<div dir="ltr"><div>So, IIUC, Ryan should have used </div><div><br></div><div> 5.91.0.1</div><div><br></div><div>as the version number on the release branch?</div><div><br></div><div>Robby</div><div><br></div></div><div class="gmail_extra">
<br><br><div class="gmail_quote">On Tue, Nov 26, 2013 at 2:38 PM, Eli Barzilay <span dir="ltr"><<a href="mailto:eli@barzilay.org" target="_blank">eli@barzilay.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="HOEnZb"><div class="h5">Yesterday, Matthew Flatt wrote:<br>
> At Mon, 25 Nov 2013 09:56:45 -0500, Ryan Culpepper wrote:<br>
> > On 11/25/2013 09:44 AM, Matthew Flatt wrote:<br>
> > > Here's the full comment:<br>
> > ><br>
> > > The version string has one of the forms:<br>
> > > X.Y<br>
> > > X.Y.Z Z != 0<br>
> > > X.Y.Z.W W != 0<br>
> > > where each X, Y, Z, W is a non-negative exact integer, Y must not<br>
> > > exceed 99, and Z or W must not exceed 999. Y>=90 means that this is<br>
> > > working towards {X+1}.0, and X.Y (Z=0, W=0) is an alpha version for<br>
> > > {X+1}.0; Z>=900 means working towards X.{Y+1}, and X.Y.Z as an<br>
> > > alpha release.<br>
> > ><br>
> > > Then intent is that when Z and W are 0, the string form of the version<br>
> > > number is just X.Y, not X.Y.Z.W.<br>
> > ><br>
> > > How about this clarification?<br>
> > ><br>
> > > ... and X.Y (i.e., Z=0 and W=0, so Z and W are<br>
> > > omitted from the string form) ...<br>
> ><br>
> > That's not the part that needs clarifying. I think that fact that the<br>
> > string form drops final zeros is clear from lines 2-4.<br>
> ><br>
> > The part that needs clarifying is how to choose the version number for<br>
> > the alpha releases leading up to version {X+1}.0. (Really, how to choose<br>
> > alpha version numbers in general, since I've had similar problems in the<br>
> > past.) From this statement, "X.Y (Z=0, W=0) is an alpha version for<br>
> > {X+1}.0" (Y>=90 already stated), I would expect that 5.91 would be a<br>
> > fine alpha version number for 6.0. Is it? If not, what should the alpha<br>
> > version number be?<br>
><br>
> I agree that "5.91" is the right alpha-version string, assuming that<br>
> it's intended as an "alpha" in the sense of our release rules (as<br>
> opposed to a "release candidate", which has a non-zero W).<br>
<br>
</div></div>These two things were originally independent: the ability to specify<br>
alpha-ness (the second .91) and release-ness (W=0), and that was used<br>
by the old build script to make some decisions for what the installers<br>
do. Assuming that this still matters, there is a problem with using<br>
"5.91" for the release process -- and instead it should be "5.91.0.1"<br>
to make it treated as a nightly build. For example, on Windows the<br>
installer for a "5.91.0.1" wouldn't grab the suffix registration, but<br>
"5.91" would which makes it bad as something that you ask people to<br>
try.<br>
<br>
BTW, this is not the same meaning of "alpha" that is used in the<br>
release checklist -- that one has the meaning of a "release<br>
candidate".<br>
<span class="HOEnZb"><font color="#888888"><br>
--<br>
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:<br>
<a href="http://barzilay.org/" target="_blank">http://barzilay.org/</a> Maze is Life!<br>
</font></span><div class="HOEnZb"><div class="h5">_________________________<br>
Racket Developers list:<br>
<a href="http://lists.racket-lang.org/dev" target="_blank">http://lists.racket-lang.org/dev</a><br>
</div></div></blockquote></div><br></div>