Naveen Patil <naveen at vinayak.engr.sgi.com> wrote:

More specifically: If you check in a text file from a Perforce client running on an OS that uses MS-DOS end-of-line conventions (MS-DOS, Windows 9x, NT), Perforce understands that CR-LF should be interpreted as EOL; if you check in a text file from an Unix Perforce client, Perforce understands that LF should be interpreted as EOL. When you then check out the file using a different Perforce client, you will get whatever is the correct EOL convention for that Perforce client's OS. I.e., Perforce does the right thing.

[I only have experience with NT and Unix Perforce clients; I assume that "Perforce does the right thing" also applies to Macintosh clients. (Macintoshes use just CR for EOL?)]

You'll get the "wrong" results only when you do something like create and edit a text file on one system, then binary copy the file to another system which uses a different EOL convention, and then proceed to check that into Perforce as a text file. E.g., if you create a text file on NT, FTP it to a Unix system, and check it in from the Unix system, then when you check it out on an NT system, you get CR-CR-LF at the end of each line, and when you check it out on an Unix system, you get CR-LF at the end of each line.