Thanks Dmitry, Jonathan and Junio for the comments. I'm happy to have fast-import be strict about its format, and have added code and tests that demand exactly one space, or an end-of-line, as necessary. I also made sure the other error messages involved with parsing datarefs are correct. Jonathan, good observation on CRLF users. If we did want to cater to them, doing it centrally in read_next_command() would be the way to go. But why bother. Regarding fixing up all end-of-line number parsing, I think the only other one is dates. Both "raw" and "now" check for garbage at end-of-line, but "rfc2822" uses a generic function that accepts junk. I'm not motivated to add a lot of code to fix that corner case. Junio, I made the commit message more clear. For the idea of combining find_mark + parse_mark, that isn't general enough for all users. This is the construct used in many places: oe = find_mark(parse_mark_ref_space(p, &x)); Also, I did the unit tests first, to make sure things were broken as I expected. You can squash it all together if you prefer. -- Pete Pete Wyckoff (2): fast-import: test behavior of garbage after mark references fast-import: tighten parsing of mark references fast-import.c | 97 ++++++++++++++---- t/t9300-fast-import.sh | 267 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 342 insertions(+), 22 deletions(-) -- 1.7.10.rc2.2.g38670 -- 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