Despite lines in 4.maf like
```
e echTel1.scaffold_309554 3896 34801 + 42008 I
```
echTel1 seems not to be in the metadata table.
So I have chosen another one.
Somewhere between file/libmagic 5.36 and 5.39, the output of `file` on a
`.csi` file changed, breaking the ``metadata_bcf`` framework test with
the following error:
```
if found_exceptions:
> raise JobOutputsError(found_exceptions, job_stdio)
E galaxy.tool_util.verify.interactor.JobOutputsError: Output output_of_input_metadata: different than expected, difference (using contains):
E ( /tmp/tmpu1a3dgjcbcf_index_metadata_test.txt v. /tmp/tmpkt78tzggbcf_index_metadata_test.txt )
E Failed to find 'gzip compressed data, extra field' in history data. (lines_diff=0):
lib/galaxy/tool_util/verify/interactor.py:1222: JobOutputsError
```
So add a requirement for the latest `file` version and update the test
accordingly.
some follow up fixes to
https://github.com/galaxyproject/galaxy/pull/9279/fileshttps://github.com/galaxyproject/galaxy/pull/9776/files
- wrong profile for the profile test tool (it used 20.01, but should
have used 20.05)
- the non-profile test tool used wrong assumptions. now fixed
and also tested against 19.09
- in the profile version `value_wrapper.value` needs to be casted
for `multiple==False` otherwise we gat `"None"` (which I do not
understand 100%)
some follow up fixes to
https://github.com/galaxyproject/galaxy/pull/9279/fileshttps://github.com/galaxyproject/galaxy/pull/9776/files
- wrong profile for the profile test tool (it used 20.01, but should
have used 20.05)
- the non-profile test tool used wrong assumptions. now fixed
and also tested against 19.09
- in the profile version `value_wrapper.value` needs to be casted
for `multiple==False` otherwise we gat `"None"` (which I do not
understand 100%)
rename entries such that they are suffixed by the column name
(execpt for dbkey).
this makes tests more specific, because this ensures that always
the correct column is used.
Marius van den Beek