I've gone through the two branches with kdiff3 and selectively backported
many of the differences between the branches. There are a number of
experiments and minor cosmetic changes which I didn't backport. One big
remaining hope I had for 6.4.4 was to backport the parallelization work
in the imagery scripts, e.g. i.landsat.rgb. I think they are pretty good
but I haven't tried them on WinGrass to learn how well MSys deals with
backgrounding jobs with "&". I had planned to merge scripts/ first, to
give them the maximum amount of testing in the stable branch, but it turned
out to be too big a job for the small windows of time I could give to it. But now I am mostly waiting on testing them in 6.5 nightly WinGrass.
The other two important tasks for 6.4.4 in my mind are to get python addon
scripts working with 6.4's g.extension.py, both for Linux and WinGrass (#1768,
just committed an experimental change to 6.5svn a few days ago; e.g. for
r.threshold), and secondly to get the Cairo driver working on WinGrass (#943).
The Cairo driver seems to need some G_spawn() pipe handling/closing in d.mon,
using SF_BACKGROUND and other popen flags which I just stumble around in the
dark with & need some assistance.
It's our busy field season right now and I'm just about to head back
out to sea, so I won't be here to work on things for the next few weeks
(... and then the holidays come).
Are there critical bugs fixed in 6.4.4svn since 6.4.3 that demand we ship
a fix for ASAP? If not, I'd like to put some energy into the above issues
first. If there is some critical need to release now, I don't mind putting
these other things off for a 6.4.5, but especially the 'running custom
python scripts' in 6.4 WinGrass seems like an important thing to have
working out of the box for end users.
On Sat, Nov 16, 2013 at 3:00 AM, Hamish <hamish_b@yahoo.com> wrote:
on the grass-user ML Markus N wrote:
(and it is time to get 6.4.4 out...)
fwiw,
I've gone through the two branches with kdiff3 and selectively backported
many of the differences between the branches. There are a number of
experiments and minor cosmetic changes which I didn't backport. One big
remaining hope I had for 6.4.4 was to backport the parallelization work
in the imagery scripts, e.g. i.landsat.rgb. I think they are pretty good
but I haven't tried them on WinGrass to learn how well MSys deals with
backgrounding jobs with "&". I had planned to merge scripts/ first, to
give them the maximum amount of testing in the stable branch, but it turned
out to be too big a job for the small windows of time I could give to it.
But now I am mostly waiting on testing them in 6.5 nightly WinGrass.
I think that this parallelization work should wait for 6.4.5 or
whatever (also given your available time).
The other two important tasks for 6.4.4 in my mind are to get python addon
scripts working with 6.4's g.extension.py, both for Linux and WinGrass (#1768,
just committed an experimental change to 6.5svn a few days ago; e.g. for
r.threshold),
and secondly to get the Cairo driver working on WinGrass (#943).
The Cairo driver seems to need some G_spawn() pipe handling/closing in d.mon,
using SF_BACKGROUND and other popen flags which I just stumble around in the
dark with & need some assistance.
The other two important tasks for 6.4.4 in my mind are to get python
addon
scripts working with 6.4's g.extension.py, both for Linux and WinGrass
(#1768,
just committed an experimental change to 6.5svn a few days ago; e.g. for
r.threshold),