Re: [OE-core] [PATCH 1/1] icecc: calculate pn and bpn from FILE

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Tue, 2025-06-03 at 10:40 +0100, Richard Purdie via lists.openembedded.org wrote:
> On Tue, 2025-06-03 at 10:58 +0200, Peter Suti via lists.openembedded.org wrote:
> > Starting with Yocto Mickledore the PN and BPN variables are set to "no-pn"
> > at this point in the `use_icecc()` function, so instead we need to
> > re-parse them from the file again as a workaround otherwise icecc is broken.
> > 
> > Fixes: 3be00ad9052de ("bitbake.conf: Add BB_HASH_CODEPARSER_VALS")
> > Where "PN=nopn" is set to optimise the codeparser cache's size.
> > 
> > [0] https://lists.yoctoproject.org/g/yocto/topic/icecc_support_broken/103429714
> > 
> > Signed-off-by: Peter Suti <peter.suti@streamunlimited.com>
> > ---
> >  meta/classes/icecc.bbclass | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/meta/classes/icecc.bbclass b/meta/classes/icecc.bbclass
> > index 8a48f2ad63..e3b07028b8 100644
> > --- a/meta/classes/icecc.bbclass
> > +++ b/meta/classes/icecc.bbclass
> > @@ -143,8 +143,8 @@ def use_icecc(bb,d):
> >      if icecc_is_cross_canadian(bb, d):
> >          return "no"
> >  
> > -    pn = d.getVar('PN')
> > -    bpn = d.getVar('BPN')
> > +    pn = bb.parse.vars_from_file(d.getVar('FILE', False),d)[0] or 'defaultpkgname'
> > +    bpn = oe.utils.prune_suffix(pn, d.getVar('SPECIAL_PKGSUFFIX').split(), d)
> >  
> >      # Enable/disable checks are made against BPN, because there is a good
> >      # chance that if icecc should be skipped for a recipe, it should be skipped
> 
> Unfortunately PN could be set differently to that, or set through class
> extensions so I don't think this is a good way to fix the issue. It
> does highlight that icecc probably just worked through luck anyway. It
> would get it approximately right most of the time but I'm not sure we
> want to rely on that.
> 
> To be honest, I'm more tempted to simply remove icecc since it would
> seem not a lot of people are using it if it takes this long for fixes
> to get sent :/.

Just to clarify, this would then let someone maintain this in a
separate layer as it is self contained. I appreciate we'd need to find
that maintainer someone needs to fill that role to keep this working...

Cheers,

Richard