https://pantsbuild.org/ logo
#announce
Title
# announce
h

hundreds-father-404

10/26/2020, 6:12 PM
Hey, what’s the thinking on 2.0.1 vs. 2.1.0? a) Are we going to still do 3-4 dev releases for 2.1.x? We would have normally done 2 releases already. b) Do we plan to treat 2.0.1 differently than normal? For most releases, we don’t end up having a patch release.
w

witty-crayon-22786

10/26/2020, 6:36 PM
i think that we should get back to time based releases.
👍 1
h

hundreds-father-404

10/26/2020, 6:36 PM
maybe only dev0 and dev1?
w

witty-crayon-22786

10/26/2020, 6:36 PM
but this is an opportunity to decide whether “monthly stable branch” is the right fit, or whether we want to do 6 weeks or something.
👍 1
h

hundreds-father-404

10/26/2020, 6:37 PM
+1 to deciding. In the 1.x series, I thought our releases were already too big. Too much barrier for users to upgrade versions
w

witty-crayon-22786

10/26/2020, 6:39 PM
hm
@hundreds-father-404: i think that i observed the opposite for most end users… people were frequently 10 or more versions behind, because they weren’t updating monthly
it’s possible that we don’t need to make a change now though, and could just stick to 1 month as the convention until we hear otherwise.
h

hundreds-father-404

10/26/2020, 6:41 PM
Agreed on not upgrading, but do we know why that was?I was under the impression because it was too much barrier to upgrade. Alternatively, the issue might have been not enough carrots in each release?
w

witty-crayon-22786

10/26/2020, 6:41 PM
@hundreds-father-404: perhaps you are primarily thinking of the lead up to 2.0, when things were changing very very rapidly? i think that post 2.0, we might expect things to be more stable, and to not have the same volume of deprecations per release
👍 1
Agreed on not upgrading, but do we know why that was?I was under the impression because it was too much barrier to upgrade.
if you go 10 months without upgrading, i think the size of each upgrade is less relevant… the total volume is the same
h

hundreds-father-404

10/26/2020, 6:42 PM
Yeah, perhaps. I’m thinking about how hard it was to go from 1.24 to 1.25, for example, when we did several things like deprecating
globs=()
and iirc deprecating the default for
--no-chroot --fast
at the same time
I think you’re right that it’ll mostly be new features + plugin API changes + performance improvements (e.g. pantsd). Not as many User API changes
h

happy-kitchen-89482

10/26/2020, 7:40 PM
I think going back to ~monthly minor releases (so starting 2.1.0 soon) makes sense,.