On Tue, Apr 23, 2013 at 4:01 PM, Dave Bouvier <d...@bx.psu.edu> wrote:
> Peter,
>
> Our approach has always been to treat each tool shed repository as a
> self-contained entity. This means that all required components are contained
> in the repository or definitions for obtaining them from remote repositories
> are contained in the repository.
>
> You'll notice that several tools that have been migrated to the tool shed
> from the distribution have 1.bed in their test-data directory. Our reasoning
> for that approach is that in the unlikely event that the distribution's
> version of 1.bed is changed, the repository version will still be valid for
> that repository's tool functional tests. Additionally, Ssuis.fasta may
> eventually be migrated out of the distribution, in which case any repository
> installed from the tool shed that depends on Ssuis.fasta will fail
> functional tests.
>
> However, I see your point about the file size, 1.9MB is a bit substantial
> for a test file. Is it possible to rewrite the test to use a smaller file?
>
>    --Dave B.

Hmm. If that is the policy then I'll probably fall over some other examples
like this later on. I think in the short term I'll just bundle Ssuis.fasta with
this tool.

In this specific case, I noticed that Ssuis.fasta was included already,
and so it made a nice medium sized test case which would exercise
the code quite well. The alternative is to hand construct multiple
smaller test cases to try and cover corner cases deliberately - much
more work but perhaps a better route in the long term.

Thanks for quickly clarifying this,

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