If a developer checks in multiple files in a single commit, is there anyway of finding out what all the files are using either bash or perl scripts on the cvs server? I've tried using the 'cvs status' command in a bash script tied into the commitinfo file, but that only works on a directory by directory basis. It may help if I give an overview of what I'm aiming to do:
We have a project which consists of a number of software components, the source for which is kept in separate subdirectories. Each separate component has associated with it a history file, so each of the subdirectories has a history.txt file in it (sometimes at the top level, sometimes in a further subdirectory). I want to use cvs to ensure that when a developer checks in a bunch of files, that he/she is also checking in an updated history file which gives an overview of his changes. It is possible for one developer to want to check in files from many different components (and so should be required to commit several different history.txt updates). What I need to do then, is have commitinfo run a script which (i) gets the names/paths of all the files being checked in, (ii) makes sure that each component which has been changed has a changed history file included in the checkin. My problem with using cvs status is that it will only give out information on multiple files that belong to the same directory. I need something like cvs status, but which outputs all the files being commited, regardless where they belong in the repository. Is their anyway of doing that? For info, the version of cvs running on the server is 1.11.5 Thanks Jason Gibbons ==================================================== ==================================================== _______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
