Fixed item 3. The translate() method in the JSPWikiMarkupParserTest was simply creating nodes and not saving them, relying on pageExists() to tell it if they had been created. But in 3.0 you can create a page without saving it -- different than 2.x, where creating automatically means the page is persisted. Forcing the page-save before the rendering speed test fixed the problem.
Anyway, it was a bug in the test code, not the core code. All good! Andrew On Sat, Oct 17, 2009 at 5:01 PM, Andrew Jaquith <[email protected]> wrote: >> I'm currently working on moves (the SNS support is really hard to do without >> screwing everything); I have something new on the hard drive already which >> may be fixing this. > > Excellent. I'll cool my jets on these until something new hits the Priha > trunk. > >> This suggests a serious bug in JSPWiki, which is creating a new page *by the >> same name* as an existing page. Since JSPWiki currently is using >> "nt:unstructured" as the page type, Priha will happily allow these so-called >> same-name-siblings - as specified by JCR spec. >> >> This functionality was turned off in 0.5.3, but it's enabled again in trunk. >> However, in this case the test will pass but JSPWiki will not work >> correctly. >> >> I think we should either switch the node type, or add an assertion for pages >> which already exist. > > Hmm. I haven't switched back the isNew() stuff in ContentManager just > yet. I'll try that first. If it doesn't fix the problem, I'll add some > existence-checking code into addPage(). > > Andrew >
