We started noticing weirdness in our lotuscript scripts shortly after updating our servers to fp3. Then the weirdness turned into very bad weirdness when our main intranet server started repeatedly crashing.
The problem appears to be a bug in how Lotuscript (and other languages?) access columnvalues (doc.columnvalues) in fp3. You can get an incorrect value or crash, depending on the structure of the data involved. We are using 64-bit domino on windows 2003, don’t know if it applies to 32-bit or not. Rollback to FP2 fixes.
Submitted to Lotus support, SPR is RDJS87KP59, hotfix pending. Just fyi, if others are seeing similar problems.
A little more info on this: lotus confirmed it is an issue in 32-bit windows as well as 64-bit. They confirmed that it may apply to other languages besides lotusscript, but this has not been tested.
Info on timeline (from me, not from lotus)–fix is not available yet, so problem will still exist in FP4 (out now) and 8.5.2 (code freeze now).
I’m having the same problem on two servers. One running 64bit version of 8.5.1FP3 and the other 32bit 8.5.1FP4 .
Haven’t been able to reproduce it on 8.5.2 yet, but can’t confirm that it’s fixed there. Will be contacting support to find out what’s going on and update here if I find out anything good.
Personally, I thought this was a very serious problem, since we also do a lot of columnvalues stuff. Good news is that I’ve just heard, from IBM and from Mr Groen above (thank you) that this is actually fixed in 8.5.2, contrary to what I was originally told (apparently they tested in a beta of 8.5.2). Perhaps at Mr Groen’s prompting, it was discovered that the spr I posted was a duplicate of spr KKOO86VE68 that was in progress.
In our case, the crash had to do with a very specific row of data in one particular view, but we were getting wrong results from other columnvalues. Googling the spr says that it has something to do with having a blank value in a column, though I’m not sure what all the prereqs for the issue are.