On Fri, March 11, 2011 4:30 am, Matt Welland wrote: > What is the plan for handling symlinks for fossil? My preference would > be for fossil to treat links as a file. Storing the pointer would be > great. Just ignoring links would also be fine by me. But actually > traversing the links is not a good idea IMHO and is a real fossil > killer. It is also not easy to work around.
Just don't ever do fossil add * Simple. fossil add `find . -type f` might be useful. I usually list new stuff with fossil extra then weed out the unwanted and turn the list into one or more adds. The whole point of symlinks is that most programs don't even know they are not what they point to. Oh, and Fossil shouldn't be storing the actual link, the stuff might be somewhere different on a different machine. > > Links are commonly used to point to large amounts of static data > outside of a build area or to reference one piece of data in multiple > places. In neither of these cases would I want my scm tool to traverse > the links and add the data. Exactly, so don't do wildcard adds (there are of course other reasons for not doing that). > A setting "ignore-links" would be an ok > compromise. > Platform dependent. BTW, there is no such thing as a no-thought no-effort SCM program. Regards, Eric _______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
