subversion-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Cooke, Mark" <mark.co...@siemens.com>
Subject RE: Commit size
Date Tue, 12 Jun 2012 15:53:26 GMT
[Please add your reply at the bottom, it makes it easier to read] 

> > 2012/6/12 Andy Levy <andy.levy@gmail.com>:
> >> On Tue, Jun 12, 2012 at 5:21 AM, Vladimir Shun'kov 
> <shunya.v@gmail.com> wrote:
> >>> Hello,
> >>>
> >>> Could you please help me with issue. Is it real to get 
> >>> commit size of user for statistic?
> >>
> >> How are you measuring size? Bytes? Lines? Number of files? 
> >> And to what end?
> >>
> >> There are a lot of problems around measuring user activity by this
> >> sort of statistic. It's easily gamed, tends to ignore the 
> >> effects of improving code in many ways[1], and in the case of
> >> binary files, it's pretty much meaningless.
> >>
> >> 1: 
> >> http://folklore.org/StoryView.py?story=Negative_2000_Lines_Of_Code.txt
> 
> On Tue, Jun 12, 2012 at 7:55 AM, Vladimir Shun'kov 
> <shunya.v@gmail.com> wrote:
> > Hello,
> >
> > I need measure in bytes. I tried use pre-commit and
> > post-commit hooks but I received not real size of
> > commit due svn uses some compression.
>
...but what do you consider to be the "real size of commit" ~ the number of bytes of ...all
affected files / ...just the changes / ...something completely different?

What are you _actually_ trying to measure (and why)?  Without knowing this, we cannot help
you.

Note: Subversion does not send all of the file(s), only the diffs (which includes all the
file if new, of course!).  This helps reduce network traffic and time.

~ mark c

Mime
View raw message