On Tue, Nov 05, 2002 at 03:11:44PM +0100, Pavel Tsekov wrote: > > >On 6 Nov 2002, Robert Collins wrote: > >> On Wed, 2002-11-06 at 01:01, Pavel Tsekov wrote: >> > On 6 Nov 2002, Robert Collins wrote: >> > >> > > Consider this: if the test copy we review was -3, and you approve that >> > > version, you should not rename it to -1, else when -2 comes out, all the >> > > testers will fail to upgrade. So -3 in review, stays as -3 when >> > > uploading. Likewise, updating -1 to -2 to -3 during the review process >> > > allows the testers to update properly. Does that make sense? >> > >> > Ok, but if one releases a new Cygwin package shouldn't it get -1, which >> > will make it first Cygwin release of this package ? And I'm under the >> > impression that xerces is a new package :) >> >> No, it should get *a* number that will then monotonically increase >> within the same vendor version number. > >Ok, I've just checked the package naming conventions to refresh my >memory. Check this: > >http://cygwin.com/setup.html#naming > >Maybe it should be updated.
I don't recall people bumping the -x numbers previously. I don't see why that's necessary. Theoretically, we are a small enough and intelligent enough group that we won't be confused when the contents of a file are updated without updating the file version. I think, in general, a new cygwin package should start out with a -1. It's hardly worth starting a long discussion about but, IMO, the -x numbers are incremented when there is a new cygwin release not when someone is providing packages for review. cgf