This is important since it blocks the upgrade of the default version of GCC from GCC 4.8 to GCC 4.9. If you cannot fix it, please change to USE_GCC=4.8 temporarily. http://package18.nyi.freebsd.org/data/101amd64-default-PR196712/2015-01-16_07h05m42s/logs/errors/seqan-1.3.1_3.log In file included from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file.h:56:0, from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/misc/misc_cmdparser.h:39, from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/apps/dfi/dfi.cpp:26: /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_generated_forwards.h: At global scope: /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_generated_forwards.h:927:66: error: invalid abstract return type 'seqan::FileFormat<TFile, TData, TMeta, void>' template <typename TFile, typename TData, typename TMeta> inline FileFormat<TFile, TData, TMeta, void> guessFileFormat(TFile & file, TData & data); // "seqan-1.3.1/seqan/file/file_format_guess.h"(65) ^ In file included from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file.h:84:0, from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/misc/misc_cmdparser.h:39, from /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/apps/dfi/dfi.cpp:26: /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:143:8: note: because the following virtual functions are pure within 'seqan::FileFormat<TFile, TData, TMeta, void>': struct FileFormat<TFile, TData, TMeta, void> ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:154:2: note: virtual void* seqan::FileFormat<TFile, TData, TMeta, void>::formatID_() const formatID_() const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:157:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::read_(TFile&, TData&) const read_(TFile & file, TData & data) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:159:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::read_(TFile&, TData&, seqan::FileFormat<TFile, TData, TMeta, void>::TSize) const read_(TFile & file, TData & data, TSize limit) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:162:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::readMeta_(TFile&, TMeta&) const readMeta_(TFile & file, TMeta & meta) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:165:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::goNext_(TFile&) const goNext_(TFile & file) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:168:2: note: virtual seqan::FileFormat<TFile, TData, TMeta, void>::TSize seqan::FileFormat<TFile, TData, TMeta, void>::length_(TFile&) const length_(TFile & file) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:171:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::write_(TFile&, TData&) const write_(TFile & file, TData & data) const = 0; ^ /wrkdirs/usr/ports/biology/seqan/work/seqan-1.3.1/seqan/file/file_format.h:173:2: note: virtual void seqan::FileFormat<TFile, TData, TMeta, void>::write_(TFile&, TData&, TMeta&) const write_(TFile & file, TData & data, TMeta & meta) const = 0; ^
Maintainer CC'd
confirmed on gcc5: http://muscles.dragonflybsd.org/bulk/bleeding-edge-potential/20150216_080917/logs/errors/seqan-1.3.1_3.log
There is a new upstream version to which I will upgrade next week hopefully!
hi, four weeks has passed. Is a new version coming? This port fails continuously in my gcc5 runs. I'd like to see it fixed
There is a new upstream version (2.0.0), but it has some build issues on FreeBSD. I am working on fixes those issues upstream and getting them in 2.0.1 (I am an upstream dev), so that a clean port of the version can be shipped without local patches (and that works with system compiler as well). Changing the current port to depend on a fixed GCC-Version is totally ok for me, if it would otherwise impede any other changes.
okay, I agree the clean 2.0.1 approach is the way to go. To remove pressure from you, I will set the port to require exactly gcc 4.8 today. It is not using clang, it always uses GCC. When the new port comes in, hopefully the system clang can build it and that all versions of gcc (including 5) can build id. When that happens, just remove the USE_GCC= line completely.
A commit references this bug: Author: marino Date: Sat Mar 21 11:53:28 UTC 2015 New revision: 381791 URL: https://svnweb.freebsd.org/changeset/ports/381791 Log: biology/segan: Limit GCC to 4.8 while waiting for update Due to questionable c++, GCC 4.9 and newer cannot build seqan. An update has been in the works that will enable clang compilers and all versions of GCC to build seqan, but it is not quite ready yet. While waiting for it, limit the version of GCC used to 4.8. PR: 196851 Changes: head/biology/seqan/Makefile
Any news on this?
> Any news on this? Current master has all FreeBSD build issues fixed and only one remaining run-time issue (hard to pin down, but not critical for a new release): http://cdash.seqan.de/index.php?project=SeqAn Upstream 2.0.1 should be released in the next two weeks. I will update the port immediately after that.
Hannes, do you have an update on this? The issue is now more than a year old and it does somewhat block the upgrade of lang/gcc.
> Hannes, do you have an update on this? I have proposed changes in #204127 that can be committed immediately that would resolve this! They introduce a new (updated) port seqan-apps and turn biology/seqan into a source-only port.
resolved via #204127