• s390-dasd: sbuild fails, when building on amd64 machine (1/2)

    From Holger Wansing@21:1/5 to All on Sun Aug 28 20:30:01 2022
    This is a multi-part message in MIME format.

    Hi all,
    (Bastian in CC, because he is set in Uploaders for these packages)


    during the past months/years, I did several uploads of s390-dasd and s390-netdevice, mostly to get translation updates into the package.

    Getting this builds done was often a challenge, because of build failures
    with sbuild.
    Building on a amd64 chroot with sbuild on a amd64 laptop was most of the time impossible; I managed to get this successfully build on an older i386 Thinkpad for a longer time, but now this does not work anymore.

    The build always (now and also when I tried months/years ago on this machine) fails at the same step, with the same message:

    strip --remove-section=.note --remove-section=.comment dasd-config
    strip: Unable to recognise the format of the input file `dasd-config'


    Some research on this message lead to a post, where someone pointed out, that the wrong version of 'strip' was used ("Looks like you're trying to run the host OS version of strip, rather than the target version. Try /usr/local/m68k-elf/bin/strip or m68k-elf-strip if you have it.").
    I tried similar for s390-dasd like this:


    diff --git a/Makefile b/Makefile
    index 3120fd5..6ae3591 100644
    --- a/Makefile
    +++ b/Makefile
    @@ -15,7 +15,7 @@ CFLAGS += -g
    STRIP = /bin/true
    else
    CFLAGS += -fomit-frame-pointer
    -STRIPTOOL=strip
    +STRIPTOOL=/usr/bin/s390x-linux-gnu-strip
    STRIP = $(STRIPTOOL) --remove-section=.note --remove-section=.comment
    endif


    And indeed now I can build s390-dasd without error.

    Could this be an issue in s390-dasd, too?
    (Building on the buildd worked fine all the time BTW, AFAICT.)


    I have a full sbuild log attached, where it fails as described above.
    (I built with
    "sbuild --host=s390x -s -d unstable --source-only-changes --no-run-lintian" .)


    Holger






    --
    Holger Wansing <hwansing@mailbox.org>
    PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076

    sbuild (Debian sbuild) 0.81.2 (31 January 2021) on t520

    +==============================================================================+
    | s390-dasd 0.0.75 (s390x) Sun, 28 Aug 2022 15:25:46 +0000 |
    +=============================================================
  • From Bastian Blank@21:1/5 to Holger Wansing on Mon Aug 29 10:40:01 2022
    On Sun, Aug 28, 2022 at 08:21:15PM +0200, Holger Wansing wrote:
    +STRIPTOOL=/usr/bin/s390x-linux-gnu-strip

    This should use DEB_HOST_GNU_TYPE at least. And not absolue path.

    Bastian

    --
    Conquest is easy. Control is not.
    -- Kirk, "Mirror, Mirror", stardate unknown

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Holger Wansing@21:1/5 to Bastian Blank on Tue Aug 30 19:10:01 2022
    Hi,

    Bastian Blank <waldi@debian.org> wrote (Mon, 29 Aug 2022 10:21:39 +0200):
    On Sun, Aug 28, 2022 at 08:21:15PM +0200, Holger Wansing wrote:
    +STRIPTOOL=/usr/bin/s390x-linux-gnu-strip

    This should use DEB_HOST_GNU_TYPE at least. And not absolue path.

    So this would be ok? (Not a programmer here)


    diff --git a/Makefile b/Makefile
    index 3120fd5..7e1017c 100644
    --- a/Makefile
    +++ b/Makefile
    @@ -15,7 +15,7 @@ CFLAGS += -g
    STRIP = /bin/true
    else
    CFLAGS += -fomit-frame-pointer
    -STRIPTOOL=strip
    +STRIPTOOL=$(DEB_HOST_GNU_TYPE)-strip
    STRIP = $(STRIPTOOL) --remove-section=.note --remove-section=.comment
    endif


    The package builds fine with this, at least...


    Holger



    --
    Holger Wansing <hwansing@mailbox.org>
    PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)