Re: rpms/mlt/devel .cvsignore, 1.4, 1.5 mlt.spec, 1.4, 1.5 sources, 1.4, 1.5

On Fri, 11 Dec 2009 12:18:23 +0100, Zarko wrote: The end-result is the same. You end up with modified files in your cvs working-copy, which you need to commit to the cvs server. And that's easy: cvs diff -u | less for a final look at what you've changed compared with what is found in the cvs repository. make clog to extract the latest .spec %changelog comment into a file "clog", which can be used as the cvs commit log message: cvs commit -F clog or cvs commit -m "some alternative log message" to commit the modifications found in all local files (with a proper log message), and if that worked without errors, make tag to apply a cvs tag to all [comitted] files of the package. If none of these steps caused an error, the new tag is ready to be used by make build Room for mistakes is given when trying to shorten this sequence with the help of macros/aliases and when not paying attention to cvs error messages. For example, "make tag" will not tag all files as appropriate, if some of the local changes have not been committed before. [...] For updates that require a new source tarball, there are two commands which upload such source archives to the lookaside cache (a server that doesn't store them in cvs) prior to modifying other files in cvs: make new-sources FILES="fu-1.0.tar.gz bar-1.0.tar.gz" to upload one or multiple files and store their file names in new local "sources" and ".cvsignore" files. Or: make upload FILES="fu-1.0.tar.gz bar-1.0.tar.gz" to upload one or multiple files and append their file names to the local "sources" and ".cvsignore" files. In either case, the changes in the "sources" and ".cvsignore" files need to be committed to cvs just like ordinary changes made in the .spec file. The "cvs commit -F clog" above also commits the changes found in those files.