[QGIS-Developer] how compatible .qml files are ?

[Nyall Dawson]

On Sat, 31 Aug 2019 at 17:02, Sandro Santilli <strk at kbt.io> wrote:
>
> On Thu, Aug 29, 2019 at 08:35:09PM +1000, Nyall Dawson wrote:
> > On Thu, 29 Aug 2019 at 20:29, Sandro Santilli <strk at kbt.io> wrote:
> > > On Thu, Aug 29, 2019 at 08:05:04PM +1000, Nyall Dawson wrote:
> > > >
> > > > It was the right move to drop this unused legacy code rather than
> > > > waste the valuable maintainer power on something completely unused.
> > >
> > > It could have been moved to a convertion tool.
> >
> > There was no volunteers to do this, and no sponsorship for work of
> > this nature => no conversion tool. ;)
>
> I'd think whoever introduced the _new_ format should have taken
> care of providing an upgrade path for those with an old format.

Not necessarily -- this was a wide group decision, with consensus from
all currently active QGIS developers and maintainers. It was deemed
unnecessary at this point in time. I strongly suspect that there's a
cultural or language effect in play here, but it's hard not to
(mis)interpret this thread as disrespecting the decisions made by the
active QGIS developers and community of the time. Just like I wouldn't
dream of jumping onto the GDAL or Qt lists and questioning design
decisions made by people with intimate knowledge of the code base and
wider context and constraints of discussions, I'd ask you to extend
the same respect and courtesy to our community and the individuals who
have made these informed decisions. These decisions were made,
accepted, and implemented -- let's not argue them retrospectively, but
accept and move on instead.

It's GREAT that you're seeking to address what you see as a feature
gap/bug in QGIS, but I just want to caution that your language is
likely mis-representing your intentions here. (Also I hope that my
tone won't be misinterpreted here either! Cross-cultural community
building is HAAAARD :O )

>
> Do you know which commit dropped the backward compatibility, exactly ?
> Note that "something completely unused" (just re-reading your quoted
> text) was actually still used by QGIS itself...

No, no idea sorry. It would have been 2-3 years ago now, and it was a
mass deletion commit, so git blame won't be much use. You may be able
to git blame qgsvectorlayer.cpp or qgsvectorlayerrenderer.cpp to find.

> It would sound like QgsProjectFileTransform::updateRevision should throw
> an exception when unable to do what's been asked for ?

In this case QgsProjectFileTransform will be a red herring - there was
never a transform in place for this code, so there won't be any clues
there.

Nyall

>
> --strk;