This question is somewhat similar to my other question, http://stackoverflow.com/questions/3455772/make-sure-files-are-converted-crlf-into-lf-in-an-update-hook-is-there-a-performa
So here is what am looking for with the below architecture. 1. MY parent repository (bare and data) is on a UNIX machine. 2. I can clone my repository in a UNIX machines. 3. I can clone my repository in a Windows machines using samba to access my parent repo.
How can i take care of CRLF issues if
The user creates clones in UNIX, maps it using samba into windows drive, the modifications are done in Windows which create a CR/LF character pair as the EOL. If the users come back to Unix and commits and pushes. How does GIT take care? or do we need to have some hooks in place?
Same as above, but the file format exceeds 8000 characters per line - a lot. Is this treated as a ASCII file with the CR stripped out on commit?
Variation of 2, but it's a binary file which has an ASCII header. Is this going to inadvertently have CR/LF's changed to LF?