Mismatch docs/code. WAS: RE: svn commit: rev 1774 - trunk/subversion/include trunk/subversion/libsvn_delta

[Sander Striker <st...@apache.org>]

Hi,

> +  /* Two-way and three-way diffs both call the first two output functions: */
> +
> +  /* If doing a two-way diff, then an *identical* data range was found
> +     between the "original" and "modified" datasources.  Specifically,
> +     the match is from ORIGINAL_START to ORIGINAL_END in the original
> +     data, and from MODIFIED_START to MODIFIED_END in the modified
> +     data.
> +
> +     If doing a three-way diff, then all three datasources have
> +     matching data ranges.  The range LATEST_START to LATEST_END in
> +     the "latest" datasource is identical to the range ORIGINAL_START
> +     to ORIGINAL_END in the original data, and is also identical to
> +     the range MODIFIED_START to MODIFIED_END in the modified
> +     data.  */
> +  void (*output_common) (void *output_baton,
> +                         apr_off_t original_start,
> +                         apr_off_t original_end,
> +                         apr_off_t modified_start,
> +                         apr_off_t modified_end,
> +                         apr_off_t latest_start,
> +                         apr_off_t latest_end);

Because I missed a rename in the code before committing we now have
_end in the .h file and _length in the .c file.  I'm going to let them
match up again, but would like some feedback to what is preferred.

Do we want ranges to be represented by:

  START - END  e.g.  0 - 3 translates to [0, 1, 2]

or:

  START, LENGTH  e.g.  3, 3 means [3, 4, 5]


Thoughts?

Sander


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Mismatch docs/code. WAS: RE: svn commit: rev 1774 - trunk/subversion/include trunk/subversion/libsvn_delta

["Kirby C. Bohling" <kb...@birddog.com>]

Sander Striker wrote:
> 
> Because I missed a rename in the code before committing we now have
> _end in the .h file and _length in the .c file.  I'm going to let them
> match up again, but would like some feedback to what is preferred.
> 
> Do we want ranges to be represented by:
> 
>   START - END  e.g.  0 - 3 translates to [0, 1, 2]
> 
> or:
> 
>   START, LENGTH  e.g.  3, 3 means [3, 4, 5]
> 
> 
> Thoughts?
> 
> Sander

	I have always thought START - END is much more natural and easy. 
However, it appears every library in exsistance seems to disagree with 
me.  Every time I have ever seen a "substring" like operation they all 
use "start, length" never "start - end" (now that I say that 10 people 
will reply with counter examples).

	I have yet to see a common C or C++ library that represents it the START 
- END way.  Of course using this argument we never will.  In the end it 
ends up making them work like memcpy(), (the third param is a number of 
bytes, not the place to stop). which makes it easier  to remember. 
That's my 2 cents.

		Kirby


> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
> 



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org