Hi all, I am setting up a local CVS archive and I found that when I imported the boost 1.33.1 distribution there are some files that get ignored by CVS. Eventually I figured out that the missing files are described by the path expression "libs/python/pyste/tests/*.cpp" relative to the top level directory. From that I found the .cvsignore in the same directory that specifies *.cpp as an ignore pattern. There are more .cvsignore files in the distribution tree, but this is the only one that filters out files for an import. Does anyone have any guidance as to how CVS imports of Boost source code distributions should be done? Should I delete all of the .cvsignore files in the distribution before importing? Should I replace them with an empty version of the file? Should I do something else? A related question for developers. Should the .cvsignore files be removed from the distribution tree before it is released to the public? Thanks, Rush
Rush Manbert wrote:
I am setting up a local CVS archive and I found that when I imported the boost 1.33.1 distribution there are some files that get ignored by CVS. Eventually I figured out that the missing files are described by the path expression "libs/python/pyste/tests/*.cpp" relative to the top level directory. From that I found the .cvsignore in the same directory that specifies *.cpp as an ignore pattern.
There are more .cvsignore files in the distribution tree, but this is the only one that filters out files for an import.
Does anyone have any guidance as to how CVS imports of Boost source code distributions should be done? Should I delete all of the .cvsignore files in the distribution before importing? Should I replace them with an empty version of the file? Should I do something else?
A related question for developers. Should the .cvsignore files be removed from the distribution tree before it is released to the public?
Now that I've had some lunch and thought about it some more, if the *.cpp files are in the .cvsignore, then it seems that they shouldn't be part of the source code distribution at all. It looks like they are artifacts of some processing that was done post-export and pre-tar. Does that make sense? - Rush
participants (1)
-
Rush Manbert