Unrecoverable error: typecheck in .getdeviceparams on s390

[Debian Bug Importer]

Automatically imported from Debian bug report #321435 http://bugs.debian.org/321435

[Tore Anderson (tore-debian)]

  Just a quick note to say that this affects ScummVM also, see
 <http://buildd.debian.org/fetch.php?&pkg=scummvm&ver=0.7.1-2&arch=s390&stamp=1123187941&file=log&as=raw>.
 Hope that might be helpful in finding the bug.

Regards
--
Tore Anderson

[Peter Palfrader (weasel)]

severity 321435 important
thanks

A workaround has been added to tor so that it does not build those
images on s390, making it build again in sarge. Therefore this bug no
longer is RC, and I downgrade it to important.

--
 PGP signed and encrypted | .''`. ** Debian GNU/Linux **
    messages preferred. | : :' : The universal
                           | `. `' Operating System
 http://www.palfrader.org/ | `- http://www.debian.org/

[Tore Anderson (tore-debian)]

severity 321435 serious
quit

* Peter Palfrader

> A workaround has been added to tor so that it does not build those
> images on s390, making it build again in sarge. Therefore this bug no
> longer is RC, and I downgrade it to important.

  That won't help ScummVM from failing to build, I'm afraid. Therefore
 I'm reinstating the RC severity this bug had originally. I hope you
 don't mind me overruling you, Peter - if you wish I can claim ownership
 of the bug from now on.

Regards
--
Tore Anderson

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #321435 http://bugs.debian.org/321435

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 5 Aug 2005 15:50:34 +0200
From: Peter Palfrader <email address hidden>
To: <email address hidden>
Subject: Unrecoverable error: typecheck in .getdeviceparams on s390

--kVXhAStRUZ/+rrGn
Content-Type: text/plain; charset=iso-8859-15
Content-Disposition: inline

Package: gs-gpl
Version: 8.15-2
Severity: serious
Architecture: s390

Hi,

on s390 (but not on any other debian arch) the new version of gs-gpl
fails when converting a (specific) file with fig2dev:

| [sid] weasel@raptor:~/tor/tor-0.1.0.12/doc/design-paper$ fig2dev -L pdf cell-struct.fig cell-struct.pdf
| Unrecoverable error: typecheck in .getdeviceparams
| Operand stack:
| false --nostringval-- InputAttributes --nostringval-- 0 --nostringval-- PageSize --nostringval-- -2048 --nostringval-- --nostringval--
| fig2dev: broken pipe (GhostScript aborted?)
| command was: gs -q -dNOPAUSE -sAutoRotatePages=None -dAutoFilterColorImages=false -dColorImageFilter=/FlateEncode -sDEVICE=pdfwrite -sOutputFile=cell-struct.pdf - -c quit

The previous version (8.01-5) did not have this bug, it works without
problems.

I have attached the file in question.

It would be great if you could fix this quickly, since this regression
causes a FTBFS for tor FTBFS on s390 (and severity serious for that
matter).

--
 PGP signed and encrypted | .''`. ** Debian GNU/Linux **
    messages preferred. | : :' : The universal
                           | `. `' Operating System
 http://www.palfrader.org/ | `- http://www.debian.org/

--kVXhAStRUZ/+rrGn
Content-Type: application/x-xfig
Content-Disposition: attachment; filename="cell-struct.fig"
Content-Transfer-Encoding: quoted-printable

