Re: [PATCH v2] [PR debug/67192] Fix C loops' back-jump location

[Andreas Arnez <arnez@linux.vnet.ibm.com>]

On Thu, Oct 29 2015, Andreas Arnez wrote:

> On Thu, Oct 29 2015, Bernd Schmidt wrote:
>> [...]
>> i.e. the breakpoint on the code inside the loop is reached before the
>> while statement itself. This may be the expected behaviour with your
>> patch, but I'm not sure it's really desirable for debugging.
>
> [...]
> Maybe we could also improve the behavior of breaking on "while (1)" by
> generating a NOP for it?  Or by using the first loop body's token
> instead?

I've basically tried the latter, and it seems to work pretty well.  It
solves all the issues discussed in this mail thread, and I haven't found
any other issues with it.  I'll post the patch separately.

>> I'd suggest you commit your original patch to fix the
>> misleading-indent problem while we sort this out.
>
> I can certainly do that.  But note that the original patch does not
> solve the misleading-indent regression caused for f2() in the new
> version of pr67192.c.  Thus the PR is not really fixed by it.

I've slightly changed the test case for that one, so I'll repost it as
well before committing it.

--
Andreas