(tried sending this last week but it was blocked for being 77 bytes
larger than the limit & it appears there are no list moderators around
to approve or decline messages so it's stuck in limbo)
The code on radsite.lbl.gov not being updated isn't much of an issue,
I'd just prefer to use that since it's a leaner package as it comes
without the CSAIL objects & other extras you get in the radiance-online
packages. It is the "official" release which I find rather surprising
that it was released without it ever being tested on Windows (3.9 built
fine & had the same install issues as 4.0 so the build breaking changes
were introduced between the 3.9 & 4.0 release). Also a lot of people are
wary about using a in-development version rather than an official
release but I guess with something like Radiance which isn't used by
Windows users in general since it needs to be built by the end user it's
not that big an issue.
The 2nd issue is that the files are listed for install in multiple
Sconscript files. They need to be in only one otherwise scons gets
confused when trying to copy the same file from 2 sources to 1
destination. The lines needing changed are
in ray/src/gen/Sconscript
from
LIBFILES = Split('illum.cal rev.cal skybright.cal surf.cal glaze1.cal
glaze2.cal clockface.hex')
to
LIBFILES = Split('illum.cal rev.cal skybright.cal glaze1.cal
glaze2.cal')
in ray/src/cv/Sconscript
from
LIBFILES = Split('source.cal tilt.cal lamp.tab window.cal')
to
LIBFILES = Split('source.cal tilt.cal window.cal')
The 3rd issue is probably just those files not being listed for install
in a Sconscript file. mkillum appears to be built by
ray/src/gen/Sconscript but it's commented out in the progs list & then
manually appended for the build. It's not manually appended for the
install though. Simple fix would be to move the manual append to the
progs list. mfg2meta seems to be the same situation in
ray/src/cv/Sconscript
4th issue is the one I'm most concerned with, everything else has a
workaround but a huge number of missing binaries is something that needs
looking in to. The code is there I guess it's just never been added to
the Scons build system for Windows builds & without spending a lot of
time learning the Radiance code on a Mac, I wouldn't know where to
start. The ones which are definitely not C-Shell Scripts or X11 window
apps are:
commake
dctimestep
dmake
evalglare
gendaylit
genklemsamp
genrhgrid
genskyvec
normtiff
ra_tiff
rmake
rvcalc
rvu
So around 20 was an overestimate & there could be some X11 window apps
in that list too but there are some major omissions like ra_tiff as
mentioned before.
Cheers,
<http://www.iesve.com/>
Palbinder Sandher
Software Deployment &
IT Administrator
T: +44 (0) 141 945 8500
F: +44 (0) 141 945 8501
http://www.iesve.com <http://www.iesve.com/>
**Design, Simulate + Innovate with the <Virtual Environment>**
Integrated Environmental Solutions Limited. Registered in Scotland No.
SC151456
Registered Office - Helix Building, West Of Scotland Science Park,
Glasgow G20 0SP
Email Disclaimer <http://www.iesve.com/disclaimer.html>
···
________________________________
From: [email protected]
[mailto:[email protected]] On Behalf Of Gregory
J. Ward
Sent: 29 June 2010 19:52
To: code development
Subject: Re: [Radiance-dev] scons
Thanks, Pally.
I don't know what to do about most of this, as I don't understand the
SCONS build system Schorsch developed. I did remove the outdated links
to the zip archives from radiance-online.org. I would have liked to
update these, but I don't know where they are supposed to get built or
why they are not.
Regarding the problems with Radiance4R0.tar.gz, this file does not have
the since-release changes to the HEAD, but the idea always is to drop
the HEAD on top of whatever you have and update it that way, so I see
nothing to do there. I have copied the 4R0 sources and libraries over
to radiance-online, but I'm afraid it's a bit of a mess right now. The
scripts aren't computing the correct patches, and a number of other
things are broken there as no one is maintaining that website at the
moment. (The usual resource allocation issues.)
On your 2nd issue below, I am looking at Sconscript and I see
clockface.hex and surf.cal mentioned only once, so maybe this has
already been fixed? As I said, I don't really understand this build
system. If you can be more specific about which lines you had to
change, that would be helpful.
I don't know what to say or do about your 3rd issue, as there could be
any number of things going wrong, and no way for me to tell from here.
Best,
-Greg