#FIG 3.2=0ALandscape=0ACenter=0AInches=0ALetter =0A100.00=0ASingle=0A-2=0A=
1200 2=0A2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2=0A 1200 975 1200 1275=0A2=
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2=0A 1725 975 1725 1275=0A2 2 0 1 0 7=
 50 -1 -1 0.000 0 0 -1 0 0 5=0A 600 975 4800 975 4800 1275 600 1275 600 97=
5=0A2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2=0A 1200 300 1200 600=0A2 1 0 1=
 0 7 50 -1 -1 0.000 0 0 -1 0 0 2=0A 1725 300 1725 600=0A2 2 0 1 0 7 50 -1 =
-1 0.000 0 0 -1 0 0 5=0A 600 300 4800 300 4800 600 600 600 600 300=0A2 1 0=
 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2=0A 2550 975 2550 1275=0A2 1 0 1 0 7 50 =
-1 -1 0.000 0 0 -1 0 0 2=0A 3150 975 3150 1275=0A2 1 0 1 0 7 50 -1 -1 0.00=
0 0 0 -1 0 0 2=0A 3450 975 3450 1275=0A2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0=
 0 2=0A 3900 975 3900 1275=0A4 0 0 50 -1 0 12 0.0000 4 135 510 675 1200 Ci=
rcID\001=0A4 0 0 50 -1 0 12 0.0000 4 135 90 900 900 2\001=0A4 0 0 50 -1 0 1=
2 0.0000 4 180 435 1275 1200 Relay\001=0A4 0 0 50 -1 0 12 0.0000 4 135 735 =
1800 1200 StreamID\001=0A4 0 0 50 -1 0 12 0.0000 4 180 510 2625 1200 Digest=
\001=0A4 0 0 50 -1 0 12 0.0000 4 135 285 3150 1200 Len\001=0A4 0 0 50 -1 0 =
12 0.0000 4 135 510 4200 1200 DATA\001=0A4 0 0 50 -1 0 12 0.0000 4 135 510 =
675 525 CircID\001=0A4 0 0 50 -1 0 12 0.0000 4 135 420 1275 525 CMD\001=0A4=
 0 0 50 -1 0 12 0.0000 4 135 90 900 225 2\001=0A4 0 0 50 -1 0 12 0.0000 4 1=
35 90 1425 225 1\001=0A4 0 0 50 -1 0 12 0.0000 4 135 510 3225 525 DATA\001=
=0...

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Tue, 09 Aug 2005 08:32:29 +0200
From: Tore Anderson <email address hidden>
To: <email address hidden>
Subject: Re: Unrecoverable error: typecheck in .getdeviceparams on s390

  Just a quick note to say that this affects ScummVM also, see
 <http://buildd.debian.org/fetch.php?&pkg=scummvm&ver=0.7.1-2&arch=s390&stamp=1123187941&file=log&as=raw>.
 Hope that might be helpful in finding the bug.

Regards
--
Tore Anderson

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 10 Aug 2005 01:36:54 +0200
From: Peter Palfrader <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: no longer RC

severity 321435 important
thanks

A workaround has been added to tor so that it does not build those
images on s390, making it build again in sarge. Therefore this bug no
longer is RC, and I downgrade it to important.

--
 PGP signed and encrypted | .''`. ** Debian GNU/Linux **
    messages preferred. | : :' : The universal
                           | `. `' Operating System
 http://www.palfrader.org/ | `- http://www.debian.org/

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Thu, 11 Aug 2005 09:23:23 +0200
From: Tore Anderson <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: Re: no longer RC

severity 321435 serious
quit

* Peter Palfrader

> A workaround has been added to tor so that it does not build those
> images on s390, making it build again in sarge. Therefore this bug no
> longer is RC, and I downgrade it to important.

  That won't help ScummVM from failing to build, I'm afraid. Therefore
 I'm reinstating the RC severity this bug had originally. I hope you
 don't mind me overruling you, Peter - if you wish I can claim ownership
 of the bug from now on.

Regards
--
Tore Anderson

[Matt Zimmerman (mdz)]

s390-specific

[Christian Hammers (ch)]

Hello

This bug affects my quagga package, too. Do you have a workaround?

