On Tue, Nov 05, 2002 at 11:03:22AM +0100, Pavel Tsekov wrote: >What should we do with this package ? It has been almost two weeks since >Lapo posted it and there isn't a single comment. Is it ok to be >uploaded ?
Well, I think my opinion is pretty consistent. Rather than say signal delivery in cygwin is "ill" and provide workarounds in processes which rely on a certain behavior, we should be fixing cygwin's signal delivery, assuming that is really the problem. So, I am always against these hand waving "works for me" patches that people like to provide as in 2 below. I know this doesn't answer the question, but that's my vote. >On Fri, 25 Oct 2002, Lapo Luchini wrote: > >> Ready at the same usual address: >> http://www.lapo.it/tmp/rsync-2.5.5-2.tar.bz2 >> http://www.lapo.it/tmp/rsync-2.5.5-2-src.tar.bz2 >> >> Revision changelog: >> 1 Compiled with gcc version 3.2 20020818 (prerelease) >> 2 Included Anthony Heading's patch to avoid dead child processes >> http://sources.redhat.com/ml/cygwin/2002-09/msg01155.html >> 3 build script updated to support digital signatures >> 4 source package now contains original's package detached gpg signature >> by the author and detached signatures on the script itself and on the >> patch, signed by me >> 5 signatures can be checked with "./rsync-2.5.5-2.sh checksig" >> 6 "./rsync-2.5.5-2.sh all" can still be used to create an unsigned >> package (with only original rsync package signature) >> 7 "SIG=1 ./rsync-2.5.5-2.sh all" can still be used to create a >> signed package (with only original rsync package signature AND >> signatures on script and patch) >> >> 1 and 2 lead me to suggest a "test" period for the package >> 3 is just a nifty feature I liked adding, but it is intended to be >> optional, that's why I have not changed the "normal" build command and >> it correctly checks (and warns) for missing /usr/bin/gnupg an the such >> >> I would like a negative or positive comment on the "improved" script, too =)