Re: DOS linefeeds in icu-3.2.tgz?

I don't have a preference with respect to line endings, but I think there
should be a clear policy on file format.
The description of how the files are compiled doesn't really answer the
question because it gives the impression that the file format might change if
equipment is changed for any reason.

I suggest picking a line ending style and commiting to it. Then if people want
other formats they know what they are  dealing with and can act accordingly. It
shouldn't seem capricious or subject to changes.

Just a suggestion...
tex

George Rhoten wrote:
> 
> Our build system downloads ICU through CVS on a Windows machine, uploads
> it to a Linux machine, uses zip to extract and convert the text files, and
> then it creates a compressed tar file from the extracted files.
> 
> It's done this way so that when CVS HEAD is downloaded, the same files are
> going to be in both zip and tar files.  People check in stuff all the
> time.  It's also faster this way.
> 
> The DOS newlines in the UTF-8 files aren't intentional, but I don't see
> them being bad either.  vim can handle it quite well.
> 
> George Rhoten
> IBM Globalization Center of Competency/ICU  San José, CA, USA
> 
> "Steven R. Loomis" <[email protected]>
> Sent by: [email protected]
> 11/23/2004 06:08 PM
> 
> To
> Deborah Goldsmith <[email protected]>
> cc
> icu list <[email protected]>
> Subject
> Re: DOS linefeeds in icu-3.2.tgz?
> 
> All of the locale files are UTF-8,  I wonder if that threw off some
> linefeed conversion somewhere.
> 
> On Nov 23, 2004, at 6:03 PM, Deborah Goldsmith wrote:
> > For some reason a lot of the locale data in the released icu-3.2.tgz
> > has DOS line breaks (CRLF) rather than Unix line breaks (LF). The CVS
> > repository has Unix line breaks. Was this intentional? I can
> > understand the .zip version having DOS line endings, but this is
> > confusing.