Hi all, For what is I think the second time (fortunately this was only on the Test Tool Shed), I have managed to upload a tar-ball to the wrong repository: 2:fae4084a0bc0 http://testtoolshed.g2.bx.psu.edu/view/peterjc/fastq_paired_unpaired It could have been a glitch, but more likely is was simply inattention on my part (most likely - multiple tab browsing has its downsides). This leads me to offer some suggestions for how to prevent this kind of user error in future: (1) After the upload, show a preview of the changes for confirmation (the drawback is this could take a long time to process for a large upload, and lead to time outs). (2) After upload, before committing the changes, compare the tool contents - adding or removing some tools is normal, replacing ALL the tools with different tools is probably an error and should be subject to user confirmation. (3) Offer a repository action to revert to a prior commit (current workaround is to re-upload the old good file, which I did as 3:81ac324efa49 in this case). Peter
On Thu, May 2, 2013 at 4:29 PM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Hi all,
For what is I think the second time (fortunately this was only on the Test Tool Shed), I have managed to upload a tar-ball to the wrong repository:
2:fae4084a0bc0 http://testtoolshed.g2.bx.psu.edu/view/peterjc/fastq_paired_unpaired
It could have been a glitch, but more likely is was simply inattention on my part (most likely - multiple tab browsing has its downsides).
This leads me to offer some suggestions for how to prevent this kind of user error in future:
(1) After the upload, show a preview of the changes for confirmation (the drawback is this could take a long time to process for a large upload, and lead to time outs).
(2) After upload, before committing the changes, compare the tool contents - adding or removing some tools is normal, replacing ALL the tools with different tools is probably an error and should be subject to user confirmation.
(3) Offer a repository action to revert to a prior commit (current workaround is to re-upload the old good file, which I did as 3:81ac324efa49 in this case).
Peter
Another even simpler change (which would also make a big difference) is to actually have the name of the current repository on the "Upload a single file or a tarball" page. Thanks, Peter
Hi Peter, This fix is now running on the test tool shed. Thanks for pointing out this flaw! Greg Von Kuster On May 2, 2013, at 12:54 PM, Peter Cock wrote:
On Thu, May 2, 2013 at 4:29 PM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Another even simpler change (which would also make a big difference) is to actually have the name of the current repository on the "Upload a single file or a tarball" page.
Thanks,
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/
On May 2, 2013, at 12:54 PM, Peter Cock wrote:
Another even simpler change (which would also make a big difference) is to actually have the name of the current repository on the "Upload a single file or a tarball" page.
Thanks,
On Thu, May 2, 2013 at 9:32 PM, Greg Von Kuster <greg@bx.psu.edu> wrote:
Hi Peter,
This fix is now running on the test tool shed. Thanks for pointing out this flaw!
Greg Von Kuster
Great - that was an easy fix but should cut down on the number of accidental commits due to human error. Let's see how long it is before someone else messes up (or I do it again - grin) and if something more advanced is needed like my first suggestions. Thanks, Peter
On Mon, May 6, 2013 at 9:55 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
On May 2, 2013, at 12:54 PM, Peter Cock wrote:
Another even simpler change (which would also make a big difference) is to actually have the name of the current repository on the "Upload a single file or a tarball" page.
Thanks,
On Thu, May 2, 2013 at 9:32 PM, Greg Von Kuster <greg@bx.psu.edu> wrote:
Hi Peter,
This fix is now running on the test tool shed. Thanks for pointing out this flaw!
Greg Von Kuster
Great - that was an easy fix but should cut down on the number of accidental commits due to human error. Let's see how long it is before someone else messes up (or I do it again - grin) and if something more advanced is needed like my first suggestions.
