FINALLY took a couple of hours late, late, late tonight and looked at the pesky little kpilot bug which caused a nasty little SIGSEGV with the vcal-conduitbase::slotProcess code during a “copy handheld to PC” sync. Turns out there was a single missing set for the fNextState in the case of ConduitAction::SyncMode::eCopyHHToPC, which as luck would have it, is exactly what I was trying to do.
So, yay me on fixing that annoying buglet which was niggling at the back of my brain for the last couple of months.
Secondarily, one of: gdb, PowerPC Linux, AAP sucks donkey elbows with regards to trying to debug this problem I was having. I could not for the life of me attach gdb to kpilot’s pid and get it to give me a decent backtrace, nor could I get it to set a breakpoint inside the source code for kpilot’s conduit libraries!@#$@#$%!!!!! I swear this used to be possible!
Tertiarily (?), both Eclipse 3.2/CDT and Kdevelop crash ENTIRELY too often for me, making them both just about darned un-useful. Had to revert to using vi and std::cout to track down this buglet!! And I was SO hopeful that CDT would finally give me a good IDE for C++, but either I’m doing something stupid or it’s still not able to resolve declarations and definitions on a consistent, useful basis.
So, in summation:
- Darnit, it feels REALLY good to get a chance to code on KDE again!
- Darnit, it was frustrating to have Eclipse/CDT and Kdevelop crash so stinking much on me
- Darnit, it’s late!
Good night, Gracie.