As Brian Dean wrote: > At this point, I'm very reluctant to re-release, unless there is a > security issue.
Sure. > Perhaps for WinAVR, a minor patch can be included to > fix it. WinAVR perhaps could do that, yes. For my win32 version, it's too late, as the zip file is already out there on savannah. I've been short in time this morning when I wrote my previous mail, so I couldn't check for any possible side-effect. Now that I've got a bit more time to check, it seems there's no security issue with that. In any place where a `z' modifier is used, there are no more other formats in the same format string, so the argument skew is harmless. So the only drawback is that a few error or diagnostic (i.e. verbose > 0) messages come out wrong. > If we really need to release again for a very high priority bug or > security fix, I'd say 5.0.1. Defer that to 5.1. I've already got a basic idea how to fix that in a portable way: . add an autoconf test to check whether %zd is supported . depending on the test result, define a macro to either "%u" or "%zu" (drop these %zd's, they are pointless as size_t is unsigned per definitionem) We should probably add a central avrdude.h file for that. This could also take up all the individual external declarations from the various files. -- cheers, J"org .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) _______________________________________________ avrdude-dev mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/avrdude-dev
