One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
Randolph
···
On Dec 15, 2008, at 12:56 PM, Axel Jacobs wrote:
Hi guys,
(Greg:)
As for the Linux README regarding libtiff, I'm all for it, but not
confident enough to write it myself, having never done it. I also
don't understand packages under Linux or Ubuntu, so I'd really prefer
that someone else submit a README to put into the distribution. I'd
be happy to proofread it and include it. I don't think there are any
serious problems remaining between the packaged libtiff and the one I
distribute, but again I don't have any Linux experience to draw on.
(Bernd:)
I could write that, but first I'll have to update my patches to the latest
HEAD and see what needs to be done at the moment. Last thing I remember is
that I talked with Greg about some tiff related patches from me, and I think
they're applied in his tree. Didn't catch up with that as I'm spending more
spare time on Lenny related stuff at the moment.
Brilliant! You guys rock!
Bernd, I know I half dumped you into this by even begging the
question. Thanks a lot for picking this up! I promise to put on my
Sherlock Holmes hat and give some good-quality dummy's level feedback.
(Thomas:)
Francesco has already written down what's necessary to
compile on OS X:
http://www.bozzograo.net/radiancewiki/doku.php
I haven't compiled Radiance myself in ages because
Greg always provides the binaries (probably first of all):
http://radsite.lbl.gov/radiance/dist/rad3R9_macosx.tar.gz
There are minor issues on OS X but they have been
covered a lot and if your student couldn't get it
to work he/she shouldn't use the source distribution
in the first place or search the in list archives
for the same problem.
Thanks for pointing this out. I was aware of Francesco's site, but did
not know the install instructions (sorry, Francesco).
Greg, do you think this would justify a link to Francesco's page on
the two download pages (radsite and radiance-online)? All the coding
gurus seem to run MacOS now. However, many Mac people are still of the
more artistic type. So even something like
'/your/preferred/path/to/the/binaries' might be too much. So where DO
I stick them? Come on... we all started off small.
Oh, and yes, I'm deliberately being a bit stubborn here.
Many thanks for you efforts and feedback
Cheerioh
Axel
_______________________________________________
Radiance-dev mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-dev
Hi Randolph,
Do you think there's a way to do this with Makefile's? One way I can think to do it is to modify the tiff target I used to have into separate targets for the local or system library. What technique did you use to detect the presence of a libtiff with LogLuv support?
-Greg
···
From: R Fritz <[email protected]>
Date: December 16, 2008 6:22:34 PM PST
One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
Randolph
Hi,
Do you think there's a way to do this with Makefile's? One way I can
think to do it is to modify the tiff target I used to have into separate
targets for the local or system library. What technique did you use to
detect the presence of a libtiff with LogLuv support?
the only ways which come into my mind to detect them automatically are not
platform independent and would require something like Gnu-Make. Even ifdef and
similar things don't work in all Make variants...
Cheers,
Bernd
···
--
Bernd Zeimetz Debian GNU/Linux Developer
GPG Fingerprint: 06C8 C9A2 EAAD E37E 5B2C BE93 067A AD04 C93B FF79
Could the test could be done in the makeall script, and the result used in Makefiles?
Randolph
···
On Dec 16, 2008, at 7:45 PM, Gregory J. Ward wrote:
Hi Randolph,
Do you think there's a way to do this with Makefile's? One way I can think to do it is to modify the tiff target I used to have into separate targets for the local or system library. What technique did you use to detect the presence of a libtiff with LogLuv support?
-Greg
From: R Fritz <[email protected]>
Date: December 16, 2008 6:22:34 PM PST
One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
Randolph
_______________________________________________
Radiance-dev mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-dev
That's kind of what I was thinking. Create different targets and stuff the appropriate one in the rmake script via makeall. Still need a reliable way to determine (on any distro) the location and contents of the tiff library. If I can find the location, I can probably use the nm command to look up the symbols and check that they're defined.
-Greg
···
From: R Fritz <[email protected]>
Date: December 17, 2008 3:44:04 PM PST
Could the test could be done in the makeall script, and the result used in Makefiles?
