Bug#881149: htslib FTBFS: test failures on i386

Control: tags -1 + patch I found the documentation of the optimization flags for GCC 7.2.0 [1]. I was then able to bisect the list of flags enabled for -O2 and determine which ones needed to be disabled in order to build htslib on i386. Patch follows. --- a/debian/rules +++ b/debian/rules @@

Bug#881149: htslib FTBFS: test failures on i386

Another FWIW, building on i386 with -O1 instead of -O2 and dropping -fno-strict-aliasing is successful. Where can one find the differences between -O1 and -O2 in GCC 7? What changed between GCC 6 and 7 would be useful too. --- a/debian/rules +++ b/debian/rules @@ -7,7 +7,10 @@ include

Bug#881149: htslib FTBFS: test failures on i386

So it turns out this is a re-hash of #865012. See upstream issue reported by Sascha: https://github.com/samtools/htslib/issues/565 Also proposed patch (which was not accepted by upstream): https://github.com/samtools/htslib/pull/571 FWIW, the proposed patch "works for me" on amd64 and i386.

Bug#881149: htslib FTBFS: test failures on i386

On 8 November 2017 at 20:21, Adrian Bunk wrote: > Assuming your CPU and kernel support 64bit, this is working. TIL, thanks!

Bug#881149: htslib FTBFS: test failures on i386

On Wed, Nov 08, 2017 at 08:08:30PM +0200, Graham Inggs wrote: > On 8 November 2017 at 19:57, Adrian Bunk wrote: > > All combinations are possible, for example: > ... > > bash: /usr/bin/hello: cannot execute binary file: Exec format error > > OK, but you can't actually execute

Bug#881149: htslib FTBFS: test failures on i386

On 8 November 2017 at 19:57, Adrian Bunk wrote: > All combinations are possible, for example: ... > bash: /usr/bin/hello: cannot execute binary file: Exec format error OK, but you can't actually execute amd64 binaries on i386, right? So I don't understand how being able to

Bug#881149: htslib FTBFS: test failures on i386

On Wed, Nov 08, 2017 at 07:44:43PM +0200, Graham Inggs wrote: >... > BTW, are you sure it is possible to install amd64 packages on i386? I > know i386 on amd64 is possible, but I've never heard this before. All combinations are possible, for example: dpkg --add-architecture s390x apt-get update

Bug#881149: htslib FTBFS: test failures on i386

On 8 November 2017 at 17:10, Adrian Bunk wrote: > The relevant change is gcc 6 -> 7. Thanks. I've just checked, and the last successful i386 build in Ubuntu was on 2017-08-04 against gcc 6. > With multiarch allowing installation of amd64 packages even on an i386 >

Bug#881149: htslib FTBFS: test failures on i386

On Wed, Nov 08, 2017 at 10:25:31AM +0200, Graham Inggs wrote: > Package: htslib > Version: 1.5-1 > Severity: serious > Tags: buster sid > > As per logs from reproducible builds [1], htslib 1.5-1 recently > (around 2017-09-25) started to FTBFS on i386 with the following test > failures: The

Bug#881149: htslib FTBFS: test failures on i386

Better link: https://tests.reproducible-builds.org/debian/history/i386/htslib.html

Bug#881149: htslib FTBFS: test failures on i386

Package: htslib Version: 1.5-1 Severity: serious Tags: buster sid As per logs from reproducible builds [1], htslib 1.5-1 recently (around 2017-09-25) started to FTBFS on i386 with the following test failures: test_vcf_api: /build/1st/htslib-1.5/test/test-vcf-api /tmp/o6gEIzeYrS/test-vcf-api.bcf