Thanks all, I'll have a look at updating this to use a POST request and writing a test around this.
Cheers, Antony On 12 December 2013 17:11, Gary Martin <[email protected]> wrote: > On 12/12/13 16:17, Olemis Lang wrote: >> >> On Thu, Dec 12, 2013 at 11:05 AM, Gary Martin >> <[email protected]>wrote: >> >>> On 12/12/13 15:25, Olemis Lang wrote: >>> >>>> On Wed, Dec 11, 2013 at 8:33 AM, Gary Martin <[email protected] >>>>> >>>>> wrote: >>>> >>>> [...] >>>> >>>> While we could resort to POSTing instead, I can see the reasoning >>>> behind >>>>> >>>>> not doing that. Pragmatism would probably trump such reasoning though. >>>>> I >>>>> think other solutions would either limit the field sizes or warn of >>>>> possible loss of data over some limit. Perhaps I have missed something >>>>> though. >>>>> >>>>> >>>>> >>>> IMO let's use POST >>>> >>>> >>> I think that is probably the correct answer too but I am tempted to >>> commit >>> the patch as is and let Antony continue with enhancements if he wishes. I >>> am unaware of any regressions from his current approach so I think this >>> is >>> already much better behaviour. >>> >>> >> >> Yes , my intention was not to suggest that current solution should not be >> committed . However at least in theory there are two aspects missing IMO >> >> 1. use POST >> 2. testing >> * though this one might be impossible *now* since js code is >> involved >> > > Fair point.. there could conceivably be tests. Anyway, I have committed the > code - my only modifications at this point have been minor style changes. > > I've opened #727 for further work. If Antony is interested in continuing, he > will, of course, be very welcome to! > > Cheers, > Gary
