Re: [PATCH v2 00/19] Various DWARF piece fixes

[Simon Marchi <simon.marchi@polymtl.ca>]

On 2017-05-09 19:45, Andreas Arnez wrote:
> This series fixes various issues with DWARF piece handling.
> Specifically it almost rewrites the functions read_pieced_value and
> write_pieced_value in multiple steps.  Test cases are added as well.
> 
> Version 1 is here:
> 
>   https://sourceware.org/ml/gdb-patches/2017-04/msg00177.html
> 
> Changes from version 1 include:
> 
> * Split up some patches further.
> 
> * Added two more fixes for memory pieces.
> 
> * Added a "merge" patch at the end.
> 
> * Introduced get_endianness convenience proc for test suite.
> 
> * Minor test case improvement.
> 
> * Comment- and some general readability improvements.

Hi Andreas,

I went over the patches that Yao hasn't replied on, namely 13, 14, 16, 
17 and 19, it all looks good to me (note the little comment on #17).  
Since it's some tricky code, I wasn't always able to convince myself of 
the correctness by just looking at it.  But I went through the test 
cases by hand, when applicable, and it gives me enough confidence that 
the code is correct (or at least more correct than the current code :)).

I'm not sure I like the merge of the two functions, already scary by 
themselves, in one scarier monster, but I understand the downsides of 
having two separate functions, so I'm ok to go with it.

I'll let Yao decide if he wants to go himself through the remaining 
patches.

Simon