Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hi, 2009/9/1 Soeren Gebbert soerengebb...@googlemail.com Hello Hamish, 2009/9/1 Hamish hamis...@yahoo.com Hi, I get two errors compiling 6.5svn. The first is because lib/gmath G_math_solvps() fails to return anything when it should return int. I think I fixed it with r38939, but please check to see if that is correct. https://trac.osgeo.org/grass/changeset/38939 Thanks, thats absolutely correct. The second is in lib/external/ccmath, eigen() is used but not declared. I suppose it needs to be added to lib/external/ccmath/ccmath_grass.h, ?? I will have a look at it this evening. Hmmm, the files ccmath.h and ccmath_grass.h should be identical, but they aren't. Sorry, i did not checked that there is still debug code present in ccmath_grass.h. The correct way would be to remove ccmath_grass.h from lib/external/ccmath and copy the ccmath.h header file while compiling the library to the dist include/grass directory naming the file ccmath_grass.h. i.e.: Makefile change $(ARCH_INCDIR)/ccmath_grass.h: ccmath_grass.h $(INSTALL_DATA) ccmath_grass.h $(ARCH_INCDIR)/ccmath_grass.h To $(ARCH_INCDIR)/ccmath_grass.h: ccmath.h $(INSTALL_DATA) ccmath.h $(ARCH_INCDIR)/ccmath_grass.h I hope this will work. I will check and test this in the next days, i have no GRASS SVN access at work so i need to wait till evening to commit a fix. :/ cheers Soeren This can be caught if you add -Werror-implicit-function-declaration to your CFLAGS. Oh, i forgot this. Thanks, Soeren cheers, Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hi, I get two errors compiling 6.5svn. The first is because lib/gmath G_math_solvps() fails to return anything when it should return int. I think I fixed it with r38939, but please check to see if that is correct. https://trac.osgeo.org/grass/changeset/38939 The second is in lib/external/ccmath, eigen() is used but not declared. I suppose it needs to be added to lib/external/ccmath/ccmath_grass.h, ?? This can be caught if you add -Werror-implicit-function-declaration to your CFLAGS. cheers, Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hello Hamish, 2009/9/1 Hamish hamis...@yahoo.com Hi, I get two errors compiling 6.5svn. The first is because lib/gmath G_math_solvps() fails to return anything when it should return int. I think I fixed it with r38939, but please check to see if that is correct. https://trac.osgeo.org/grass/changeset/38939 Thanks, thats absolutely correct. The second is in lib/external/ccmath, eigen() is used but not declared. I suppose it needs to be added to lib/external/ccmath/ccmath_grass.h, ?? I will have a look at it this evening. This can be caught if you add -Werror-implicit-function-declaration to your CFLAGS. Oh, i forgot this. Thanks, Soeren cheers, Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Dear developers, i have submitted the changes to the grass6 development branch (a patch for grass7 is in development). I have tested the patch several times and hope to not break anything in grass ... . You can test the patched grass version with the latest grass testsuite. It ships additional tests for the ccmath, gmath and gpde libraries. http://www-pool.math.tu-berlin.de/~soeren/grass/GRASS_TestSuite/GRASS_Testsuite_0.2.0.14.tar.bz2 To compile the gmath and gpde libraries tests you have to switch to the library test directory and type make. Doing so the modules test.gmath.lib and test.gpde.lib are available to test the libraries. The gmath library test also ships some benchmarks for solver and blas 2,3 functions. The next step will be the replacement of the lib/gmath/lu.c code with the ccmath version including the patch of several modules. And the rewrite of several modules which are using self made matrix/vector functionality with the gmath vector and matrix functions. regards Soeren 2009/8/28 Hamish hamis...@yahoo.com Hamish wrote: I added a new lib/external/README.license file to explain things. AFAICT the SUBMITTING file and RFCs do not require adjustment. Soeren: Ok, i will put ccmath into lib/external together with the lgpl.license file and a README. actually there is nothing about the GPL or the implications of it in the SUBMITTING file, probably it is a good idea to add something there after all, with a be aware that files in lib/external/ ... rider. I don't think technical SVN restrictions are workable, but I'm open to being proved wrong by a good idea. Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hello Hamish, i am sorry for the delay, i was very busy in the last days ... 2009/8/21 Hamish hamis...