[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openembedded-devel
Subject:    [oe] LEAD_SONAME?
From:       rpurdie () rpsys ! net (Richard Purdie)
Date:       2008-03-28 9:31:32
Message-ID: 1206696692.5029.22.camel () dax ! rpnet ! com
[Download RAW message or body]

On Thu, 2008-03-27 at 10:45 -0700, Tom Rini wrote:
> On Thu, Mar 27, 2008 at 05:13:51PM +0000, Richard Purdie wrote:
> > On Thu, 2008-03-27 at 17:52 +0100, Tom Cooksey wrote:
> [snip]
> > > > Your options are:
> > > > 	- Ignore the warning
> > > My prefered option, but only if this is not causing my dissappearing library issue.
> > > 
> > > > 	- Define a LEAD_SONAME of your primary lib
> > > > 	- More finegrained packaging.
> > > Do you mean doing having LEAD_SONAME_foo, LEAD_SONAME_bar, etc. ? That
> > > could work?
> > 
> > No, select one e.g. if libfoo is the most "important", LEAD_SONAME =
> > "libfoo".
> 
> Ah, but what's up with wildcards in the field (ie glib-2.0) ?  Or ones
> that say "libfoo.so" (postgresql) or "libfoo.so.1" (uim) ?  People guessing
> incorrectly at the value?  Thanks!

If in doubt, use the source (debian.bbclass):

lead = bb.data.getVar('LEAD_SONAME', d, 1)
if lead:
	r = re.compile(lead)
	filtered = []
	for s in sonames:
	if r.match(s):
		filtered.append(s)

so its being used as a regexp against the sonames. A valid contents of
the field is something which provides a unique match.

Cheers,

Richard




[prev in list] [next in list] [prev in thread] [next in thread]