- () TestCaseAutoNestedVerbatim
- () TestCaseAutoSearch
- I think these are just incomplete tests, but i can't be sure
- () TestCaseAutoSearchOrder
- Search ordered by date created - looks like test case is broken by design. no ,v files commited for it.
- ()TestCases.TestCaseAutoInternalTags
- variables that are expected to be linked, and aren't - eg WEB, Bugs
- actually, it looks like the testcases here are just plain wrong, but I simply don't know. - it looks from my testing that we have successfully replicated Cario's stupidity wrt Wiki words that are directly preceeded by a > not being linked.
- ()TestCases.TestCaseAutoIncludes
- () TestCaseAutoIncludeAttachment
- no idea, but look slike there's a missing file
- () TestCaseAutoInOutPre
- no idea, does this require a plugin ?
ok, assuming we don't want to be bug compatible, and that all the testcase failures above are crap testcases there are some bugs to fix.
- if you put a set statement withing a verbatim, it is still used (and is should not be)
- when you include a topic that has a twiki variable withing angle brackets, they are not expended, making me thing that the topic rendering code is not being re-used, leading to inconsistencies
i'm going to take a break and see what pops up.
SD
All "Auto" testcases as of just prior to beta 3 were passing, and should remain that way.
CC
I don't think they are crap testcases. I crafted them all to explicitly test for behaviours that existed in Cairo that I thought were a bit silly but i didn't want to break.
CC
I checked the above tests again, and marked those that still fail.
RA
if you look at the first test in
TestCaseAutoNestedVerbatim there is
no way that this test should pass, and whats more, the expected result is just plain wrong, as it implies that
TWikiVariables get expended within verbatim blocks. This is why i desperatly want more information on what bug is making the test run tell us that the test passed.
-- SD
Sven, read my lips;
don't read the rendered testcase topic - you will just get confused. Only two useful data points exists; first, the 'expected' and 'actual' blocks in the raw testcase, and second, the result of
running the testcase. The rendered topic will do wierd things sometimes due to the strange way rendering is done. The verbatim testcase looks correct to me; it was crafted to be equivalent to cairo. What's the
specific issue you have with it?
CC
closing it, as its due to me being a stupid pedantic moron, that did get confused exactly like crawford predicted -- SD