Well that didn't take long - I just accidentally stuffed up an upload to http://toolshed.g2.bx.psu.edu/view/peterjc/venn_list - the sooner the quick fix to the upload form makes it to the live Tool Shed the better. I'd also urge the related issue of visually distinguishing the main and test Tool Sheds to be prioritised: https://trello.com/card/-/506338ce32ae458f6d15e4b3/817 Regards, Peter
Peter, I've updated the test tool shed with new code to also display (sandbox) in the page title attribute. --Dave B. On 5/6/13 14:11:10.000, Peter Cock wrote:
On Mon, May 6, 2013 at 9:55 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
On May 2, 2013, at 12:54 PM, Peter Cock wrote:
Another even simpler change (which would also make a big difference) is to actually have the name of the current repository on the "Upload a single file or a tarball" page.
Thanks,
On Thu, May 2, 2013 at 9:32 PM, Greg Von Kuster <greg@bx.psu.edu> wrote:
Hi Peter,
This fix is now running on the test tool shed. Thanks for pointing out this flaw!
Greg Von Kuster
Great - that was an easy fix but should cut down on the number of accidental commits due to human error. Let's see how long it is before someone else messes up (or I do it again - grin) and if something more advanced is needed like my first suggestions.
Well that didn't take long - I just accidentally stuffed up an upload to http://toolshed.g2.bx.psu.edu/view/peterjc/venn_list - the sooner the quick fix to the upload form makes it to the live Tool Shed the better.
I'd also urge the related issue of visually distinguishing the main and test Tool Sheds to be prioritised: https://trello.com/card/-/506338ce32ae458f6d15e4b3/817
Regards,
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/
On Tue, May 7, 2013 at 5:28 PM, Dave Bouvier <dave@bx.psu.edu> wrote:
Peter,
I've updated the test tool shed with new code to also display (sandbox) in the page title attribute.
--Dave B.
Thanks - that will help. I've not looked how you did this, but ideally this text and the visual style would be user-configurable (for anyone hosting their own local Tool Shed). Regards, Peter
The text content is specified by the brand option in your .ini file. For visual styles, you have the option of specifying a different location for the CSS in your .ini file, with the static_style_dir option. Note that if you're running behind a proxy configured to serve static files directly, that will also have to be updated to reflect any changes you've made. --Dave B. On 5/7/13 12:39:07.000, Peter Cock wrote:
On Tue, May 7, 2013 at 5:28 PM, Dave Bouvier <dave@bx.psu.edu> wrote:
Peter,
I've updated the test tool shed with new code to also display (sandbox) in the page title attribute.
--Dave B.
Thanks - that will help.
I've not looked how you did this, but ideally this text and the visual style would be user-configurable (for anyone hosting their own local Tool Shed).
Regards,
Peter
On Thu, May 2, 2013 at 4:29 PM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Hi all,
For what is I think the second time (fortunately this was only on the Test Tool Shed), I have managed to upload a tar-ball to the wrong repository:
2:fae4084a0bc0 http://testtoolshed.g2.bx.psu.edu/view/peterjc/fastq_paired_unpaired
It could have been a glitch, but more likely is was simply inattention on my part (most likely - multiple tab browsing has its downsides).
Oops, I did it again: http://toolshed.g2.bx.psu.edu/view/peterjc/get_orfs_or_cdss
This leads me to offer some suggestions for how to prevent this kind of user error in future:
(1) After the upload, show a preview of the changes for confirmation (the drawback is this could take a long time to process for a large upload, and lead to time outs).
(2) After upload, before committing the changes, compare the tool contents - adding or removing some tools is normal, replacing ALL the tools with different tools is probably an error and should be subject to user confirmation.
(3) Offer a repository action to revert to a prior commit (current workaround is to re-upload the old good file, which I did as 3:81ac324efa49 in this case).
Peter
On the bright side, the small improvement we discussed back in May to have the current repository name on the "Upload a single file or a tarball" page is definitely helping. Am I still the only person sometimes careless enough to mess this up and need an automated safety net? Thanks, Peter
participants (3)
-
Dave Bouvier -
Greg Von Kuster -
Peter Cock