[EMAIL PROTECTED] writes:
>
> The simplistic assumption that:
>
> - "If I'm editing files on Windows I want the OLD MS-DOS
> conventions"
> - "If I'm editing files on Unix I want Unix conventions"
>
> Is not really sufficient in today's world.
That not the assumption that CVS makes. The assumption CVS makes is:
"If I'm working on platform X, then I want text files to be in the
canonical form for platform X." If all you ever do is work on Unix and
MS-DOS/Windows, then the canonical text file formats are similar enough
that it's tempting to try to convince all your tools to handle both
forms, but CVS runs on a much wider range of systems, some of which have
text file conventions that are completely different. Trying to have CVS
understand all the different conventions is a lost cause.
> Where-as the Emacs session on Win2000 is quite happy.
>
> I think that 'at least' names of entities in the CVS repository should
> be held (or at least parsed into) canonical form. These are (almost)
> opaque cookies the CVS server stores on the client system, not client data.
If CVS treated those as binary files instead of text files (which is
what you're proposing, whether you know it or not), then on some
platforms you wouldn't be able to (conveniently) use a text editor to
fix them when someone moves the repository. I'm not sure I'm willing to
pay that price.
-Larry Jones
That's the problem with nature. Something's always stinging you
or oozing mucus on you. -- Calvin
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
