Philip Martin wrote:
>Branko ÄŒibej <brane@xbc.nu> writes:
>
>
>
>>The intent of svn:eol-style was to get coherent results (diffs, that
>>is) when people use all those editors out there that can't preserve
>>the eol style.
>>
>>
>
>You say "all those editors", are there lots of them? I find it hard
>to imagine using such a tool.
>
Sadly, such things exist, especially in the Windows world. But even
worse, most editors on Windows will blindly use \r\n, regardless of the
prevailing style in the file.
>Do they change anything else? Spaces to tabs, say, or trailing whitespace?
>
Such things happen, too, but are less common.
>Even if there are (Windows?) editors that, say, accept LF line endings
>and write CRLF line endings, is that a big problem? After the first
>checkin, the one that translates the file, subsequent checkins aren't
>really a problem. Or are we attempting to handle scenarios where
>there are two broken editors, one doing LF-to-CRLF and the other doing
>CRLF-to-LF? How likely is that?
>
In any environment where you're editing on more than one platform. I
expect most Unix editors write just \n, again regardless of the
prevailing style. The result is a mess.
Not to mention many, many Unix C compilers that will turn a blind eye to
line continuation if they happen to see backslash-CR-LF.
>Do we have any real-life examples where svn:eol-style fixes a problem?
>Why should it be Subversion's job to solve this problem anyway?
>
>
Because that's what people expect of a cross-platform version management
tool. Sad, but true.
--
Brane Čibej <brane_at_xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 12 21:22:54 2003