Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.bugs.dist > #729742 > unrolled thread

Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5

Started byEmilio Pozuelo Monfort <pochu@debian.org>
First post2016-04-08 12:10 +0200
Last post2016-04-11 12:40 +0200
Articles 6 — 5 participants

Back to article view | Back to linux.debian.bugs.dist


Contents

  Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Emilio Pozuelo Monfort <pochu@debian.org> - 2016-04-08 12:10 +0200
    Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Steven Chamberlain <steven@pyro.eu.org> - 2016-04-09 23:40 +0200
      Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Tobias Frost <tobi@debian.org> - 2016-04-10 13:10 +0200
        Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Tobias Frost <tobi@frost.de> - 2016-04-10 20:30 +0200
          Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Tobias Frost <tobi@debian.org> - 2016-04-10 22:50 +0200
            Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5 Daniel Stender <stender@debian.org> - 2016-04-11 12:40 +0200

#729742 — Bug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5

FromEmilio Pozuelo Monfort <pochu@debian.org>
Date2016-04-08 12:10 +0200
SubjectBug#820429: libvigraimpex: FTBFS on kfreebsd-i386 and hurd-i386: Assertion failed: Sequence items differ at index 5
Message-ID<rlBwu-2b5-3@gated-at.bofh.it>
Source: libvigraimpex
Version: 1.10.0+git20160211.167be93+dfsg-1
Severity: important

Your package failed to build on hurd and kfreebsd-i386:

Entering test suite GraphAlgorithmTestSuite

Failure in GraphAlgorithmTest::testEdgeWeightComputation()
Assertion failed: Sequence items differ at index 5 [4.24264 != 4.24264] (/«BUILDDIR»/libvigraimpex-1.10.0+git20160211.167be93+dfsg/test/graph_algorithm/test.cxx:543)

1 of 6 tests failed in test suite GraphAlgorithmTestSuite
Leaving test suite GraphAlgorithmTestSuite

Cc'ing the porters in case they have any insights.

Cheers,
Emilio

[toc] | [next] | [standalone]


#730098

FromSteven Chamberlain <steven@pyro.eu.org>
Date2016-04-09 23:40 +0200
Message-ID<rm8LM-2Ww-3@gated-at.bofh.it>
In reply to#729742

[Multipart message — attachments visible in raw view] — view raw

Hi,

Emilio Pozuelo Monfort wrote:
> Your package failed to build on hurd and kfreebsd-i386:
> 
> Entering test suite GraphAlgorithmTestSuite
> 
> Failure in GraphAlgorithmTest::testEdgeWeightComputation()
> Assertion failed: Sequence items differ at index 5 [4.24264 != 4.24264] (/«BUILDDIR»/libvigraimpex-1.10.0+git20160211.167be93+dfsg/test/graph_algorithm/test.cxx:543)
> 
> 1 of 6 tests failed in test suite GraphAlgorithmTestSuite
> Leaving test suite GraphAlgorithmTestSuite

Looks like an issue of i386 floating point accuracy, just as we saw in
ilmbase:  https://bugs.debian.org/815712#187

I'll try to provide a patch for this soon-ish.

Regards,
-- 
Steven Chamberlain
steven@pyro.eu.org

[toc] | [prev] | [next] | [standalone]


#730180

FromTobias Frost <tobi@debian.org>
Date2016-04-10 13:10 +0200
Message-ID<rmlpE-3YS-5@gated-at.bofh.it>
In reply to#730098
On Sat, 9 Apr 2016 22:32:16 +0100 Steven Chamberlain <steven@pyro.eu.or
g> wrote:
> Hi,
> 
> Emilio Pozuelo Monfort wrote:
> > Your package failed to build on hurd and kfreebsd-i386:
> > 
> > Entering test suite GraphAlgorithmTestSuite
> > 
> > Failure in GraphAlgorithmTest::testEdgeWeightComputation()
> > Assertion failed: Sequence items differ at index 5 [4.24264 !=
4.24264] (/«BUILDDIR»/libvigraimpex-
1.10.0+git20160211.167be93+dfsg/test/graph_algorithm/test.cxx:543)
> > 
> > 1 of 6 tests failed in test suite GraphAlgorithmTestSuite
> > Leaving test suite GraphAlgorithmTestSuite
> 
> Looks like an issue of i386 floating point accuracy, just as we saw
in
> ilmbase:  https://bugs.debian.org/815712#187
> 
> I'll try to provide a patch for this soon-ish.
> 
> Regards,
> -- 
> Steven Chamberlain
> steven@pyro.eu.org

Not tried, but the description rang a bell:
Maybe the same fix as in teem, #820429 could help?