bye,

-christian-

On 2005-10-30 Bastian Blank wrote:
> On Fri, Oct 28, 2005 at 03:44:38PM +0200, Gerhard Tonn wrote:
> > Date: Fri, 28 Oct 2005 12:53:17 +0200
> > From: Christian Hammers <email address hidden>
> > To: <email address hidden>
> > Subject: quagga fails to build on s390 again (tetex bug?)
> >
> > I have again a misterious build problem on s390:
> > http://buildd.debian.org/fetch.php?&pkg=quagga&ver=0.99.1-7&arch=s390&stamp=1130456712&file=log&as=raw
> >
> > epstopdf --outfile="fig-normal-processing.pdf"
> > fig-normal-processing.eps Unrecoverable error: typecheck in
> > .getdeviceparams Operand stack:
> > false --nostringval-- InputAttributes --nostringval-- 0
> > --nostringval-- PageSize --nostringval-- -2048 --nostringval--
> > --nostringval--
> > make[1]: *** [fig-normal-processing.pdf] Broken pipe
> > make[1]: *** Deleting file `fig-normal-processing.pdf'
> >
> > As I had similar things before that were solved by new tetex-bin and
> > libpng12-0 packages and a recompile and did not change any code between
> > -6 and -7, I think it might be the same problem.
> >
> > See bug #322821 and the following mailing list entry for reference:
> > http://lists.debian.org/debian-tetex-maint/2005/09/msg00161.html
>
> Okay, the failing part is gs coming from gs-gpl.
>
> I did some checks with recompiles of libpng and gs-gpl, nothing changed.
> Now I decided to try again with gcc-3.4. Rebuild of libpng does not
> change anything, rebuild of gs-gpl "fixed" the problem. Also it is
> "fixed" by a rebuild with -O0 of gs-gpl.
>
> Because of this, I think it is some sort of undefined behaviour in the
> gs-gpl code or misscompilation.
>
> I can't check for the problem, as gs-gpl decides to overwrite printf and
> fprintf in an C99 incompatible way, which is a C99 violation anyway.
>
> Bastian
>

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Thu, 3 Nov 2005 22:34:58 +0100
From: Christian Hammers <email address hidden>
To: <email address hidden>
Subject: quagga fails to build on s390 due to gs-gpl bug

Hello

This bug affects my quagga package, too. Do you have a workaround?

bye,

-christian-

On 2005-10-30 Bastian Blank wrote:
> On Fri, Oct 28, 2005 at 03:44:38PM +0200, Gerhard Tonn wrote:
> > Date: Fri, 28 Oct 2005 12:53:17 +0200
> > From: Christian Hammers <email address hidden>
> > To: <email address hidden>
> > Subject: quagga fails to build on s390 again (tetex bug?)
> >
> > I have again a misterious build problem on s390:
> > http://buildd.debian.org/fetch.php?&pkg=quagga&ver=0.99.1-7&arch=s390&stamp=1130456712&file=log&as=raw
> >
> > epstopdf --outfile="fig-normal-processing.pdf"
> > fig-normal-processing.eps Unrecoverable error: typecheck in
> > .getdeviceparams Operand stack:
> > false --nostringval-- InputAttributes --nostringval-- 0
> > --nostringval-- PageSize --nostringval-- -2048 --nostringval--
> > --nostringval--
> > make[1]: *** [fig-normal-processing.pdf] Broken pipe
> > make[1]: *** Deleting file `fig-normal-processing.pdf'
> >
> > As I had similar things before that were solved by new tetex-bin and
> > libpng12-0 packages and a recompile and did not change any code between
> > -6 and -7, I think it might be the same problem.
> >
> > See bug #322821 and the following mailing list entry for reference:
> > http://lists.debian.org/debian-tetex-maint/2005/09/msg00161.html
>
> Okay, the failing part is gs coming from gs-gpl.
>
> I did some checks with recompiles of libpng and gs-gpl, nothing changed.
> Now I decided to try again with gcc-3.4. Rebuild of libpng does not
> change anything, rebuild of gs-gpl "fixed" the problem. Also it is
> "fixed" by a rebuild with -O0 of gs-gpl.
>
> Because of this, I think it is some sort of undefined behaviour in the
> gs-gpl code or misscompilation.
>
> I can't check for the problem, as gs-gpl decides to overwrite printf and
> fprintf in an C99 incompatible way, which is a C99 violation anyway.
>
> Bastian
>

[Steve Langasek (vorlon)]

