integer or float parameters - regression on trunk?

Hi all, I've found what I consider to be a regression from Daniel Blankenberg's change:
http://bitbucket.org/galaxy/galaxy-central/changeset/1f4d6b25a84e
Fix for Integer/FloatToolParameter.get_html_field() when 'value' is provided as an integer/float. Fixes an issue seen when saving workflows: If an integer or float tool parameter is changed to a value of 0 or 0.0 and saved, the form field would be redisplayed using the default tool value; and not the value that is now saved in the database.
If have a tool parameter like this: <param name="dummy" type="integer" value="" label="Test integer with empty default"/> Prior to Dan's change, the HTML would show with the value as specified (empty string). After Dan's change, the HTML page for the tool loads with "0" shown, which is not what I wanted (we can get that with value="0"). I think the old functionality is more useful - couple it with a validator and this forces the user to pick a value rather than always having a valid default. However, what I really want to include optional="true" value="" and allow the value to be an integer or the empty string. In this use-case the HTML should start with the given default shown (empty string as value). See also: http://lists.bx.psu.edu/pipermail/galaxy-dev/2010-October/003455.html Does that make sense? Regards, Peter

Hi Peter, Sorry about that, changeset 4463:9d9f0de20f4f should restore the ability to not specify an initial value for integer and float tool parameters. Let us know if you encounter further issues. I've opened a ticket on enhancing additional tool parameters to be optional: http://bitbucket.org/galaxy/galaxy-central/issue/403/enhance-tool-parameters... Thanks for using Galaxy, Dan
Hi all,
I've found what I consider to be a regression from Daniel Blankenberg's change:
http://bitbucket.org/galaxy/galaxy-central/changeset/1f4d6b25a84e
Fix for Integer/FloatToolParameter.get_html_field() when 'value' is provided as an integer/float. Fixes an issue seen when saving workflows: If an integer or float tool parameter is changed to a value of 0 or 0.0 and saved, the form field would be redisplayed using the default tool value; and not the value that is now saved in the database.
If have a tool parameter like this:
<param name="dummy" type="integer" value="" label="Test integer with empty default"/>
Prior to Dan's change, the HTML would show with the value as specified (empty string). After Dan's change, the HTML page for the tool loads with "0" shown, which is not what I wanted (we can get that with value="0").
I think the old functionality is more useful - couple it with a validator and this forces the user to pick a value rather than always having a valid default.
However, what I really want to include optional="true" value="" and allow the value to be an integer or the empty string. In this use-case the HTML should start with the given default shown (empty string as value). See also: http://lists.bx.psu.edu/pipermail/galaxy-dev/2010-October/003455.html
Does that make sense?
Regards,
Peter

On Thu, Oct 21, 2010 at 7:28 PM, Daniel Blankenberg <dan@bx.psu.edu> wrote:
Hi Peter,
Sorry about that, changeset 4463:9d9f0de20f4f should restore the ability to not specify an initial value for integer and float tool parameters. Let us know if you encounter further issues.
Thanks - I'll check this out tomorrow.
I've opened a ticket on enhancing additional tool parameters to be optional: http://bitbucket.org/galaxy/galaxy-central/issue/403/enhance-tool-parameters...
I've subscribed to the issue. In general, should I file enhancement issues for things like this if the mailing list discussion goes quiet? Peter

On Thu, Oct 21, 2010 at 9:33 PM, Peter <peter@maubp.freeserve.co.uk> wrote:
On Thu, Oct 21, 2010 at 7:28 PM, Daniel Blankenberg <dan@bx.psu.edu> wrote:
Hi Peter,
Sorry about that, changeset 4463:9d9f0de20f4f should restore the ability to not specify an initial value for integer and float tool parameters. Let us know if you encounter further issues.
Thanks - I'll check this out tomorrow.
That has fixed it, thank you for the very prompt reply. Peter
participants (2)
-
Daniel Blankenberg
-
Peter