> On Fri, 2008-02-08 at 16:10 +1300, Amos Jeffries wrote:
>
> The "same()" operations used in those tests may be different from a
> "same_MD5s()" operation.
>
> Try inserting an empty line in a middle of a function and recompile. You
> will get a different MD5. Stripping the executable does not help in my
> tests......................
Yes this is true. I tried exactly the same tests to evaluate the
formatter, but they did not work. I believed that striping the code will
solve the problem but nothing...
In general I was not able to find a safe way to test it. I am just looking
the diffs searching for bad formated code ...
(Also I must note here that the bazaar repository was very-very helpful
here, branches/revert/diff etc on local repositories are really fast, but
OK it is an other story :-) )
>................... Stripping empty lines fron sources does not help if
the formatter
> moves brackets and such, changing the number of lines. Does our
> formatter do that?
>
Yes it does. For example the code:
void some_function(){
}
convert to:
void some_function()
{
}
