Hi Andrew, Thanks for the patch. Appreciate the contribution. The fix for BEEHIVE-1118 looks good. There was one minor issue I came across for the diff of the generated validation xml files and the expected results. When I run your test I get a difference in the ordering of the <msg> elements associated to a field rule. Looking at the code, the rules are stored in an ArrayList and we just iterate through them when writing the xml file... so I'm not sure yet why there's a difference with your expected result, but can look into it.
Carlin On 7/5/06, Andrew McCulloch <[EMAIL PROTECTED]> wrote:
Hi Carlin, the work went a bit quicker than I had anticipated... I just attached a patch to BEEHIVE-1118 in JIRA that always diffs the pageflow-validation files. It doesn't look like any of the other compiler bvts are using validation so no additional tests had to be modified. Please take a look whenever you have a chance. On 7/5/06, Andrew McCulloch <[EMAIL PROTECTED]> wrote: > > Thanks Carlin. I will move forward with the intent of always diffing the > pageflow validation config files. With a little luck I should have a patch > by tomorrow for you to review. > > --Andrew > > > On 6/30/06, Carlin Rogers <[EMAIL PROTECTED]> wrote: > > > > Thanks for looking at this Andrew! Appreciate your contributions. > > > > My vote would be to always diff the generated page flow validation > > files. > > > > Carlin > > > > On 6/30/06, Andrew McCulloch < [EMAIL PROTECTED]> wrote: > > > > > > I am working on a patch for BEEHIVE-1118 which involves the order > > that > > > form validations are performed. The bug looks like it comes down to > > the > > > way > > > the pageflow-validation-*.xml file is generated. In addition to the > > three > > > testRecorder drts that need to change with this fix I am writing a new > > > compiler bvt test case. > > > > > > Currently the compiler bvts only diff the struts-config*.xml files > > > > > against expected results. I have a patch to the JUnit class that > > performs > > > the file comparisons to make diffing the generated validation files > > > mandatory (like he struts config files are), however, I wanted to know > > if > > > this is the best choice. It would also be easy to modify the test to > > only > > > diff the file if an expected result file is present. This would only > > miss > > > the case where a validation file is generated unnecessarily. It would > > > > > however avoid changing the other compiler bvts that may use Jpf > > validation > > > annotations already. Alternatively I could add some sort of flag or > > > properties file to trigger the diffing of the validation files only > > when > > > desired. This choice would be extensible and be easily adjusted to > > > handled > > > any other generated files. > > > > > > If this seems like a worthwhile change to the netui compiler bvts > > I > > > would like to hear how others think it would be best to handle > > this. My > > > initial thought is that it would be best to always diff the pageflow > > > validation files. > > > > > > --Andrew > > > > > > > > > > >
