Re: [galaxy-dev] Tool Shed packages for BLAST+ binaries
Hi Dave, I think this is a new regression from the Test Tool Shed installation framework, http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus/f2478dc77ccb Tool test results Automated test environment Tools missing tests or test data Installation errors - no functional tests were run for any tools in this changeset revision Tool dependencies TypeNameVersion blast+ package 2.2.27 Error ("Invalid file %s specified, ignoring set_environment_for_install action.",/ToolDepsTest/blast+/2.2.27/iuc/package_blast_plus_2_2_27/eab09bc4d63e/env.sh) Is that really a comma in the path? Is this a simple typo in a config file? Note this tool is using: http://testtoolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_27/eab09bc... (using platform specific actions) -- Meanwhile, over on the main Tool Shed, which I would like to update to use arch-specific actions for the packages (since that is now in the stable Galaxy releases), and switch from BLAST+ 2.2.26 to 2.2.27, we have a different install failure: http://toolshed.g2.bx.psu.edu/view/devteam/ncbi_blast_plus/70e7dcbf6573 still using 2.2.26 via shell magic, http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26/40c69b76b46... /bin/sh: 1: blastn: not found Peter
Peter, Thanks for bringing this to our attention, we're working on fixing a number of issues with the test framework, and hope to have more information for you tomorrow. --Dave B. On 2013-11-11 06:11, Peter Cock wrote:
Hi Dave,
I think this is a new regression from the Test Tool Shed installation framework,
http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus/f2478dc77ccb
Tool test results Automated test environment Tools missing tests or test data Installation errors - no functional tests were run for any tools in this changeset revision Tool dependencies TypeNameVersion blast+ package 2.2.27 Error ("Invalid file %s specified, ignoring set_environment_for_install action.",/ToolDepsTest/blast+/2.2.27/iuc/package_blast_plus_2_2_27/eab09bc4d63e/env.sh)
Is that really a comma in the path? Is this a simple typo in a config file?
Note this tool is using: http://testtoolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_27/eab09bc... (using platform specific actions)
--
Meanwhile, over on the main Tool Shed, which I would like to update to use arch-specific actions for the packages (since that is now in the stable Galaxy releases), and switch from BLAST+ 2.2.26 to 2.2.27, we have a different install failure:
http://toolshed.g2.bx.psu.edu/view/devteam/ncbi_blast_plus/70e7dcbf6573 still using 2.2.26 via shell magic, http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26/40c69b76b46... /bin/sh: 1: blastn: not found
Peter
On Tue, Nov 12, 2013 at 3:33 AM, Dave Bouvier <dave@bx.psu.edu> wrote:
Peter,
Thanks for bringing this to our attention, we're working on fixing a number of issues with the test framework, and hope to have more information for you tomorrow.
--Dave B.
Hi Dave, Good news, the BLAST+ tests appear to have all passed on the Test Tool Shed, http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus/f2478dc77ccb Tool test results Automated test environment Time tested: ~ 5 hours ago System: Linux 3.8.0-30-generic Architecture: x86_64 Python version: 2.7.4 Galaxy revision: 11318:7553213e0646 Galaxy database version: 117 Tool shed revision: Tool shed database version: Tool shed mercurial version: Tests that passed successfully Tool id: blastxml_to_tabular Tool version: blastxml_to_tabular Test: test_tool_000000 (functional.test_toolbox.TestForTool_testtoolshed.g2.bx.psu.edu/repos/peterjc/ncbi_blast_plus/blastxml_to_tabular/0.0.11) ... Tool id: ncbi_tblastx_wrapper Tool version: ncbi_tblastx_wrapper Test: test_tool_000000 (functional.test_toolbox.TestForTool_testtoolshed.g2.bx.psu.edu/repos/peterjc/ncbi_blast_plus/ncbi_tblastx_wrapper/0.0.21) Curiously however, this no longer seems to be complaining about the BLAST+ tools without any tests - a new bug? Over on the main Tool Shed, the binary installation seems to be failing (still using the bash script magic - is the test system still missing bash, or is there a different problem here?). Here too, there doesn't seem to be any mention of the tools missing tests. At this point (given it is working on the Test Tool Shed), I think it should be safe to update the BLAST+ packages to use the new architecture/os specific <action> tags (a recent feature which is now supported in the stable Galaxy releases): http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26 http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_27 http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_28 Any objections? Thanks, Peter
On Thu, Nov 14, 2013 at 4:35 PM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Hi Dave,
Good news, the BLAST+ tests appear to have all passed on the Test Tool Shed, http://testtoolshed.g2.bx.psu.edu/view/peterjc/ncbi_blast_plus/f2478dc77ccb
...
Over on the main Tool Shed, the binary installation seems to be failing (still using the bash script magic - is the test system still missing bash, or is there a different problem here?). Here too, there doesn't seem to be any mention of the tools missing tests.
At this point (given it is working on the Test Tool Shed), I think it should be safe to update the BLAST+ packages to use the new architecture/os specific <action> tags (a recent feature which is now supported in the stable Galaxy releases):
http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_26 http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_27 http://toolshed.g2.bx.psu.edu/view/iuc/package_blast_plus_2_2_28
I've just done that on the main Tool Shed, fingers crossed that will give a green light on the overnight test results for the BLAST+ wrappers. Peter
participants (2)
-
Dave Bouvier -
Peter Cock