kicad-developers team mailing list archive

[Jon Evans <jon@xxxxxxxxxxxxx>]

Hi Rigo, here's my thoughts, other devs might have more to add:

> Most significantly, the separate icon set. While everyone is looking
forward to having this small feature, it's still kinda offending the term
feature "freeze". Though we weren't at string freeze, new features are
already annoying enough to translators. Are we getting other features
before V6 launch?

Several features were marked as feature freeze exceptions by the lead
development team, as they were deemed important features that were in
development and were being planned around for V6.  Support for dark color
themes was included here as lack of it makes KiCad pretty unusable in
desktop environments with dark window colors.

There are two other new features that are on the list of exceptions but
aren't in yet:

- Arc support in polygons
- New Python API

> Though we weren't at string freeze, new features are already annoying
enough to translators

The development team does not consider feature freeze to mean string
freeze, which is why we have that separate concept.  I certainly appreciate
the efforts of translators who work to keep up with development, but since
we are not at a string freeze, I think it would also be acceptable if some
translators wanted to just wait for the string freeze so they didn't have
to possibly do work twice.  Of course our use of Weblate now makes this a
little easier.

> Many (or most) critical/high priority issues are unresolved crashes,
usually hard to reproduce, without need-info tag. Are they slowing down the
version iteration?

Some of these may be missing a needed "need-info" tag.  Generally, crash
bugs in KiCad fall into two categories: ones that we can easily reproduce,
which are generally fixed very quickly, and ones that we cannot reproduce.
The latter are still important to fix if we have evidence that they are
caused by a KiCad problem (some hard-to-reproduce crash bugs can be caused
by various other system configuration things that are outside our
control).  Generally, as long as the reporter is able to consistently
reproduce the crash bug in a supported environment, we will hold the
release until all such bugs are fixed, but there may be exceptions.

If there are crash bugs that can't be reproduced by anyone (even outside
the development team) reliably enough to understand what causes them, it
might be a different story.  I think each situation has to be analyzed
independently.

> Will we get more frequent updates like 6.1 6.2 adding small features step
by step instead of doing all the things in one milestone?

Currently the plan is to stick with the major version numbering for new
features (i.e. going straight to 7.0) but to do fewer big features per
major version release.  We would all like to see more frequent releases.

-Jon

On Sat, Mar 20, 2021 at 11:02 AM RigoLigo RLC <rigoligo03@xxxxxxxxx> wrote:

> I have a few questions about the current KiCad development.
>
> 1. new features after feature freeze?
>
> Most significantly, the separate icon set. While everyone is looking
> forward to having this small feature, it's still kinda offending the term
> feature "freeze". Though we weren't at string freeze, new features are
> already annoying enough to translators. Are we getting other features
> before V6 launch?
>
> 2. unresolved mysterious crashes
>
> Many (or most) critical/high priority issues are unresolved crashes,
> usually hard to reproduce, without need-info tag. Are they slowing down the
> version iteration? I mean, KiCad crashes are not new, not to mention those
> that are hard to tackle. Reducing data loss with autosave and such would be
> enough for current status of KiCad IMO. Of course crashes that have known
> reasons must be fixed. But what are we going to do with these mysterious
> crash issues? Wait till they're resolved before we launch RC1?
>
> 3. Development iteration...
>
> Currently KiCad will push big version leaps for all the features
> developers add into it. I know it's got reasons (difficulty of maintaining
> two branches, plugin API consistency) but a big leap requires a LONG wait.
> For OSS users this might not be a big problem but in Chinese community
> there are companies that use KiCad as alternatives. They dare not to use
> nightlies(even it's stable, it's not a stable release, and files are
> incompatible), but are still eager to have nice new features. Will we get
> more frequent updates like 6.1 6.2 adding small features step by step
> instead of doing all the things in one milestone?
>
> I'd like to know from you guys.
>
> Rigo
> _______________________________________________
> Mailing list: https://launchpad.net/~kicad-developers
> Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~kicad-developers
> More help   : https://help.launchpad.net/ListHelp
>