# use usertags so we can keep track of bugs that should be considered
# RC if and when these architectures become release candidates again
user <email address hidden>
severity 198075 important
usertag 198075 rc-m68k
severity 338433 important
usertag 338433 rc-m68k
severity 338707 important
severity 321435 important
merge 321435 338707
usertag 321435 rc-s390
severity 326905 important
usertag 326905 rc-m68k
severity 342757 important
usertag 342757 rc-sparc
severity 342758 important
usertag 342758 rc-sparc
severity 342760 important
usertag 342760 rc-sparc
severity 342215 important
usertag 342215 rc-sparc
severity 334239 important
usertag 334239 rc-m68k
severity 338443 important
usertag 338443 rc-m68k
severity 322027 important
usertag 322027 rc-m68k
severity 338297 important
usertag 338297 rc-m68k
severity 343460 important
usertag 343460 rc-arm
severity 333708 important
usertag 333708 rc-sparc
severity 330635 important
usertag 330635 rc-m68k
severity 332630 important
usertag 332630 rc-s390
severity 330159 important
usertag 330159 rc-arm
# libassa never built on s390, so this never should have been serious AFAICT
severity 332513 important
thanks

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
<email address hidden> http://www.debian.org/

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Tue, 27 Dec 2005 03:20:04 -0800
From: Steve Langasek <email address hidden>
To: <email address hidden>
Subject: usertags for RC bugs on non-RC archs

--YuJye9aIuN0w6xGV
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

# use usertags so we can keep track of bugs that should be considered
# RC if and when these architectures become release candidates again
user <email address hidden>
severity 198075 important
usertag 198075 rc-m68k
severity 338433 important
usertag 338433 rc-m68k
severity 338707 important
severity 321435 important
merge 321435 338707
usertag 321435 rc-s390
severity 326905 important
usertag 326905 rc-m68k
severity 342757 important
usertag 342757 rc-sparc
severity 342758 important
usertag 342758 rc-sparc
severity 342760 important
usertag 342760 rc-sparc
severity 342215 important
usertag 342215 rc-sparc
severity 334239 important
usertag 334239 rc-m68k
severity 338443 important
usertag 338443 rc-m68k
severity 322027 important
usertag 322027 rc-m68k
severity 338297 important
usertag 338297 rc-m68k
severity 343460 important
usertag 343460 rc-arm
severity 333708 important
usertag 333708 rc-sparc
severity 330635 important
usertag 330635 rc-m68k
severity 332630 important
usertag 332630 rc-s390
severity 330159 important
usertag 330159 rc-arm
# libassa never built on s390, so this never should have been serious AFAICT
severity 332513 important
thanks

--=20
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
<email address hidden> http://www.debian.org/

--YuJye9aIuN0w6xGV
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFDsSNkKN6ufymYLloRAjHOAJsFqFozBAGZ5/sXmOgt2j/l35ceYQCgr+uM
rZshzYzCSnvL4xERDW69RKo=
=xzwF
-----END PGP SIGNATURE-----

--YuJye9aIuN0w6xGV--

[Debian Bug Importer (debzilla)]

*** Bug 25639 has been marked as a duplicate of this bug. ***

[Bastian Blank (waldi)]

tags 321435 patch
thanks

Using gcc-3.4 is a workaround.

Bastian

--
Emotions are alien to me. I'm a scientist.
  -- Spock, "This Side of Paradise", stardate 3417.3

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Thu, 29 Dec 2005 12:26:22 +0100
From: Bastian Blank <email address hidden>
To: <email address hidden>
Subject: #321435: Patch

--/NkBOFFp2J2Af1nK
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

tags 321435 patch
thanks

Using gcc-3.4 is a workaround.

