[cvsnt] Re: cvswrappers

Vermes, Istvan IVermes at Graphisoft.hu
Wed Oct 20 11:19:47 BST 2004


Hi again, and thanks for the quick answer!

Tony Hoyle wrote:
> 
> Vermes, Istvan wrote:
> > Sometimes it works well, sometimes not.
> >
> > Then we put the last row to the beginning of the file. It works at
times
> > as well, but not in all case.
> >
> > We doesn't able to determine how does it work in this case?
> 
> You can't.  The order of reading of cvswrappers is completely random.
> It actually varies depending on the version of the client you're
using,
> and whether there's a local .cvswrappers file.  Different
> implementations/cvs versions do it in different ways.
> 
> It's far better to have the server configured 'normally' by defining
> binary files, and have local wrappers on the mac machines.  Also
forget
> about the '-m' options, as they're redundant (I probably should
properly
> depreciate them).
> 
> The Mac clients shpuld determine what is binary on that platform
> automatically.  It's not something the server really needs to handle
> (I'd expect that MacCvsX does, since I've not heard of others having
the
> problem).

Ok, the order of reading of cvswrappers isn't defined.
But how does the processing of one cvswrapper work inside my file (or if
I have exactly one cvswrapper file), when there is a redundant rule?

My problem with the Mac is that I store a lot of MacOS binary files in
the repository. It's ok that the MacCvs client can determine the type of
the files in a workfolder. But if I run an 'update' command for example,
it's important for the repository to know which files are binary. (I
don't want to merge binary files of course.)

Vermes Istvan



More information about the cvsnt mailing list