thanks. Your answers helped me a lot. May I ask some further
questions to ensure I got everything right?
Am 2007-05-28 um 11:16 schrieb Ryan Schmidt:
> On May 28, 2007, at 03:35, Holger Schulz wrote:
>> after installing Subversion via fink on MacOS X 10.4.9 and
>> running some test, I have some questions which some of you
>> might be able to answer.
>> 1. Okay, the data in my repositories will mainly be LaTeX files,
>> since I need these for my work. Typesetting will produce PDF file,
>> which shouldn't be merged but overwritten. Does subversion know that
>> overwriting is necessary in these cases ordoes I have to tell it
>> somehow? Same question for MS-Word documents and similar which I have
>> to use sometimes in collaboration with some colleagues.
> Subversion does handle binary files just fine, but it will not
> blindly overwrite data. If, for example, two people independently
> modify a PDF or Word document, then there will be a conflict, which
> a developer will have to resolve by examining both versions of the
> file to determine what to do.
At this time my reps will only be used by myself.
> With text files, you would get a line-by-line diff, but it can't
> show you that with binary files since it doesn't know how to
> interpret binary formats.
Does that mean, that subversion does recognize which data is binary
and which can be understood as text? Or do have to tell subversion
what is what?
> Note that many will recommend not storing generated files in the
> repo. In your case, if the PDFs are generated from the LaTeX files,
> then store the LaTeX files in the repo along with a script that
> generates the PDFs, but not the PDFs themselves.
And that script would generate the PDF file on checkout? I'm not sure
if I understood this point.
> Others will give you different advice, hinging on the fact that the
> software needed to generate the PDFs may be hard to install, or
> that its version may change and its output might thus change, and
> that one would want a record of the way the PDFs looked with the
> old software.
Perhaps sometimes it is easier to check out the PDF files only, e.g.
to print them without having to typeset them.
>> 2. Currently I have a G4 desktop which holds the repositories. I also
>> have a PowerBook which I you for work at home and from outside. At
>> home I connect the desktop from the book via its Bonjour name
>> <something>.local, to connect it within my local net. From outside I
>> have to use a global name (in this case via dyndns). So the problem
>> appears that the same repository might apear under two different
>> names. How can I tell the working copy that its repository has
> In the working copy:
> svn switch --relocate $OLDPREFIX $NEWPREFIX
> where $OLDPREFIX and $NEWPREFIX are the parts of the URL that
> changes. For example, if you're relocating from svn://foo.local/
> repo to http://foo.dyndns.org/repo then you could issue
> svn switch --relocate svn://foo.local http://foo.dyndns.org
I'll try that.
> It is more portable and safer to back up the output of "svnadmin
> dump" which works independently of the processor architecture and
> the version of BerkeleyDB, and can be done safely regardless of
> whether anyone is reading or writing at the time. Also, you can do
> incremental dumps with this utility, which could be useful if
> you're doing permanent backups to CD.
Sounds good. I'll try that.
To unsubscribe, e-mail: firstname.lastname@example.org
For additional commands, e-mail: email@example.com
Received on Mon May 28 14:15:41 2007