[Please don't Cc me on nmh-workers posts, BTW. I don't need 2 copies.]
Scott Blachowicz <[EMAIL PROTECTED]> writes:
> > > so bouncing between even & odd release numbers over time seems weird.
> >
> > There's no bouncing between them. People with FTP access get the
> > even-numbered releases. People with CVS access can get the odd-numbered
> > developer versions in between as well.
>
> I guess I was thinking more of the Linux releases...don't they do
> something like have 1.1.x be various "development" versions with 1.0.x
> being "release" versions? Or something like that?
Oh -- I'm sorry. Maybe the Linux kernel versioning system is not a good
comparison to make -- I only meant to take their idea of "odd" being
developer versions and "even" being release versions, not anything more than
that.
> I could see cases where you might want to have multiple "development" version
> numbers to mark some sort of major change without releasing it (ie.
> 1.0.3 and 1.0.4 could both be internal version numbers).
Not too likely, but feasible, yeah. I guess to keep with the "odd number at
the end" thing I'd have to suggest 1.0.3.1 (or 1.0.3a) instead of 1.0.4 in
that case. Kind of awkward, I realize, but I think that'd be a pretty rare
case.
Different-numbered beta releases (e.g. nmh-0.28pre1, pre2, etc.) make sense
when the beta releases are being doled out a tarball at a time by a master
developer, but when the development source is available continuously via
CVS, I don't see too much reason for it.
-----------------------------------------------------------------------
Dan Harkless | To prevent SPAM contamination, please
[EMAIL PROTECTED] | do not post this private email address
SpeedGate Communications, Inc. | to the USENET or WWW. Thank you.
