|
From: | John W. Eaton |
Subject: | Re: 4.4 Release Progress |
Date: | Sat, 7 Apr 2018 19:12:20 -0400 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.6.0 |
On 04/07/2018 05:12 PM, Mike Miller wrote:
On Sat, Apr 07, 2018 at 15:37:21 -0400, John W. Eaton wrote:Yes. I already added a 4.3.90 version number in the bug tracker.So I take it you like the idea of using the x.y.9x numbers for alpha/beta/rc/whatever-you-want-to-call-it?
We've used that before, and I prefer it to appending -rcN to the version number.
I was thinking about this again, and it doesn't quite follow your ideal of the version number being completely unambiguous. With Octave 5.0.0, if I remember right, you are planning on tagging 5.1.0 as a release, then immediately incrementing the version to 5.1.1 for stable bug fixes. So there will be exactly one revision in the hg history that has a version number of 5.1.0.
I hadn't really thought about it, but I guess it makes sense that there would be just one revision in the hg archive corresponding to actual releases. Then X.Y.1 is the number for subsequent development of the X.Y.0 release.
Do you think it's important to do the same for release candidates? IOW, should there be exactly one revision in the hg history when the version number is 4.3.90?
I don't know that it matters as much. But if we did that, then each release candidate would be an even number? 4.3.9{0,2,4,...} are release candidates and 4.3.9{1,3,5,...} would be used for and changes in between? Hmm. I don't really know what is best.
jwe
[Prev in Thread] | Current Thread | [Next in Thread] |