@yahoo.com Hamish wrote: by the way, I notice that we already ship the bwidget LGPL library trunk/lib/external/bwidget/ That is for Tcl which is redundant for grass 7, actually that's used by tcl NVIZ so still in use. wx NVIZ is not yet to the point where it can replace the tcl version. (btw, we should rename visualization/nviz2/ to be wxNViz or something; the tcl version is already version 2.2) I added a new lib/external/README.license file to explain things. AFAICT the SUBMITTING file and RFCs do not require adjustment. Ok, i will put ccmath into lib/external together with the lgpl.license file and a README. Additionally i need to modify the lib/Makefile to put the directory external before the gmath library. But i think that is not a problem. Best regards Soeren Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hello Glynn, i am sorry for the delay, the last days were very stressing for me ... 2009/8/21 Glynn Clements gl...@gclements.plus.com Soeren Gebbert wrote: 3.) I can place an ATTENTION.txt file in the ccmath library folder, which explains the licensing issue. You can't assume that someone modifying the file will have viewed the contents of the folder. When I open a source file, it's often from grep or compilation output (C-x ` or selecting a specific line of output), or via M-. (find-tag, i.e. locate the point in the source where a given symbol is defined). If you want to ensure that the information is seen by anyone modifying the file, the information needs to go into the file itself. What would be better still would be if subversion could restrict the ability to modify those files, or at least send a reminder email if the files are modified. Can we modify the svn access to lib/external? As Hamish suggested i would like to place the ccmath library into this directory, because other LGPL libraries are located there. Best regards Soeren An alternative would be to put the library in e.g. lib/nonGPL/ccmath so that the pathname serves as a reminder. -- Glynn Clements gl...@gclements.plus.com ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Soeren Gebbert wrote: Additionally i need to modify the lib/Makefile to put the directory external before the gmath library. But i think that is not a problem. It should be okay to put the external subdirectory first, as it shouldn't depend upon anything in GRASS. -- Glynn Clements gl...@gclements.plus.com ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hamish wrote: I added a new lib/external/README.license file to explain things. AFAICT the SUBMITTING file and RFCs do not require adjustment. Soeren: Ok, i will put ccmath into lib/external together with the lgpl.license file and a README. actually there is nothing about the GPL or the implications of it in the SUBMITTING file, probably it is a good idea to add something there after all, with a be aware that files in lib/external/ ... rider. I don't think technical SVN restrictions are workable, but I'm open to being proved wrong by a good idea. Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Soeren: IMHO it may be difficult to re-license the ccmath library under the GPL ... . Hamish: why? Term 3 of the LGPL exists to describe exactly how to do that. Soeren: I am not concerned about the technical way, but about agreement of the original author of ccmath. Maybe i am wrong, but AFAICT as the copyright holder he has to agree that we can change the license of his code. The authors already gave consent to do that when they released the code under the terms of the LGPL, which explicitly allows that. The terms you see laid out in their LPGL.txt file are /their/ terms, not the FSF's. As a matter of good will it could be nice to drop them a note to thank them and say what we have done, but I don't think we need their permission to do it. None the less, my 2c vote would be to keep it as LGPL and put it in the existing lib/external/ directory. Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Soeren Gebbert wrote: 3.) I can place an ATTENTION.txt file in the ccmath library folder, which explains the licensing issue. You can't assume that someone modifying the file will have viewed the contents of the folder. When I open a source file, it's often from grep or compilation output (C-x ` or selecting a specific line of output), or via M-. (find-tag, i.e. locate the point in the source where a given symbol is defined). If you want to ensure that the information is seen by anyone modifying the file, the information needs to go into the file itself. What would be better still would be if subversion could restrict the ability to modify those files, or at least send a reminder email if the files are modified. An alternative would be to put the library in e.g. lib/nonGPL/ccmath so that the pathname serves as a reminder. -- Glynn Clements gl...