Randolph
On Dec 16, 2008, at 7:45 PM, Gregory J. Ward wrote:
Hi Randolph,
Do you think there's a way to do this with Makefile's? One way I can think to do it is to modify the tiff target I used to have into separate targets for the local or system library. What technique did you use to detect the presence of a libtiff with LogLuv support?
-Greg
From: R Fritz <[email protected]>
Date: December 16, 2008 6:22:34 PM PST
One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
Randolph
(Sorry it's taken so long to get back to you on this)
Looks to me like you could just compile and link a small test program which references libtiff and TIFFInitSGILog()--if it builds, libtiff with logluv support is in a default library directory. I think that's sufficient. If you want something more specific, I could work out a script that would identify the system's type and queries the package manager for specific file locations, but it seems like overkill to me.
Randolph
Gregory J. Ward wrote:
···
That's kind of what I was thinking. Create different targets and stuff the appropriate one in the rmake script via makeall. Still need a reliable way to determine (on any distro) the location and contents of the tiff library. If I can find the location, I can probably use the nm command to look up the symbols and check that they're defined.
-Greg
From: R Fritz <[email protected]>
Date: December 17, 2008 3:44:04 PM PST
Could the test could be done in the makeall script, and the result used in Makefiles?
Randolph
On Dec 16, 2008, at 7:45 PM, Gregory J. Ward wrote:
Hi Randolph,
Do you think there's a way to do this with Makefile's? One way I can think to do it is to modify the tiff target I used to have into separate targets for the local or system library. What technique did you use to detect the presence of a libtiff with LogLuv support?
-Greg
From: R Fritz <[email protected]>
Date: December 16, 2008 6:22:34 PM PST
One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
Randolph
_______________________________________________
Radiance-dev mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-dev
Randolph Fritz wrote:
(Sorry it's taken so long to get back to you on this)
Looks to me like you could just compile and link a small test program
which references libtiff and TIFFInitSGILog()--if it builds, libtiff
with logluv support is in a default library directory. I think that's
sufficient.
That exactly what automake and co do... and imho that's sufficient.
···
--
Bernd Zeimetz Debian GNU/Linux Developer
GPG Fingerprint: 06C8 C9A2 EAAD E37E 5B2C BE93 067A AD04 C93B FF79
R Fritz wrote:
One of the things I tried in my experimental Radiance build scripts was testing for the presence of a system libtiff with logluv support and using it. Is that something that's planned for 4.0?
The SCons build system does that now. If it finds a compatible
libtiff on the standard paths (or on a path defined in the
platform config file) it will use that, otherwise it builds the
included one as before. I've only tested it on linux, but it
should work on any platform, including Windows/cygwin/etc.
-schorsch
···
--
Georg Mischler -- simulations developer -- schorsch at schorsch com
+schorsch.com+ -- lighting design tools -- http://www.schorsch.com/
Hi Greg,
one main problem I still see for people who'd like to use the system's libtiff
are the tiff headers shipped in src/common:
0 bzed@think:~/workspace/debian/recluse/pkg-radiance/src/common(upstream-cvs)$
ll tiff*
-r--r--r-- 1 bzed bzed 30663 2008-08-28 00:57 tiff.h
-r--r--r-- 1 bzed bzed 19161 2008-08-28 00:57 tiffio.h
-r--r--r-- 1 bzed bzed 410 2008-08-28 00:57 tiffvers.h
0 bzed@think:~/workspace/debian/recluse/pkg-radiance/src/common(upstream-cvs)$
I think it makes more than sense to use the tiff headers shipped with the
distribution's libtiff package instead of the old versions included in the
Radiance source. Could they be dropped from src/common and only be copied in
there in case the libtiff from the Radiance source is being built?
The other option would be to change the Makefile which does the libtiff build
job to let libtiff install the tiff headers into a special subdirectory, so
this directory could be added to the CPPFLAGS with -I per default, but it
would be only used if there's something in it.
Cheers,
Bernd
···
--
Bernd Zeimetz Debian GNU/Linux Developer
GPG Fingerprint: 06C8 C9A2 EAAD E37E 5B2C BE93 067A AD04 C93B FF79