> Thanks so much for the patch. Looking at the source I noticed some > small things (care to build a bike shed with me? :). > > Many of the test EXPECT's have %d changed to hard coded line numbers and > others have %s changed to hard coded file names. In some of them, > that's all that changed. I assume that will be cleaned up before > committing.
Yes, this can be tidied up, I'd like to get some feedback on how far we should go with the behavior change, before I attack it again.: For non-numeric value used to access string offsets: a) still breaks BC due to array dereferencing ISSET: returns true , issues warnings (this is rather strange behavior) READ: warnings only b) fix isset, and return first character ISSET: return false (no warning) READ: warning + return first character (similar to current behavior except dereferenced strings) c) fix isset, and return empty strings ISSET: return false (no warning) READ: warning + return empty string All of these break BC in subtle ways, I think only b) or c) are viable options, a) without the isset warnings does not really solve much unless the dereferencing is reverted, and it is used to warn people of an impending behavior change in the next release... > > There are some extraneous whitespace changes in the source. > I tried to apply the patch to 54 and trunk via "patch < bug.diff" > (which usually works fine) and all hunks failled. Not sure what I'm > doing wrong. I've only done it against trunk, not sure what else has changed in 54 that would cause it to fail though. > > Thanks, > > --Dan > > -- > T H E A N A L Y S I S A N D S O L U T I O N S C O M P A N Y > data intensive web and database programming > http://www.AnalysisAndSolutions.com/ > 4015 7th Ave #4, Brooklyn NY 11232 v: 718-854-0335 f: 718-854-0409 -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
