[profanity] amount of releases per year

Michael Vetter jubalh at iodoru.org
Thu May 14 11:16:37 CEST 2020


> i wonder how many releases we should have per year? like openbsd does
> well with 2 releases per year.

I'm aiming at a major release every +/- 6 months.
https://github.com/profanity-im/profanity/releases shows that so far
this works well. Often being even quicker: 5 months.

> At least, i feel it would make sense to make a release soon, as quite
> a few user-visible things have happened since february:

We use milestones to manage what we want in a new release and when it
is finished: https://github.com/profanity-im/profanity/milestones

So you can see that for 0.9.0 currently 13 tasks are still open:

This is only a rough estimate, as I often work under different things
under the hood and don't track it in GH.
Also if I feel a release is taking too long I delay some issues to the
next release and craft a new release now.

> That means, one release roughly every 3-4 months or so is good, i
> think.

With few active developers a release every 3 months will be hard. Also
there are only 2 people supporting Profanitys development with money
(but many making requests) which means I can only spend few time on it.

We have some brave users running Profanity from git (thanks guys!) and
giving them the opportunity to run latest git for some weeks is also
essential in finding the bugs early on since they don't use all
features every single day.

The 6 months release is only a guideline (that I aim for), but in the
end we have a similar approach to Debian: it's ready when it's ready :-)

Other info that might be of interest:
- I was hoping to have experimental MAM support in 0.9.0 but not sure
  whehter it will happen. Some rough things work alreday since a couple
  of weeks though, but it's not finished yet.

- 0.9.0 is planned to have some nice new features (like the ones you
  mentioned in your mail plus the ones still tracked in the milestone)
  and after it is released I hope that pasis and I can port libmesode
  things to libstrophe and then finally deprecating libstrophe. I also
  hope that pasis can wait to release a new libstrophe until this is
  done so we dont have to do backporting work for libmesode.

- If this happens then the next release (either 0.10.0 or 1.0 depending
  on features and stability) is going bump some library version, like
  requiring the latest libstrophe (because we will switch to some
  functions that will only be in the new release), and newer glib. So
  0.9.0 will run still nice on older distros but the release after will
  probably need a newer distro.

- We will also remove some old Profanity code that is used to transform
  some very old configs for that new release.

Hope this helps,

More information about the profanity mailing list