threads / discuss / 24872

Treating similar blank lines between hunks as part of the hunks

Subject: Treating similar blank lines between hunks as part of the hunks

## tl;dr

2 messages between Aug 26, 2010 and Aug 27, 2010.

replies: 1people: 2as markdown or json

Timur Tabi· Aug 26, 2010, 20:19 UTC · lore

Is there any way to convince git-diff to treat blank lines between two hunks as part of the hunks for diffing? That is, if I have these two files:

----- A B C

D E -----

and

----- A F G

H E -----

git-diff will do this:

 A
-B
-C
+F
+G

-D
+H
 E

Instead, I would prefer it did this:

 A
-B
-C
-
-D
+F
+G
+
+H
E

This second diff is easier to read, especially when the number of hunks is large, but the size of each hunk is small. This can happen in source code when one function is replaced with a completely different function. Git sees the blanks in each function as being common, but everything else is completely different, so the diff is unreabable.

-- 
Timur Tabi
Linux kernel developer at Freescale

Jonathan Nieder· Aug 27, 2010, 02:13 UTC · re: Timur Tabi · lore

Re: Treating similar blank lines between hunks as part of the hunks

Timur Tabi wrote:

Show 29 quoted lines

> git-diff will do this:
> 
>  A
> -B
> -C
> +F
> +G
> 
> -D
> +H
>  E
> 
> Instead, I would prefer it did this:
> 
>  A
> -B
> -C
> -
> -D
> +F
> +G
> +
> +H
> E
> 
> This second diff is easier to read, especially when the number of
> hunks is large, but the size of each hunk is small.  This can happen
> in source code when one function is replaced with a completely
> different function.

Not sure if it would help in your example, but have you tried "git diff --patience"?