geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeremy Boynes" <jer...@coredevelopers.net>
Subject RE: CR/LF cleansing
Date Sun, 31 Aug 2003 18:18:12 GMT
What seems to be happening is that people are submitting tarballs in
patches that contain DOS format files (CRLF). When these are expanded by
people running Unix or MacOS, the CRLF is kept and the file committed as
a Unix file with this is. vi often shows these files with a ^M at the
end of the line but I doubt GUI editors (e.g. IDEs) do so.

So two things:
1) if you are running on Windows, please ensure you are using a text
mode CVS
   e.g. WinCVS or Cygwin cvs on a non-binary drive
2) if you are running on Unix, Linux or MacOS, please check all patch
tarballs
   for DOS-style CRLF before committing them

Thanks
Jeremy

> -----Original Message-----
> From: Dain Sundstrom [mailto:dain@coredevelopers.net] 
> Sent: Sunday, August 31, 2003 10:29 AM
> To: geronimo-dev@incubator.apache.org
> Subject: Re: CR/LF cleansing
> 
> 
> I ran into this my self the other day.  I committed a bunch of new 
> files that were posted in zip format, and when I looked at the files 
> using my editor, I did not notice the line ending problems.  I also 
> didn't notice that all of the files had execute permissions 
> on, so for 
> now on I'll double check everything before I commit it at the command 
> line with 'ls' and 'od'.
> 
> -dain
> 
> On Sunday, August 31, 2003, at 10:36 AM, Daniel S. Haischt wrote:
> 
> > Jason Dillon wrote:
> >
> > [...]
> >
> >> It is not just a client issue.  Patches that come in have bunk line
> >> endings too.
> >
> > i consider this a client _only_ issue, presuming the statement ...
> >
> >  Developer.getCurrentCVSClient().equals(CygwinCVSClient);
> >
> > evaluates to true. why am i stating that?
> >
> > quote [1]:
> >
> >  'I suspect this is because the CVS cygwin client has no idea  it's 
> > running on Windows'
> >
> > if the above statement is true and a developer creates a diff for 
> > example according to the CVS Book [2] using ...
> >
> >  cvs diff -u [file to be patched]
> >
> > she, the developer, will use the Cygwin cvs binary which 
> thinks that 
> > ASCII files are stored with UNIX LF line endings but actually all 
> > drives are mounted in binary mode or vice versa (ASCII files are 
> > asumed to have DOS CRLF but all drives are mounted in text mode).
> >
> > so the essential point of the above statements is that if 
> the client 
> > is not configured accordingly, it will always mess up files - no 
> > matter whether someone did a 'cvs commit' or 'cvs diff'.
> >
> > regards
> >
> > daniel s. haischt
> > --
> >
> > references:
> >
> > [1]: http://www.dehora.net/journal/archives/000330.html
> > [2]: http://cvsbook.red-bean.com/cvsbook.html#diff
> >
> >
> >
> >
> >
> 
> /*************************
>   * Dain Sundstrom
>   * Partner
>   * Core Developers Network
>   *************************/
> 
> 


Mime
View raw message