[GRASS-dev] Re: [GRASS GIS] #843: v.digit broken on new WinGrass
release
GRASS GIS
trac at osgeo.org
Wed Dec 23 17:08:53 EST 2009
#843: v.digit broken on new WinGrass release
---------------------------+------------------------------------------------
Reporter: cnielsen | Owner: grass-dev at lists.osgeo.org
Type: defect | Status: new
Priority: critical | Milestone: 6.4.0
Component: Vector | Version: unspecified
Resolution: | Keywords: wingrass,v.digit
Platform: MSWindows XP | Cpu: x86-64
---------------------------+------------------------------------------------
Comment (by glynn):
Replying to [comment:9 marisn]:
> > Huh? That matches the error message, but v.digit shouldn't be using
lib/form. It has its '''own embedded copy''' of the form library.
> "Own embedded copy" is a good reason for trouble. Before r38954 form lib
was out of sync. If lib/form was broken before r38954 then something else
was broken too (nviz? d.what?).
By "embedded", I mean "in-process". The form library in lib/form consists
of a custom "wish" which has a command for executing SQL, and a library
which spawns this as a slave process. The version in v.digit doesn't use a
separate child process; the form code has been merged into the main
v.digit program (which is also a custom wish).
The separate form library is still used by nviz and d.what.vect.
> Unfortunately I have no idea about sockets/pipes on windows. According
to lib/form history, Glynn should have an idea how sockets/pipes/etc.
works on Windows.
I don't have much of an idea either. I do know that Windows doesn't
understand Unix-domain sockets (so no socketpair()). It does have a
_pipe() function which is roughly equivalent to the Unix version, but the
lack of fork() means that creating a slave process is nothing like on
Unix.
--
Ticket URL: <https://trac.osgeo.org/grass/ticket/843#comment:10>
GRASS GIS <http://grass.osgeo.org>
More information about the grass-dev
mailing list