On 29 May 2012 21:24, Dobrica Pavlinusic <dpavlin@rot13.org> wrote:
On Tue, May 29, 2012 at 10:00:12AM +0200, Paul Poulain wrote:
Le 29/05/2012 09:50, Nicole Engard a écrit :
Just to add in my opinion here.
As someone who has been in Paul's situation, where a library (that has paid us to write them code) has been testing (and sometimes using the code in production) and has confirmed that things work I agree that putting a sign off in their name should be an okay practice. I also agree though that someone not from my company should QA the patch - that extra set of outside eyes is essential. Thanks Nicole, I'm "happy" to see BibLibre is not the only one facing this kind of problem.
I would just like to add that we are in same situation with EAN-13 barcode support[1]. They have running it in production, signing off patches is somewhat of high bar for them.
1: http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=6448
Is logging into bugzilla and commenting on the bug too much of a high bar also? Or, how about using gerrit, to sign off, eg http://gerrit.workbuffer.org:8080/#/c/7/ If they were logged in, they could mark that verified. Or use the Biblibre sandboxes to do something similar? I think making it easier for people to do the first sign off is a good thing, of course I'd still like to see ideally 2 other disinterested parties checking it before it was pushed too. Chris