What would be the consequences of using the 6.50 POIDE, headers, libs etc... with the 6.00 compiler?
I've tried it experimentally... but I'm wondering if there's a hidden "gotcha" waiting...
You're trying to get around the optimization bug you posted? ???
Ralf
Quote from: Bitbeisser on December 17, 2011, 12:54:12 AM
You're trying to get around the optimization bug you posted? ???
Ralf
Yep... I was wondering if dropping in the earlier compiler was a safe thing to do until the problem is fixed ... I can go back to 6.00 entirely, but then I have to rebuild all my project files because of the imbedded versions... So it seemed easier to just swap out POCC and go from there...
Sorry to bump this... but it's a priority for me to find this out....
Anyone?
Sorry, I can't tell you more than what you might have found out yourself.
Just using the IDE should certainly be fine, don't think that there is a problem with using libs or headers unless something in the calling convention would have changed...
Btw, have you tried to use an earlier 6.50 release to find out when this issue was possibly introduced?
Ralf
Quote from: Bitbeisser on December 17, 2011, 10:06:00 PM
Sorry, I can't tell you more than what you might have found out yourself.
Just using the IDE should certainly be fine, don't think that there is a problem with using libs or headers unless something in the calling convention would have changed...
Btw, have you tried to use an earlier 6.50 release to find out when this issue was possibly introduced?
Ralf
Apparently it began with 6.50 and has been there since the beginning. RC1 had it... Also note a similar problem has been cited on this forum before but never fully tracked down.
I don't think it's a flaming diasater or anything like that... but since it was first raised on a different forum I've found a few spots in my own code where pointer manipulations have gone wrong, quite possibly because of it.
So if possible, I'd like to side step the problem and with the current snails pace of updates, I thought reverting the compiler back a version might be a good solution... although I can also just compile new code with the optimizations off for now.
Addendum.... I've done some further experimenting and updated the bug report to note that it only appears to be a problem with pointers to string literals... which is sufficiently obscure that I can ignore it for the time being.