@gclements.plus.com ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hello Glynn, 2009/8/15 Glynn Clements gl...@gclements.plus.com Hamish wrote: As for including LGPL code in the main trunk I personally don't have a problem with that as it reflects the upstream license/author's wishes, is new code, and is used as a library. By my reading the RFC2 doc nor 'g.version -c' the COPYING file need any adjustment to cover this. But a formal decision on that may be a matter for the PSC. As long as everything in the lib/ccmath/ dir falls under the same (GPL compatible) license and the LGPL.txt file is placed in that dir I'd be happy. If it's being included in the main GRASS tree, it would be simpler to relicense it as GPL. So we have to convince the author Daniel A. Atkinson to relicense the ccmath library as GPL code? The only email address i have found is: dan...@aol.com. I personally prefer to keep the LGPL license for ccmath. The main thing is to ensure that each file has a header stating that it is licensed under the LGPL. In the absence of such a header, we can't reasonably assume that any additions to the code are licensed under the LGPL. I will ensure this. Best regards Soeren -- Glynn Clements gl...@gclements.plus.com ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Soeren wrote: if there are no objections against the changes and the new LGPL'd ccmath library? I have no comment either way on the technical side of the library changes beyond wondering if it will peacefully interact with the (mostly unused) existing --with-blas --with-lapack ./configure switches. As for including a dependent library in GRASS I personally don't have a problem with that as it does not seem to be widely available or have much infrastructure of its own (beyond a e.g. Freshmeat.net writeup page*). i.e. not a fork because there isn't really any new version to keep in sync with, and users would have a hard time finding a package for it. As for including LGPL code in the main trunk I personally don't have a problem with that as it reflects the upstream license/author's wishes, is new code, and is used as a library. By my reading the RFC2 doc nor 'g.version -c' the COPYING file need any adjustment to cover this. But a formal decision on that may be a matter for the PSC. As long as everything in the lib/ccmath/ dir falls under the same (GPL compatible) license and the LGPL.txt file is placed in that dir I'd be happy. Hamish [*] a web search for ccmath found a most interesting auto-georegistration tool for applying/projecting a 3D registration onto 2D photographs: gipfel. Check out the .avi demo. http://io.debian.net/~tar/debian/gipfel/ Note the track overlay is from a x,y,z text file, very neat. reminds me a bit of i.points.auto, Stereo[1], and autopano-sift[2] software. [1] http://grass.osgeo.org/gdp/stereo-grass/index.html [2] http://user.cs.tu-berlin.de/~nowozin/autopano-sift/ ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Hamish wrote: As for including LGPL code in the main trunk I personally don't have a problem with that as it reflects the upstream license/author's wishes, is new code, and is used as a library. By my reading the RFC2 doc nor 'g.version -c' the COPYING file need any adjustment to cover this. But a formal decision on that may be a matter for the PSC. As long as everything in the lib/ccmath/ dir falls under the same (GPL compatible) license and the LGPL.txt file is placed in that dir I'd be happy. If it's being included in the main GRASS tree, it would be simpler to relicense it as GPL. The main thing is to ensure that each file has a header stating that it is licensed under the LGPL. In the absence of such a header, we can't reasonably assume that any additions to the code are licensed under the LGPL. -- Glynn Clements gl...@gclements.plus.com ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
Re: [GRASS-dev] gmath/gpde Patch for grass6.5 and grass7
Soeren wrote: In case the developers don't have any objections against the patch, i would like to see them commited to the svn repositories. Because i do not have write permissions on the osgeo grass reporitory, a developer with write acces must do this please. since you were already accepted as a developer and had CVS write access on the old server, all you need to do now to activate SVN write access on the new server is post to the grass-psc mailing list giving your osgeo ID and that you have read+understand+accept the RFC2 document. see https://trac.osgeo.org/grass/wiki/HowToContribute#WriteaccesstotheMainGRASS-SVNrepository cheers, Hamish ___ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev