Bug#834862: [Debian-med-packaging] Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
Hi, >>> I’ll see what I can do. However, to be honest I’m not sure I have the time >>> to do porting work on software in which using hardware optimised approaches >>> is probably a main selling point… I would be more than happy to raise the >>> issue with upstream though. >> >> OK, thanks! If there's no realistic prospect of supporting other >> architectures, you could also restrict the Architecture field >> accordingly, so that the excluded autobuilders won't bother trying to >> touch the package. > > Sure, I’ll just wait for upstream’s reply to the GitHub issue I raised. If > there’s no success after some time then I’ll restrict the architectures. > Since HISAT2 is very similar to Bowtie2 as far as the build system and > implementation is concerned, I guess that is also a viable option, as that’s > how it’s done in that package. FYI, since there hasn’t been any reaction from upstream I’m just going to adjust the architecture field soon. Cheers Sascha
Bug#834862: [Debian-med-packaging] Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
Hi Aaron, >> I’ll see what I can do. However, to be honest I’m not sure I have the time >> to do porting work on software in which using hardware optimised approaches >> is probably a main selling point… I would be more than happy to raise the >> issue with upstream though. > > OK, thanks! If there's no realistic prospect of supporting other > architectures, you could also restrict the Architecture field > accordingly, so that the excluded autobuilders won't bother trying to > touch the package. Sure, I’ll just wait for upstream’s reply to the GitHub issue I raised. If there’s no success after some time then I’ll restrict the architectures. Since HISAT2 is very similar to Bowtie2 as far as the build system and implementation is concerned, I guess that is also a viable option, as that’s how it’s done in that package. Cheers Sascha signature.asc Description: Message signed with OpenPGP using GPGMail
Bug#834862: [Debian-med-packaging] Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
Sascha Steinbiss writes: > I’ll see what I can do. However, to be honest I’m not sure I have the time to > do porting work on software in which using hardware optimised approaches is > probably a main selling point… I would be more than happy to raise the issue > with upstream though. OK, thanks! If there's no realistic prospect of supporting other architectures, you could also restrict the Architecture field accordingly, so that the excluded autobuilders won't bother trying to touch the package. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Bug#834862: [Debian-med-packaging] Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
Hi Aaron, many thanks for the follow-up. > This change helped, but non-x86 builds are still failing due to the > unavailability of cpuid.h on other architectures: > > processor_support.h:17:22: fatal error: cpuid.h: No such file or directory > # include > > Could you please conditionalize this directive appropriately? Also, > please make sure that any amd64 assembly is conditionalized out in favor > of portable C/C++ on other architectures. I’ll see what I can do. However, to be honest I’m not sure I have the time to do porting work on software in which using hardware optimised approaches is probably a main selling point… I would be more than happy to raise the issue with upstream though. Cheers Sascha signature.asc Description: Message signed with OpenPGP using GPGMail
Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
notfixed 834862 2.0.4-2 found 834862 2.0.4-2 thanks Sascha Steinbiss writes: >* Remove arch-restrictive compiler flags. > Closes: #834862 Hi, Sascha. This change helped, but non-x86 builds are still failing due to the unavailability of cpuid.h on other architectures: processor_support.h:17:22: fatal error: cpuid.h: No such file or directory # include Could you please conditionalize this directive appropriately? Also, please make sure that any amd64 assembly is conditionalized out in favor of portable C/C++ on other architectures. Thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Bug#834862: hisat2: FTBFS on non-x86: -msse2 (and often -m32) unrecognized
Source: hisat2 Version: 2.0.4-1 Severity: important Justification: fails to build from source Builds of hisat2 for non-x86 architectures failed due to its usage of -msse2, and in some cases also -m32. Even where supported, these options can limit portability, or in the case of -m32, select the wrong ABI. g++: error: unrecognized command line option ‘-m32’ g++: error: unrecognized command line option ‘-msse2’ Could you please take a look? Thanks!
