Jacob Keller <jacob.e.keller@xxxxxxxxx> writes: > From: Jacob Keller <jacob.keller@xxxxxxxxx> > > Modify logic of check_refname_component and add a new disposition > regarding "*". Allow refspecs that contain a single "*" if > REFNAME_REFSPEC_PATTERN is set. Change the function to pass the flags as > a pointer, and clear REFNAME_REFSPEC_PATTERN after the first "*" so that > only a single "*" is accepted. > > This loosens restrictions on refspecs by allowing patterns that have > a "*" within a component instead of only as the whole component. Also > remove the code that hangled refspec patterns in check_refname_format > since it is now handled via the check_refname_component logic. > > Now refs such as `for/bar*:foo/bar*` and even `foo/bar*:foo/baz*` will > be accepted. This allows users more control over what is fetched where. > Since users must modify the default by hand to make use of this > functionality it is not considered a large risk. Any refspec which > functioned before shall continue functioning with the new logic. Thanks. Now I can read the changes to the code in these two commits and see that they both make sense ;-) The above description seem to use "ref" and "refspec" rather liberally, so I'll rewrite some parts of it to clarify while queuing. By the way, have you run test suite before sending this (or any previous round of this) patch? This seems to break t5511-refspec.sh for me. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html