pulling my hair out over Leopard
Hi everyone, I'm trying to get leopard and boost to play nice, so that I can use dicomlib for another app. Apparently, dicom lib requires boost, because it gives 116 error messages if you try to compile in xcode without it. To repro: 1) download dicomlib from http://dicomlib.swri.ca/joomla/index.php?option=com_remository&Itemid=26&func=fileinfo&id=32 2) put those files into an xcode project 3) watch as they don't compile with 116 errors all related to boost I've tried: 1) including the boost header files in the xcode project. xcode doesn't care unless I change all the < to ", which is painful, time consuming, and frankly, stupid. 2) installing boost at /usr/local/include using ./configure make sudo make install as an administrator. Of course, the boost library is handily titled 'boost-1_34_1' instead of something like 'boost', which means that I have to soft link it using ln -s. Neither the regular install nor the ln option gets around the 115 compile errors. 3) I tried bjam -sTOOLS=darwin and bjam -sTOOLS=darwin -sBUILD=framework, neither of which work, both of which give errors that ld -h is an unrecongized option. I don't know where the linker options are defined, so maybe just getting rid of -h would do it, but that switch probably means something, so just wholesale removing a compiler switch may not be the best move. So here's the real question: How do I make this work? I've lost about 24 hours to this, and I just don't see how it's gonna work. In the meantime, I'm guessing I'm gonna go with http://www.apple.com/downloads/macosx/math_science/quickdicomtoolkit.html but knowing my luck, that will require something similar as well.... Thanks for your help, Mark
On Nov 11, 2007, at 10:26 PM, Mark Roden wrote:
Hi everyone,
Hi Mark, [...]
2) installing boost at /usr/local/include using ./configure make sudo make install as an administrator. Of course, the boost library is handily titled 'boost-1_34_1' instead of something like 'boost', which means that I have to soft link it using ln -s. Neither the regular install nor the ln option gets around the 115 compile errors.
I suggest you don't use configure / make and use bjam instead. bjam --layout=system install Cheers, -- Benoit Sigoure aka Tsuna EPITA Research and Development Laboratory
Hi Benoit, Thanks for the tip-- that approach yields ...failed updating 54 targets... ...skipped 12 targets... ...updated 5589 targets... So is it safe to assume that those 54 failed targets are unimportant? Thanks, Mark On 11/11/07, Benoit Sigoure <tsuna@lrde.epita.fr> wrote:
On Nov 11, 2007, at 10:26 PM, Mark Roden wrote:
Hi everyone,
Hi Mark,
[...]
2) installing boost at /usr/local/include using ./configure make sudo make install as an administrator. Of course, the boost library is handily titled 'boost-1_34_1' instead of something like 'boost', which means that I have to soft link it using ln -s. Neither the regular install nor the ln option gets around the 115 compile errors.
I suggest you don't use configure / make and use bjam instead. bjam --layout=system install
Cheers,
-- Benoit Sigoure aka Tsuna EPITA Research and Development Laboratory
_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
On Nov 12, 2007, at 12:17 AM, Mark Roden wrote:
Hi Benoit,
Thanks for the tip-- that approach yields ...failed updating 54 targets... ...skipped 12 targets... ...updated 5589 targets...
So is it safe to assume that those 54 failed targets are unimportant?
I don't know, I guess it depends. What are these 54 targets? Didn't you get the same 54 failures beforehand? The --layout=system thingy is only meant to ask bjam to install the libraries without encoding the toolset name and other information in the name of the libraries (which seemed to be your problem, right?). -- Benoit Sigoure aka Tsuna EPITA Research and Development Laboratory
No, that's actually some 6 more failures than with the previous use of bjam I said in the original message, and the xcode project still doesn't work. Thanks tho. Mark On 11/11/07, Benoit Sigoure <tsuna@lrde.epita.fr> wrote:
On Nov 12, 2007, at 12:17 AM, Mark Roden wrote:
Hi Benoit,
Thanks for the tip-- that approach yields ...failed updating 54 targets... ...skipped 12 targets... ...updated 5589 targets...
So is it safe to assume that those 54 failed targets are unimportant?
I don't know, I guess it depends. What are these 54 targets? Didn't you get the same 54 failures beforehand? The --layout=system thingy is only meant to ask bjam to install the libraries without encoding the toolset name and other information in the name of the libraries (which seemed to be your problem, right?).
-- Benoit Sigoure aka Tsuna EPITA Research and Development Laboratory
_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
Hello, I want to know if there's a "MAXIMUM INDEPENDENT SET" algorithm in the lib. Thanks. Max
Hello Doug， I see. Thanks for your reply. It seems to be known as a NP-C problem to compute the independent sets with minimum number of sets. It's equivalent to another problem/description that coloring a graph with a minimum number of colors. I've already found a graph coloring algorithm in the boost lib, it works quite well, both in efficiency and the fact it's quite close to the optimization result. Thanks. ------------------ Max 2007-11-14 ------------------------------------------------------------- 发件人：Douglas Gregor 发送日期：2007-11-14 10:23:09 收件人：boost-users@lists.boost.org 抄送： 主题：Re: [Boost-users] [BGL]MAXIMUM INDEPENDENT SET On Nov 11, 2007, at 5:11 PM, LoadCom wrote:
Hello,
I want to know if there's a "MAXIMUM INDEPENDENT SET" algorithm in the lib.
We do not have any algorithms that compute maximal independent sets; sorry! - Doug _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
Mark, dicomlib doesn't support Mac OS X, however I was able to make it place nice with boost as all it needs are some simple headers. If you really want to get it to work on OS X then your going to have to hack a little bit on dicomlib/socket as it has no idea what MACH, DARWIN, etc are... But only 16 errors that seem easily portable socket and endian related items. I created an Xcode project and the structure should look like so: dicom.xcodeproj boost_1_34_1 dicomlib dicomlib/socket Download: http://homepage.mac.com/jolix/FileSharing12.html ~Julian Cain On Nov 11, 2007, at 1:26 PM, Mark Roden wrote:
Hi everyone,
I'm trying to get leopard and boost to play nice, so that I can use dicomlib for another app. Apparently, dicom lib requires boost, because it gives 116 error messages if you try to compile in xcode without it.
To repro: 1) download dicomlib from http://dicomlib.swri.ca/joomla/index.php?option=com_remository&Itemid=26&func=fileinfo&id=32 2) put those files into an xcode project 3) watch as they don't compile with 116 errors all related to boost
I've tried: 1) including the boost header files in the xcode project. xcode doesn't care unless I change all the < to ", which is painful, time consuming, and frankly, stupid. 2) installing boost at /usr/local/include using ./configure make sudo make install as an administrator. Of course, the boost library is handily titled 'boost-1_34_1' instead of something like 'boost', which means that I have to soft link it using ln -s. Neither the regular install nor the ln option gets around the 115 compile errors. 3) I tried bjam -sTOOLS=darwin and bjam -sTOOLS=darwin - sBUILD=framework, neither of which work, both of which give errors that ld -h is an unrecongized option. I don't know where the linker options are defined, so maybe just getting rid of -h would do it, but that switch probably means something, so just wholesale removing a compiler switch may not be the best move.
So here's the real question: How do I make this work? I've lost about 24 hours to this, and I just don't see how it's gonna work.
In the meantime, I'm guessing I'm gonna go with http://www.apple.com/downloads/macosx/math_science/quickdicomtoolkit.html
but knowing my luck, that will require something similar as well....
Thanks for your help, Mark
_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
participants (5)
-
Benoit Sigoure -
Douglas Gregor -
Julian C -
LoadCom -
Mark Roden