From: Bryce H. <br...@os...> - 2003-04-25 21:35:26
|
On Fri, 25 Apr 2003, Nick Jennings wrote: > On Fri, Apr 25, 2003 at 12:07:58PM -0700, Rob Helmer wrote: > > On Fri, Apr 25, 2003 at 10:51:59AM -0700, Bryce Harrington wrote: > > > > > For some reason that I could not determine, the javascript for one of > > > the tests in repository.js always fails. This is the > > > isEmpty(document.repository.repo_root.value) validation command. The > > > other validation commands work fine. I tried it on Konqueror and > > > Netscape 4.x. I ended up just commenting it out. > > > > I only test SW with IE and Mozilla, last I checked it worked in > > Konq but it could use testing. I'll try to reproduce this. > > Hmm, I didn't catch this in the first email. Bryce, if it's possible > to use a fairly recent Mozilla, at least for cross-testing purposes to > find the problem, that would be great. I also tested with Mozilla, forgot to mention that. Probably not terribly recent though. I believe most of your browser-dependency issues would disappear if use of javascript in the entry forms was avoided, especially for the submit buttons. There are also some techniques for making javascript that'll run on a wider variety of browsers. > > > I was getting errors when trying to update user info and it wasn't > > > "setting". On line 2977 it sets the user params via: > > > > > > $user->{'personal'}->{"$param"} = param("$param") || ''; > > > > > > However there was no $user->{'personal'} value in the $user hash at this > > > point so it failed. To fix this I added above the loop: > > > > > > $user->{personal} = { }; > > > > Hmm... haven't seen this before. > > My guess is the hash structure did not exist because it wasn't reading > the config file correctly, if at all. This was my take as well. Bryce |