RE: Re: Re: Detecting

On 12/03/2015 10:29, Timur Khanipov wrote:
>> If the line ending style happened to match the native line endings, they show as not
>> modified, correct? It's only when the actual line endings in the file are
>> not the native style that the file is marked modified, correct?
>
> Incorrect. As I mentioned in the initial post and my first reply to Stefan, the 'modified' status (the one which I consider improper) appears only with CRLF files and CRLF is the native line ending for Windows.

Which is correct behaviour. Because CRLF line endings will be converted to LF endings on commit, and then reconverted back to CRLF on checkout-on-Windows. This is a file content modification.

Technically I suppose it should *also* show Modified for files with LF endings on Windows, because if you commit those they will actually convert to CRLF endings.

eol-style is a weird property, since it alters file content as well. So changing this property is rarely a pure property change only.

>> I think if a line ending actually in use does not match the line ending SVN
>> expects for that file, that marking it modified makes sense.
>
> I disagree that it makes sense. When a user sets svn:eol-style to 'native', pending modifications of line endings in the working copy are not interesting for review in any way (because the user explicitly told Subversion not to track line ending changes).

That is completely the opposite of what that means. If you want SVN to not care about line endings, then omit the property. If you set the property then you are explicitly telling SVN to care about line endings and change them in the specified ways when files are committed and checked out.

I have a policy to never set that property, because I'd much prefer to preserve file content completely unaltered rather than have surprise EOL changes. All of my editors can cope with non-native EOL styles anyway.

------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=4061&dsMessageId=3104630

To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2015-03-12 00:37:02 CET