On Mon, 3 May 2004 12:23:06 +0100, John Maddock wrote
Switching testing to the release branch is not a problem (at least for us), but a full-fledged release also involves other administrative work, in particular preparing and uploading the archives, tagging the CVS, etc. -- http://tinyurl.com/2meuh. And in general, somebody needs to manage it. If that's going to be taking care of, we'll take care of Win32 regressions. Otherwise, I'd say patches are a way to go, until 1.32.
Well I'm willing to put some time into this, or I wouldn't have suggested it, however there's no point unless there's a general desire for an update, so I would like be interested in what other Boost authors feel about this.
I have one nasty date_time bug that I've fixed since 1.31 which would be worthy of including in a patch release. If authors only include critical bugfixes, the number of regressions should be minimal and hence the release work should be minimized. OTOH, if there is a bit more discipline with 1.32 (that is, the release process doesn't take 3 months) 1.31.1 would only have a life of a couple months... Jeff