[cvsnt] Re: Escape sequences in $Id$ tags

[Michael Wojcik]

> From: cvsnt-bounces at cvsnt.org 
> [mailto:cvsnt-bounces at cvsnt.org] On Behalf Of Glen Starrett
> Sent: Thursday, 27 April, 2006 20:26
> 
> Dan Troesser wrote:
> > Have you ever worked around this problem before in CVSNT?
Apparently,
> > because my id starts with u and in this case is preceded with a \
from the
> > domain name, the compiler thinks it's an escape code.  Can we tell
CVSNT to
> > ignore domain names?
> 
> The domain name is going to be significant in larger networks and 
> shouldn't just be ignored.  Maybe a substitution of / for \ in the 
> username string would be sufficient?  It would still parse for human 
> readers, and shouldn't be trouble for most compilers.  

That seems like the best option to me.  A conforming C implementation
won't care about (apparent) Unicode / ISO 10646 escapes in a comment
(ISO 9899:1999 6.4.3 and 6.4.9), but apparently some Java compilers do.

Note that many (though not all) Windows applications are happy to use a
forward slash instead of a backslash between domain and username anyway,
just as most Windows applications other than cmd.exe built-ins are happy
to take forward slashes as path separators.  So substituting a forward
slash in the expansion of the Id (and Author, Header, Locker, and Log)
keywords wouldn't be all that odd.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus