[GRASS-dev] v.generalize: flag -r removed?

Hi all.
Apparently the flag -r has been recently removed. May I know the reason
in a few lines? I have to adjust the QGIS the plugin, and I prefer to
understand before doing.
Thanks.
--
Paolo Cavallini: http://www.faunalia.it/pc

On Sat, Sep 24, 2011 at 3:21 PM, Paolo Cavallini <cavallini@faunalia.it> wrote:

Hi all.
Apparently the flag -r has been recently removed. May I know the reason
in a few lines? I have to adjust the QGIS the plugin, and I prefer to
understand before doing.

Until recently, v.generalize did not work with boundaries, more
generally, did not fully adhere to the grass vector topology model.
The -r flag was one of the components causing damage, further on it
duplicated functionality covered be v.clean and made the interface
even more complex. Therefore it was removed while overhauling the
module.

Markus M

Il giorno sab, 24/09/2011 alle 15.48 +0200, Markus Metz ha scritto:

Until recently, v.generalize did not work with boundaries, more
generally, did not fully adhere to the grass vector topology model.
The -r flag was one of the components causing damage, further on it
duplicated functionality covered be v.clean and made the interface
even more complex. Therefore it was removed while overhauling the
module.

Thanks a lot Markus for the clarification.
Fixed in QGIS plugin.
All the best.
--
Paolo Cavallini: http://www.faunalia.it/pc

Paolo wrote:

> Apparently the flag -r has been recently removed. May I know
> the reason in a few lines? I have to adjust the QGIS the
> plugin, and I prefer to understand before doing.

Markus M:

Until recently, v.generalize did not work with boundaries,
more generally, did not fully adhere to the grass vector
topology model.
The -r flag was one of the components causing damage, further
on it duplicated functionality covered be v.clean and made the
interface even more complex. Therefore it was removed while
overhauling the module.

since this flag/feature was released in an official stable
release it should not have been removed for future versions of
the stable branch. If the feature was in fact too broken to be
allowed to live, it should be replaced with a no-op -r flag with
a description of

    flag->description =
        _("This does nothing. It is retained for backwards compatibility");
    flag->guisection = _("Unused");

any user script, addon module, documentation written since 6.0.0
should continue to work with any version of grass 6 >= than the
one it was written for, with at most needing a no-modification
recompile. for years we have shown good discipline in this regard;
doing so respects the users' confidence to invest their energy
(once) and get on with their work, and makes grass a more
dependable platform to base their infrastructure/work flow on.

this is a blocker for 6.4.2.

thanks,
Hamish

On Sat, Sep 24, 2011 at 6:38 PM, Hamish <hamish_b@yahoo.com> wrote:

Paolo wrote:

> Apparently the flag -r has been recently removed. May I know
> the reason in a few lines? I have to adjust the QGIS the
> plugin, and I prefer to understand before doing.

Markus M:

Until recently, v.generalize did not work with boundaries,
more generally, did not fully adhere to the grass vector
topology model.
The -r flag was one of the components causing damage, further
on it duplicated functionality covered be v.clean and made the
interface even more complex. Therefore it was removed while
overhauling the module.

since this flag/feature was released in an official stable
release it should not have been removed for future versions of
the stable branch. If the feature was in fact too broken to be
allowed to live, it should be replaced with a no-op -r flag with
a description of

flag->description =
_("This does nothing. It is retained for backwards compatibility");
flag->guisection = _("Unused");

any user script, addon module, documentation written since 6.0.0
should continue to work with any version of grass 6 >= than the
one it was written for, with at most needing a no-modification
recompile. for years we have shown good discipline in this regard;
doing so respects the users' confidence to invest their energy
(once) and get on with their work, and makes grass a more
dependable platform to base their infrastructure/work flow on.

this is a blocker for 6.4.2.

The -r flag is restored in r48471,2 as suggested. IMHO, v.generalize
should not have been included in its original form in the
releasebranch in the first place.

Markus M

thanks,
Hamish