Compiling the current version of OpenSees

motalaat · Post by **motalaat** » Wed Aug 02, 2006 5:13 pm

I just updated my OpenSees using CVS. I used to have an OpenSees that was compiling wonderfully on VC++6.0. Now, I ran into a ubiquitous compile-tine error. The compiler did not understand (and neither did I), the lines containing "<<<<<<<<<<SomeNunmber", or "=========SomeClassName" in files such as TclModelBuilderUniaxialMaterialCommand.cpp. After much thinking and frustrating trials about it, and a peek at version of an updated file I had from its author which was identical but did not have the "<<<<<" and "======" strings, I took a look at the source code online. There were no such maddening strings, and after commenting them out the compiler finished successfully. I thought I'd pass on the word that some files do get corrupt in this manner while transferred via CVS for what it's worth, so that other developers may not get as frustrated as I did.

Post by **fmk** » Wed Aug 02, 2006 6:27 pm

the maddening <<<<<<<<<<<< is CVS telling you that conflicts exist, i.e. you have modified
lines of code code that has been also modified and checked in by someone into CVS .. they
are not corrupted, it is cvs actually doing what it is supposed to do .. if you want the version
at the website, just delete the file and do a cvs update again .. suggest you look at the cvs manual for a full and better explanation.

motalaat · Post by **motalaat** » Wed Aug 02, 2006 6:52 pm

This was very relieving to know. I'm glad I'll have the option to review what gets overwritten. Yes, I did add a few lines to some "main" interpreter files to link them to my own interpreter files. Although, btw, one of those files which gave a couple of <<<<< errors, TransformationDOF_Group.cpp, I never before touched or even knew it existed, but there muct be some other valid and simple explanation that doesn't matter much at the moment. Thanks, Frank. This really was educational.