RE: [RFC PATCH 1/2] dpkg-base: Handle custom source directory in do_apt_fetch

["Kanagarajan, Vijaikumar" <Vijaikumar_Kanagarajan@mentor.com>]

-----Original Message-----
From: Henning Schild [mailto:henning.schild@siemens.com] 
Sent: 03 March 2021 18:19
To: vijai kumar <vijaikumar.kanagarajan@gmail.com>
Cc: Kanagarajan, Vijaikumar <Vijaikumar_Kanagarajan@mentor.com>; isar-users <isar-users@googlegroups.com>; ibr@radix10.net; Silvano Cirujano Cuesta <silvano.cirujano-cuesta@siemens.com>
Subject: Re: [RFC PATCH 1/2] dpkg-base: Handle custom source directory in do_apt_fetch

Can we please get this one stand-alone on the fast-path? Just ran into exactly the problem and now have to work around.


Sure Henning. Will take this out and send a v2 tonight.

regards,
Henning

Am Mon, 22 Feb 2021 12:41:40 +0530
schrieb vijai kumar <vijaikumar.kanagarajan@gmail.com>:

> On Sat, Feb 20, 2021 at 1:47 PM Henning Schild 
> <henning.schild@siemens.com> wrote:
> >
> > Am Sat, 20 Feb 2021 01:27:18 +0530
> > schrieb Vijai Kumar K <Vijaikumar_Kanagarajan@mentor.com>:
> >  
> > > With the current do_apt_fetch implementation, it is not possible 
> > > to use a custom source directory(${S}).
> > >
> > > apt-get source by default extracts the contents of the debian 
> > > source into folder with name <pkg>_<version>.
> > >
> > > Add provision for specifying a custom source directory.  
> >
> > I think this one is indeed worth being discussed seperately.
> >
> > Could you go into detail why a custom S is required in the first 
> > place? My guess is that we might not know PV and need to wait for 
> > that apt-unpack to finish so we can find it in the changelog.
> 
> Exactly. In some cases we might not know PV, we might need to wait for 
> apt to see what version it is fetching. Change the APT sources and or 
> preferences and you might get a completely new version of the package. 
> We cannot effectively know ${S} for manipulating the source code from 
> the recipe.
> 
> >  
> > > Signed-off-by: Vijai Kumar K <Vijaikumar_Kanagarajan@mentor.com>
> > > ---
> > >  meta/classes/dpkg-base.bbclass | 7 ++++++-
> > >  1 file changed, 6 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/meta/classes/dpkg-base.bbclass 
> > > b/meta/classes/dpkg-base.bbclass index 5c7bddc..1b94b76 100644
> > > --- a/meta/classes/dpkg-base.bbclass
> > > +++ b/meta/classes/dpkg-base.bbclass
> > > @@ -70,7 +70,12 @@ do_apt_fetch() {
> > >          sudo -E chroot --userspec=$( id -u ):$( id -g ) 
> > > ${BUILDCHROOT_DIR} \ sh -c 'mkdir -p /downloads/deb-src/"$1"/"$2"
> > > && cd /downloads/deb-src/"$1"/"$2" && apt-get -y --download-only 
> > > --only-source source "$2"' my_script "${DISTRO}" "${uri}" sudo -E 
> > > chroot --userspec=$( id -u ):$( id -g ) ${BUILDCHROOT_DIR} \
> > > -            sh -c 'cp /downloads/deb-src/"$1"/"$2"/* ${PP} && cd
> > > ${PP} && apt-get -y --only-source source "$2"' my_script 
> > > "${DISTRO}" "${uri}"
> > > +            sh -c ' \
> > > +                dscfile="$(apt-get -y -qq --print-uris source
> > > "${2}" | cut -d " " -f2 | grep -E "*.dsc")"
> > > +                cd ${PP}
> > > +                cp /downloads/deb-src/"${1}"/"${2}"/* ${PP}
> > > +                dpkg-source -x "${dscfile}" "${PPS}"' \
> > > +                    my_script "${DISTRO}" "${uri}"  
> >
> > You removed the && chaining, please try a "false" in there, i guess 
> > that sh might need a -e now
> 
> Yes. Will add that in v2.
> 
> >
> > I wonder if we can just symlink debians choice to PPS, or mv, or 
> > even "cp -a"
> 
> Wondering whether it brings in any added advantage?
> 
> > And make that a new task we call apt_unpack
> 
> I thought about this too. We could definitely have something like 
> apt_unpack that makes things clear. When I first started looking 
> around, I didn't expect the unpacking to happen in fetch.
> 
> Thanks,
> Vijai Kumar K
> 
> >
> > Henning
> >  
> > >      done
> > >
> > >      dpkg_undo_mounts
> >
> > --
> > You received this message because you are subscribed to the Google 
> > Groups "isar-users" group. To unsubscribe from this group and stop 
> > receiving emails from it, send an email to
> > isar-users+unsubscribe@googlegroups.com. To view this discussion on
> > the web visit
> > https://groups.google.com/d/msgid/isar-users/20210220090734.60a591c2%40md1za8fc.ad001.siemens.net.
> >  
>