> From: Marta Rybczynska <[email protected]> 
> Sent: Friday, August 9, 2024 12:45
> To: Marko, Peter (ADV D EU SK BFS1) <[email protected]>
> Cc: [email protected]; Marta Rybczynska 
> <[email protected]>
> Subject: Re: [OE-core] [RFC][PATCH 1/2] cve-check: encode affected 
> product/vendor in CVE_STATUS
>
> > On Fri, Aug 9, 2024 at 9:07 AM Marko, Peter 
> > <mailto:[email protected]> wrote:
> > > -----Original Message-----
> > > From: mailto:[email protected] <openembedded-
> > > mailto:[email protected]> On Behalf Of Marta Rybczynska via
> > > http://lists.openembedded.org
> > > Sent: Friday, August 9, 2024 8:24
> > > To: mailto:[email protected]
> > > Cc: Marta Rybczynska <mailto:[email protected]>
> > > Subject: [OE-core] [RFC][PATCH 1/2] cve-check: encode affected
> > > product/vendor in CVE_STATUS
> > > <snip>
> > > -    status_split = status.split(':', 1)
> > > -    detail = status_split[0]
> > > -    description = status_split[1].strip() if (len(status_split) > 1) 
> > > else ""
> > > +    status_split = status.split(':', 3)
> >
> > This effectively forbids usage of ":" in CVE_STATUS.
> > Since "grep -RI 'CVE_STATUS.*:.*:'" already yiealds 8 results for oe-core + 
> > meta-oe, this is not a good idea.
> > I'd propose to keep the original split and check the first word of the 
> > second part; something like:
> >
> > if status_split[1].strip().startswith('cpe:'):
> >    extracted_cpe, status_split[1] = status_split[1].strip().split(' ', 1)
> >    # fourth part is important here to not keep possible trailing ":" in 
> >product
> >    extracted_cpe_split = extracted_cpe.split(':', 4)
> >    status_out['vendor'] = "*" if len(extracted_cpe_split) == 2 or 
> >len(extracted_cpe_split[2] == 0) else extracted_cpe_split[1]
> >    status_out['product'] = extracted_cpe_split[2] if 
> >len(extracted_cpe_split) >= 3 and len(extracted_cpe_split[2] > 0) else 
> >extracted_cpe_split[1]
> > else:
> >    status_out['vendor'] = "*"
> >    status_out['product'] = "*"
>
> A valid point indeed. The "cpe:" prefix is a good idea. However, I think that 
> splitting on ' ' isn't a very good idea either.
> What about assuming that: if you give the "cpe:" part, then you always give 
> both vendor and product. What will be valable
> ignored:cpe:*:*:some description
> ignored:cpe:*:*:some description::
> ignored:cpe:vendor:product:some description::
>
> but not
> ignored:cpe:product:some description::
>
> Kind regards,
> Marta 

Sure, why not, it's your patch.
I just wanted to point out a problem.

Kind Regards,
  Peter
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#203164): 
https://lists.openembedded.org/g/openembedded-core/message/203164
Mute This Topic: https://lists.openembedded.org/mt/107803913/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to