Lee Rhodes wrote: > I picked 4.2.0 because according to gcc.gnu.org it is the "current > release". Although, since I have to operate in a cygwin environment I am > having my doubts. The 4.2 branch suffered from a long delay post-branch before release, and so it diverged significantly from mainline. Since it's a release branch it's in regression-fix only mode which means a lot of new and great stuff that's gone into 4.3 since the branch (which was so long ago - 8 months or so) will never be in 4.2. This divergence got so bad that there was some talk of just skipping 4.2 as a release, but too many people had spent too much time backporting stuff into 4.2 from HEAD that this was immediately ruled out. Anyway, my point is that even though 4.2 is the current stable release, it's quite far removed from the latest in terms of features and development. For example, there was a massive upgrade of the copy of libtool in HEAD, going from a hacked up forked local copy from 2001 to a current upstream libtool top-of-tree version. This is particularly relevant to Cygwin/MinGW because it gives us a fighting chance of finally getting functional shared target libraries working. Since this happened fairly recently and it's not a fix for a regression there's little chance that it will ever get backported to 4.2, and so 4.2 will probably be forever broken on Cygwin/MinGW in this area, which means it's not a very attractive candidate for the Cygwin and MinGW teams to release stable binaries. Danny Smith has mentioned that he intends to release a 4.2 MinGW package, but implicit in that is that he would either backport important things from 4.3, or locally patch problems (for example his out-of tree method for building libgcc_s.dll manually by hand), so it's still 4.2 "plus". That is all just a long winded way of trying to explain why people that care about gcc on Cygwin/MinGW tend to focus on HEAD and why the 4.2 branch isn't all that interesting. > How do I find/apply Dave Korn's patch to stdio.h / newlib? Is it > incorporated into any Cygwin contributed snapshots, or is it a pure manual > source edit? Sigh. No, you shouldn't go manually editing anything. Yes, the fixes are the Cygwin snapshots, and this is precisely why I said you need to use a snapshot back three or four messages ago at the beginning of this thread. :) Brian
