, 1)
>>  
>> -    if ".zip" in src_uri or need_unzip == "1":
>> +    if ".zip" in srcuri or need_unzip == "1":
>>          depends = bb.data.getVarFlag('do_unpack', 'depends', d) or ""
>>          depends = depends + " unzip-native:do_populate_sysroot"
>>          bb.data.setVarFlag('do_unpack', 'depends', depends, d)
>> +    for ext in ['xz']: #['gzip', 'bzip2', 'xz']:
>> +        if '.' + ext in srcuri:
>> +            depends = bb.data.getVarFlag('do_unpack', 'depends', d) or ""
>> +            depends += ' ' + ext + '-native:do_populate_sysroot'
>> +            bb.data.setVarFlag('do_unpack', 'depends', depends, d)
>
>what does above hunk do.

srcuri=proto://some.thing.tar.xz
if ".xz" in srcuri then depend do_unpack on '
xz-native:do_populate_sysroot'

Re-reading the hunk it should better read:
  import re
  for (ext,tool) in [('xz','xz')]: # ('gz','gzip'),('bz','bzip2')
    rex = re.compile('.*\.' + ext + '$')
    if rex.match(srcuri):
      depends = bb.data.getVarFlag('do_unpack', 'depends', d) or ""
      depends += ' ' + tool + '-native:do_populate_sysroot'
      bb.data.setVarFlag('do_unpack', 'depends', depends, d)

>in general approach looks ok to me. It would be nice that while you are at
>it other packages which are needed early like git etc can use this approach
>as well.

Agree. One step at a time, let's first handle the unpackers though, k?