[postgis-devel] extras/tiger_geocoder/upgrade_geocoder.bat and CRLF

Regina Obe lr at pcorp.us
Mon Aug 7 14:11:33 PDT 2017


Disregard my note.  Maybe I was catching things mid change or I had a typo in my testing.

Strk latest commit for .editorconfig works for me.  CRLF is preserved.

-----Original Message-----
From: Regina Obe [mailto:lr at pcorp.us] 
Sent: Monday, August 07, 2017 5:06 PM
To: 'PostGIS Development Discussion' <postgis-devel at lists.osgeo.org>
Subject: RE: [postgis-devel] extras/tiger_geocoder/upgrade_geocoder.bat and CRLF


> In my case .editorconfig only runs if I open the file with my editor, which didn't happen when I noticed the change. So it must have been .gitattribute that changed the file from LF to CRLF.

> I've no idea what svn:eol-style is supposed to do, but I'm pretty sure that regina committed the file with LF line endings.

> Regina: can you run some tests with and w/out the .editorconfig change ?
> Does "svn diff" spots the difference between a LF and CRLF version ?

> --strk;
Yap it's probably my fault.

I'm editing with VS Code which is editor config aware.  

When I svn update the file fresh it comes down with a CRLF line break After editing with VS Code, it changes it to LF.

So that explains why it's an issue on git.  I guess git is seeing my commit as LF and maintaining that but the .gitattributes is forcing on pull to change it to a CRLF and so a further Git pull whatever sees the local file as different from server.

Now editing the editorconfig
Just putting in a section for bat does not seem sufficient because the 

[*]
end_of_line = lf

seems to dominate.

If I remark out the 

#end_of_line = lf 

at the top and add a section for bat with

end_of_line = crlf

it works, my editor preserves the CRLF.

Do we really need the?

end_of_line = lf







More information about the postgis-devel mailing list