Since there was some confusion about what we are talking about, see below. Obviously the diff output in my `1' test cases is prefereable. Note that `git diff' does the same thing as diff -p (and it doesn't even need a -p option to do it).
I also observe that by default, emacs wants to indent the label by 1 character - even though usually it likes to align labels to the LHS of the enclosing block. Presumably for this reason. Ian. $ cat t.c function(){ code; out: context1; context2; context3; context4; will be edited; context5; context6; context7; } $ perl -pe 's/will be/was/' <t.c >u.c $ perl -pe 's/out:/ $&/' <t.c >t1.c $ perl -pe 's/will be/was/' <t1.c >u1.c $ diff -up t.c u.c --- t.c 2018-11-27 16:14:33.614941668 +0000 +++ u.c 2018-11-27 16:15:39.020244882 +0000 @@ -5,7 +5,7 @@ out: context2; context3; context4; - will be edited; + was edited; context5; context6; context7; $ diff -up t1.c u1.c --- t1.c 2018-11-27 16:15:52.424512251 +0000 +++ u1.c 2018-11-27 16:15:57.064604805 +0000 @@ -5,7 +5,7 @@ function(){ context2; context3; context4; - will be edited; + was edited; context5; context6; context7; $ diff --git a/t.c b/t.c index e0a0315..d09f948 100644 --- a/t.c +++ b/t.c @@ -5,7 +5,7 @@ out: context2; context3; context4; - will be edited; + was edited; context5; context6; context7; diff --git a/t1.c b/t1.c index d7ef2c4..bc8bba1 100644 --- a/t1.c +++ b/t1.c @@ -5,7 +5,7 @@ function(){ context2; context3; context4; - will be edited; + was edited; context5; context6; context7; _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel