Re: line-ending conversion and keyword substitution

At 01:46 PM 12/13/2001 -0500, Greg Hudson wrote:
>Are you actually envisioning a vector graphics program which saves data
>by appending to the data file it loaded? Of course not; when William
>saves his work the vector graphics program is going to write out a valid
>data file, from scratch.

Oh, so the CRLF is not valid data? I was thinking it would do something in
the program (and therefore be maintained), like set properties you weren't
expecting.

Just like an image where a portion of the image becomes corrupted but isn't
noticed. Or a Word document where some properties are changed randomly and
they come back to bite you when you least expect it. Or a JAR file where
the contents of a class become corrupted.

I guess that in your scenario, though, your style of corruption is indeed
better than mine since it won't survive an edit.

>[Regarding the .dsp file case requiring an explicit CRLF style:]
>I was skeptical too, but I don't think it's uncommon.

I'm not saying requiring CRLF within a .dsp is uncommon, but how often does
someone edit a .dsp on a Unix box? When they do, how often do they get a
mixture of line endings in the file? If they do, how often do they expect
their versioning system to fix the problem?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:53 2006