Bug List: (This bug is not in your last search results)   Show last search results      Search page      Enter new bug
Bug#: 217873
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: Steve Arnold <nerdboy@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: Martin von Gagern <Martin.vGagern@gmx.net>
Add CC:
CC:
Remove selected CCs
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
grass-6.2.3-html-nonviz.patch grass-6.2.3-html-nonviz.patch patch Martin von Gagern 2008-04-16 01:21 0000 505 bytes Details | Diff
grass-6.2.3-html-nviz-fix.patch grass-6.2.3-html-nviz-fix.patch patch Martin von Gagern 2008-04-16 01:22 0000 287 bytes Details | Diff
grass-6.2.3.ebuild 6.2.3 ebuild take 1 text/plain Martin von Gagern 2008-04-16 01:39 0000 6.91 KB Details
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 217873 depends on: Show dependency tree
Bug 217873 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)


Not eligible to see or edit group visibility for this bug.






View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2008-04-15 21:51 0000
There are several new versions of GRASS, the latest stable release being 6.2.3,
but also several release candidates for 6.3.0. I'd like to see an ebuild at
least for the first, preferably for the latter as well.

I tried to build the current grass 6.2.1-r1 from portage, but couldn't because
the ebuild complains about tcl/tk built with USE=threads, which I'm not willing
to sacrifice. So I investigated the matter. Seems like that part of the ebuild
originated in the grass-6.1.0.ebuild from October 2006 by nerdboy and was
carried along ever since. I'm not sure how this relates to the following
message from the upstream ChangeLog:

2006-03-16 23:41  cedric

        * visualization/nviz/src/: draw.c, exag.c, lights.c, map_obj.c,
          position.c, togl_flythrough.c, volume.c:
          NVIZ runs with TCL 8.4 Threaded: Memory allocated with Tcl_Alloc,
          such as the result of Tcl_Merge should be freed using
          TCL_DYNAMIC.

To me that sounds like threaded tcl should be supported.
The corresponding changes http://trac.osgeo.org/grass/changeset/19285 were
already included in the 6.1.0 release. Any suggestions where the information
about this incompatibility came from and whether it is still a problem?

I'll try adapting the grass-6.2.1.ebuild to 6.2.3 and see if I can get it to
run. But as I have never before worked with GRASS, I might be ill suited to
evaluate whether the result works as expected.

------- Comment #1 From Martin von Gagern 2008-04-16 01:21:13 0000 -------
Created an attachment (id=149880) [details]
grass-6.2.3-html-nonviz.patch

Derived from grass-6.2.1-html-nonviz.patch, this patch applies more easily, as
it contains no context around the lines to be removed. An alternative would be
to grep for all lines not containing nviz. An updated patch requires less
modifications to the ebuild, though.

------- Comment #2 From Martin von Gagern 2008-04-16 01:22:37 0000 -------
Created an attachment (id=149882) [details]
grass-6.2.3-html-nviz-fix.patch

Derived from grass-6.2.1-html-nviz-fix.patch, this patch applies more easily,
as the context should stay unchanged for longer than that of the previous patch
did. Also the patch is simpler and less in style with the rest of that script.

------- Comment #3 From Martin von Gagern 2008-04-16 01:39:47 0000 -------
Created an attachment (id=149885) [details]
6.2.3 ebuild take 1

Changes so far:
 * moved checks from src_unpack to pkg_setup
 * disabled checks for non-threaded tcl/tk
 * disables some patches and sed scripts that seem unneccessary
 * disabled elibtoolize as it seems to work without
 * moved patch application from src_compile to src_unpack
 * using shipped start script instead of ebuild-generated one

So far for today, I'll continue when I find time again, in the meantime this
should be a good starting point, as it does compile and install here. After
installation I hit http://intevation.de/rt/webrt?serial_num=5328 and decided
I'll have to learn some GRASS first.

Things yet to do in the ebuild:
 * remove commented out portions from older ebuild
 * try to get rid of env file, to avoid clobbering the environment
   Can anybody provide reasons for the different environment settings therein?
 * consider installing to /opt instead of /usr, but read FHS first

------- Comment #4 From Steve Arnold 2008-05-01 08:06:13 0000 -------
Added to the tree, modified somewhat...

Bug List: (This bug is not in your last search results)   Show last search results      Search page      Enter new bug