Bugzilla – Full Text Bug Listing
| Summary: | [xpath2] Need to Refactor Promotion rule implementation | ||
|---|---|---|---|
| Product: | [WebTools] WTP Source Editing | Reporter: | David Carver <d_a_carver> |
| Component: | wst.xpath | Assignee: | Jesper Moller <jesper> |
| Status: | RESOLVED FIXED | QA Contact: | David Carver <d_a_carver> |
| Severity: | normal | ||
| Priority: | P3 | ||
| Version: | 3.1 | ||
| Target Milestone: | 3.2 M4 | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| Whiteboard: | |||
|
Description
David Carver
Assigning inbox items to triaged since these have all be triaged. I'm tackling this now (and a number of other type promotion issues along the way) Any status update on this, we have about another week before the 3.2M2 cut off date. Progress + regression. I was interrupted by other work (Denmark is now live as the first country with 100% digital registration of real estate titles and mortgages, it just went into production. XML heavy!) Also I found a wad of tests that were expecting the wrong error code (XPST0013), so I'm dealing with that., but I will make the M2 deadline for this bug. (In reply to comment #4) > Progress + regression. > I was interrupted by other work (Denmark is now live as the first country with > 100% digital registration of real estate titles and mortgages, it just went > into production. XML heavy!) > > Also I found a wad of tests that were expecting the wrong error code > (XPST0013), so I'm dealing with that., but I will make the M2 deadline for this > bug. Cool thanks for the update. Yeah I know about those tests expecting the wrong error code, I fixed several the other day in LocalNameFromQNameFuncTest. This wasn't supposed to happen! I thought the cut-off date was by the end of this week. Sorry 'bout that. Anyway, I've been sidetracked by the negative test expectations. There are hundreds of test expectations that are not set correctly (which is taken by the test harness as XPST0003). Once fixed, they revealed that most type errors were reported as XPST0003 as well, which is very wrong. So I'm fixing the error messages first. WTP likes a stabilization period in general before a stable build. Unfortunately this means you don't really have a full six weeks of development. It's a week ramp up (planning, 4 weeks dev, one week ramp down). Not my preferred method, but that is what wtp follows. Anyways, sounds like you could use the extra time anyways. Jesper just checking to see how this is going. We have about 2 weeks before we hit freeze again. I think this is the last big set of unit tests we need to get working. Renaming summary to better reflect scope of the work. Jesper do you have a patch that you could attach to this bug, maybe I can help finish it up over the next week or so. Feel free to attach mylyn context as well. Are you committing to these at the moment? Previous comment was meant for Dave... Bugzilla is not Google Wave. :-) Sorry for the late response, I've fixed up these so they pass the test suite. So the rest of the work is re-implementing the promotion and sub-type rules correctly. I'm going to change the title to reflect the remaining work, and remove the blocking on bug 262765, as it no longer blocks that bug. Latest code is in head. I'm done touching this portion of the code. We have a hand full of tests that still need to pass (about 12). Will take a look at fn:collection and maybe fn:in-scope-prefixes during 3.2M4. Moving to 3.2M4 since we are approaching the ramp down of 3.2M3. If you can get it in for 3.2M3 that is great otherwise shoot for 3.2M4. yeah, it's finally done. Just have to fix up copyrights. Committing Tuesday. Woohoo, committed. I think Hudson ate it without spitting it back out. Actually..yellow build. https://build.eclipse.org/hudson/view/WTP/job/cbi-wtp-wst.xsl.psychopath/lastBuild/testReport/org.eclipse.wst.xml.xpath2.processor.testsuite.functions/SeqAVGFuncTest/test_fn_avg_mix_args_066/ 1 test is failing. Re-Opening...I won't fix, so you get the CI point once it is fixed. Yep, I was too trigger happy, I really thought I had run the tests properly locally as well. the test was somehow wrong, FORG0001 is the correct error code. Fixed in CVS. You can re-spin it if you want. It'll respin itself every 10 minutes if there are new changes. Should be green soon then. And you'll get the points. :) Just check it in a half hour, if it's green, resolve the bug. build back to green resolving. |