On Wed, Aug 14, 2013 at 11:56 AM, Peter Cock <p.j.a.c...@googlemail.com> wrote:
> Hi Greg,
>
> I'm hoping you (or Dave) can throw a little light on why the
> NCBI BLAST+ nightly tests are not working again yet:
>
> http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus/949fa0294c0d
>
> e.g. Fatal error: Exit code 127 () /bin/sh: 1: blastn: not found
> /bin/sh: 1: blastn: not found
>
> As of that commit, the ncbi_blast_plus repository depends on
> the new IUC repository package_blast_plus_2_2_26 which
> now depends on the IUC repository package_boost_1_53,
> which depends on Bjoern's package_bzlib_1_0, i.e. this chain:
>
> http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus ->
> http://testtoolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26 ->
> http://testtoolshed.g2.bx.psu.edu/view/iuc/package_boost_1_53 ->
> http://testtoolshed.g2.bx.psu.edu/view/bgruening/package_bzlib_1_0
>
> Sadly the (Test) Tool Shed isn't showing me any install logs,
> so I am hoping *you* have access to them.
>
> I've also requested individual package repositories (i.e. type
> "Tool dependency definition") be tested nightly for installation
> problems:
> https://trello.com/c/HVGrShnC/1042-tool-shed-should-test-installation-of-packages
>
> Meanwhile, I have tried this locally - the dependencies seem
> to install, but the BLAST+ package says "Installed, missing
> tool dependencies". It looks like the BLAST+ make fails,
> but I don't know how/if I can see this from the Galaxy Admin
> web-interface.
>
> I can see the logs at the command line,
>
> $ tail -n 30 
> blast+/2.2.26+/iuc/package_blast_plus_2_2_26/9ff9bfdcd04f/INSTALLATION.log
> make[1]: Leaving directory
> `/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/GCC412-Debug64/build'
> #############################################
>
> #############################################
> cd c++ && ./configure --with-boost=$BOOST_ROOT_DIR
> --prefix=/opt/galaxy-dist-shed-tools/blast+/2.2.26+/iuc/package_blast_plus_2_2_26/9ff9bfdcd04f
> && make && make install
> STDERR
> basename: missing operand
> Try `basename --help' for more information.
> configure: WARNING: Untested Boost version; may prove incompatible.
> configure: WARNING: If so, please re-run this script with the flag
> --without-boost.
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/db/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/db/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/ctools/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/ctools/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/misc/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/misc/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/gui/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/gui/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/sample/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/sample/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/internal/Makefile.in]
> Error 1 (ignored)
> make[1]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/internal/Makefile.in]
> Error 1 (ignored)
> make[2]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/corelib/test/Makefile.in]
> Error 1 (ignored)
> make[2]: 
> [/opt/galaxy-dist/database/tmp/tmp-toolshed-mtdutx4hN/ncbi-blast-2.2.26+-src/c++/src/corelib/test/Makefile.in]
> Error 1 (ignored)
> /bin/sh: /usr/bin/touch: No such file or directory
> make[3]: *** [sources.usr] Error 127
> make[2]: *** [sources.usr.locked] Error 2
> make[1]: *** [all_r.real] Error 5
> make: *** [all] Error 2
> #############################################
>
> I just checked the BOOST bit worked by manually compiling
> BLAST+,
>
> $ echo 
> 'BOOST_ROOT_DIR=/opt/galaxy-dist-shed-tools/boost/1.53.0/iuc/package_boost_1_53/a72f8efe9201/boost;
> export BOOST_ROOT_DIR'
> $ ./configure --with-boost=$BOOST_ROOT_DIR
> --prefix=/opt/galaxy-dist-shed-tools/blast+/2.2.26+/iuc/package_blast_plus_2_2_26/9ff9bfdcd04f
> ...
> ******* CONFIGURATION SUCCESSFUL *******
> $ make
> ...
> /bin/sh: /usr/bin/touch: No such file or directory
> ...
> make: *** [all] Error 2
>
> So sign of the basename error, but the touch problem persists. Fixed with:
>
> $ sudo ln -s /bin/touch /usr/bin/touch
>
> Having created the touch alias, manual compilation appears to be
> working (it just takes a long long time). I intend to repeat this via the
> Galaxy Admin page ...

Failed to compile,

...
make[4]: Entering directory
`/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/GCC412-Debug64/build/app/blast'
/usr/bin/g++  -c  -Wall -Wno-format-y2k -g -fPIC  -D_DEBUG
-D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64
-DNCBI_WITHOUT_MT
-I/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/GCC412-Debug64/inc
-I/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/include
/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/src/app/blast/blastn_app.cpp
-o blastn_app.o
/usr/bin/g++  -Wl,--enable-new-dtags -g -Wl,-E  blastn_app.o
-Wl,-rpath,/opt/galaxy-dist-shed-tools/blast+/2.2.26+/iuc/package_blast_plus_2_2_26/9ff9bfdcd04f/lib
-Wl,-rpath-link,/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/GCC412-Debug64/lib
-L/home/pjcock/Downloads/ncbi-blast-2.2.26+-src/c++/GCC412-Debug64/lib
-lblast_app_util -lblastinput-static -lncbi_xloader_blastdb-static
-lncbi_xloader_blastdb_rmt-static -lxblastformat-static
-lalign_format-static -ltaxon1-static -lblastdb_format-static
-lgene_info-static -lxalnmgr-static -lblastxml-static -lxcgi-static
-lxhtml-static -lxblast-static -lxalgoblastdbindex-static
-lcomposition_adjustment-static -lxalgodustmask-static
-lxalgowinmask-static -lseqmasks_io-static -lseqdb-static
-lblast_services-static -lxobjutil-static -lxobjread-static
-lcreaders-static -lxnetblastcli-static -lxnetblast-static
-lblastdb-static -lscoremat-static -ltables-static -lxalnmgr-static
-lncbi_xloader_genbank-static -lncbi_xreader_id1-static
-lncbi_xreader_id2-static -lncbi_xreader_cache-static
-lxconnext-static -lxconnect-static -ldbapi_driver-static
-lncbi_xreader-static -lid1-static -lid2-static -lseqsplit-static
-lxcompress-static -lxobjmgr-static -lgenome_collection-static
-lseqset-static -lseqedit-static -lseq-static -lseqcode-static
-lsequtil-static -lpub-static -lmedline-static -lbiblio-static
-lgeneral-static -lxser-static -lxutil-static -lxncbi-static -lz
-lbz2   -ldl -lnsl -lrt -lm  -o blastn
collect2: ld returned 1 exit status
make[4]: *** [blastn] Error 1

Presumably this machine is missing one dependency or other?

> However, this doesn't help with knowing what might be going wrong
> during the (Test) Tool Shed nightly tests.

I'm still curious how this is failing on the Test Tool Shed.

However, I'm going back to the binary download route... it just
seems like a more robust & reproducible solution all round:

http://testtoolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26/c85236cd100d

This seems to work OK for 64 bit Linux, but once again, fingers
crossed for the next nightly Tool Shed tests...

Peter
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/

Reply via email to