On 09/04/2010 03:49, Gary Funck wrote: > I'm still wondering: Do GCC developers routinely > test their patches on MIPS, ARM, and S390 platforms > (for example)? I signed up for the 'cfarm' and don't see > an S390 there, and some of the secondary targets > look like they might be really SLOW?
Well, MIPS and ARM are well supported by simulators, which is how a lot of people test patches that are going to affect the core compiler and potentially have side effects on many platforms, e.g. when merging a big branch with some major new functionality to trunk. But very many patches are either clearly target-specific, or clearly generic enough or simple enough that it's seen as fair to apply them after testing on common/popular platforms and fix any side-effects that show up on peculiar or uncommon hosts as they manifest. HEAD is implicitly unstable (except during the stage 3 stabilisation process before a release of course.) cheers, DaveK