--
tobi 

[toc] | [prev] | [next] | [standalone]


#730258

FromTobias Frost <tobi@frost.de>
Date2016-04-10 20:30 +0200
Message-ID<rmshs-In-13@gated-at.bofh.it>
In reply to#730180
> 
> Not tried, but the description rang a bell:
> Maybe the same fix as in teem, #820429 could help?
> 

Ähm,  #704206,  of course

[toc] | [prev] | [next] | [standalone]


#730290

FromTobias Frost <tobi@debian.org>
Date2016-04-10 22:50 +0200
Message-ID<rmusW-2ny-3@gated-at.bofh.it>
In reply to#730258
On Sun, 10 Apr 2016 20:25:40 +0200 Tobias Frost <tobi@frost.de> wrote:

I tested below patch on fischer.debian.org; and did not see the error
(tried 4 builds)
In the hope this helps...

--
tobi

Regards.
diff -Nru libvigraimpex-1.10.0+git20160211.167be93+dfsg/debian/rules
libvigraimpex-1.10.0+git20160211.167be93+dfsg/debian/rules
--- libvigraimpex-1.10.0+git20160211.167be93+dfsg/debian/rules  2016-
03-23 20:13:15.000000000 +0000
+++ libvigraimpex-1.10.0+git20160211.167be93+dfsg/debian/rules  2016-
04-10 20:12:04.000000000 +0000
@@ -6,11 +6,6 @@
 CXXFLAGS := $(shell env DEB_CXXFLAGS_MAINT_APPEND=-pipe dpkg-
buildflags --get CXXFLAGS; dpkg-buildflags --get CPPFLAGS)
 LDFLAGS := $(shell env DEB_LDFLAGS_MAINT_APPEND='-Wl,--as-needed -Wl,-
z,now' dpkg-buildflags --get LDFLAGS)

-ifeq ($(DEB_HOST_ARCH),i386)
-  CFLAGS += -ffloat-store
-  CXXFLAGS += -ffloat-store
-endif
-
 # support for Ubuntu development
 ifeq ($(DEB_HOST_ARCH),ppc64el)
   CFLAGS := $(subst -O3,-O2,$(CFLAGS))
@@ -29,6 +24,16 @@
   CXXFLAGS += --param ggc-min-expand=20
 endif

+ifneq (,$(filter $(DEB_HOST_ARCH_CPU), i386))
+  CFLAGS += -ffloat-store
+  CXXFLAGS += -ffloat-store
+endif
+ifneq (,$(filter $(DEB_HOST_ARCH_CPU), powerpc ppc64 s390x))
+  CFLAGS += -ffp-contract=off
+  CXXFLAGS += -ffp-contract=off
+endif
+
+
 ifneq ($(wildcard
/usr/lib/$(DEB_HOST_MULTIARCH)/hdf5/serial/libhdf5.so),)
   export CMAKE_INCLUDE_PATH=/usr/include/hdf5/serial
   export CMAKE_LIBRARY_PATH=/usr/lib/$(DEB_HOST_MULTIARCH)/hdf5/serial

[toc] | [prev] | [next] | [standalone]


#730400

FromDaniel Stender <stender@debian.org>
Date2016-04-11 12:40 +0200
Message-ID<rmHqa-4eH-21@gated-at.bofh.it>
In reply to#730290
On Sun, 10 Apr 2016 22:39:39 +0200 Tobias Frost <tobi@debian.org> wrote:
> +ifneq (,$(filter $(DEB_HOST_ARCH_CPU), i386))
> +Â Â CFLAGS += -ffloat-store
> +Â Â CXXFLAGS += -ffloat-store
> +endif
> +ifneq (,$(filter $(DEB_HOST_ARCH_CPU), powerpc ppc64 s390x))
> +Â Â CFLAGS += -ffp-contract=off
> +Â Â CXXFLAGS += -ffp-contract=off
> +endif

Yup, that works. Thank you very much!

Daniel

-- 
4096R/DF5182C8
http://www.danielstender.com/blog/

[toc] | [prev] | [standalone]


Back to top | Article view | linux.debian.bugs.dist


csiph-web