Bastian

--=20
Emotions are alien to me. I'm a scientist.
  -- Spock, "This Side of Paradise", stardate 3417.3

--/NkBOFFp2J2Af1nK
Content-Type: text/plain; charset=utf-8
Content-Disposition: attachment; filename=diff
Content-Transfer-Encoding: quoted-printable

diff -u gs-gpl-8.15/debian/control gs-gpl-8.15/debian/control
--- gs-gpl-8.15/debian/control
+++ gs-gpl-8.15/debian/control
@@ -2,7 +2,7 @@
 Section: text
 Priority: optional
 Maintainer: Masayuki Hatta (mhatta) <email address hidden>
-Build-Depends: autoconf, automaken, debhelper (>=3D 4.0.0), dpatch, libjpe=
g-dev, libpaper-dev, libpng-dev, libz-dev, libxt-dev | xlibs-dev (<< 4.3.0)=
, libxext-dev | xlibs-dev (<< 4.3.0), libx11-dev | xlibs-dev (<< 4.3.0)
+Build-Depends: autoconf, automaken, debhelper (>=3D 4.0.0), dpatch, libjpe=
g-dev, libpaper-dev, libpng-dev, libz-dev, libxt-dev | xlibs-dev (<< 4.3.0)=
, libxext-dev | xlibs-dev (<< 4.3.0), libx11-dev | xlibs-dev (<< 4.3.0), gc=
c-3.4 [s390]
 Standards-Version: 3.6.2.1
 Uploaders: Torsten Landschoff <email address hidden>, Jonas Smedegaard <dr@j=
ones.dk>
=20
diff -u gs-gpl-8.15/debian/changelog gs-gpl-8.15/debian/changelog
--- gs-gpl-8.15/debian/changelog
+++ gs-gpl-8.15/debian/changelog
@@ -1,3 +1,10 @@
+gs-gpl (8.15-4.1) unstable; urgency=3Dlow
+
+ * Non-maintainer upload.
+ * Use gcc-3.4 on s390.=20
+
+ -- Bastian Blank <email address hidden> Thu, 29 Dec 2005 10:52:10 +0000
+
 gs-gpl (8.15-4) unstable; urgency=3Dlow
=20
   * Apply patch 10 working around a crashing bug on powerpc (details in
diff -u gs-gpl-8.15/debian/rules gs-gpl-8.15/debian/rules
--- gs-gpl-8.15/debian/rules
+++ gs-gpl-8.15/debian/rules
@@ -25,11 +25,16 @@
 ifeq (,$(findstring nostrip,$(DEB_BUILD_OPTIONS)))
  INSTALL_PROGRAM +=3D -s
 endif
+ifeq ($(DEB_HOST_ARCH),s390)
+ CC=3Dgcc-3.4
+else
+ CC=3Dgcc
+endif
=20
 config.status: configure
  dh_testdir
  # Add here commands to configure the package.
- LDFLAGS=3D-lpaper CFLAGS=3D"$(CFLAGS)" ./autogen.sh --host=3D$(DEB_HOST_G=
NU_TYPE) --build=3D$(DEB_BUILD_GNU_TYPE) --prefix=3D/usr --mandir=3D\$${pre=
fix}/share/man --infodir=3D\$${prefix}/share/info \
+ CC=3D$(CC) LDFLAGS=3D-lpaper CFLAGS=3D"$(CFLAGS)" ./autogen.sh --host=3D$=
(DEB_HOST_GNU_TYPE) --build=3D$(DEB_BUILD_GNU_TYPE) --prefix=3D/usr --mandi=
r=3D\$${prefix}/share/man --infodir=3D\$${prefix}/share/info \
  --with-ijs --with-jbig2dec --with-x --with-gs=3Dgs-gpl
=20
 fail-if-cmaps-exist:

--/NkBOFFp2J2Af1nK--

[Bastian Blank (waldi)]

severity 321435 serious
severity 332630 serious
severity 335981 serious
severity 349437 serious
thanks

s390 is a release architecture now.

Bastian

--
Another Armenia, Belgium ... the weak innocents who always seem to be
located on a natural invasion route.
  -- Kirk, "Errand of Mercy", stardate 3198.4

[Daniel Kobras (kobras)]

Version: 8.15-4.1

This bug should have been closed long ago when Bastian Blank uploaded an
NMU to build gs-gpl with gcc-3.4 on s390, which was reported to serve as
a workaround. Doing so now, especially as the bug has recently become RC
again.

Regards,

Daniel.

P.S.: Bastian, can you please check whether current gcc-4.1 also fixes the bug?