Re: [PATCH] Negative repeat count for 'x' command

[Toshihito Kikuchi <k.toshihito@yahoo.de>]

Hi Pedro,

I've sent V3 patch today.

> BTW, aren't these tests dependent on the host's charset?  (show host-charset)

Thank you for pointing this out. Yes, then I added 'set host-charset ASCII'
before running string testcases.

> An idea to generalize the x/i tests to all archs would be to
> let go of the asm, and instead write the test function in C.  Then
> you'd first use forward x/i to store a few line's instructions in
> a list/array, and afterwards you'd disassemble backwards, comparing
> with the expected instructions stored in the stored list/array.
> I wonder whether that'd work.

This was great suggestion.  I realized .exp test scripts can support complex
logics more than expected.  I added a new section 'backward disassemble general'
to do that.

A new section "address zero boundary" is also added to test around address zero.
 Since I don't have non-MMU machines, I verified it using by setting
/proc/sys/vm/mmap_min_addr to 0.

Thanks,
Toshihito