On Sun, Apr 8, 2012 at 11:52 AM, Manuel López-Ibáñez
<lopeziba...@gmail.com> wrote:
> On 8 April 2012 18:35, Gabriel Dos Reis <g...@integrable-solutions.net> wrote:
>> On Sun, Apr 8, 2012 at 11:13 AM, Manuel López-Ibáñez
>> <lopeziba...@gmail.com> wrote:
>>> On 8 April 2012 06:09, Jason Merrill <ja...@redhat.com> wrote:
>>>> On 04/07/2012 06:29 PM, Manuel López-Ibáńez wrote:
>>>>
>>>>> +getenv_columns (void)
>>>>
>>>>
>>>> I had been thinking to check COLUMNS once at the beginning of compilation;
>>>> I
>>>> don't think the value can change while the compiler is running since we
>>>> don't respond to SIGWINCH. And let's use this value in
>>>> c_common_initialize_diagnostics, too.
>>>>
>>>
>>> To focus on the issue at hand, and independently of whether any bugs
>>> exist or not, and whether I am capable or not to fix them. What you
>>> are proposing is to change the current default of no line-wrapping to
>>> line-wrap at getenv("COLUMNS"), isn't it?
>>
>> yes.
>
> OK, so to do this, I have to figure out where the pretty-printer is
> initialized in each FE. For C++ there are two different places:
>
> init_error() initializes cxx_pp
global_dc is the global diagnostic context -- that is used by almost
every front-end (except Ada and Fortran?)
cxx_pp is the C++ front-end-specific global pretty printer: this is where
we dump C++ ASTs as strings. It is mostly used only by the various
tree -> string functions that are occasionally called from the debugger
or by the front-end for things like __PRETTY_FUNCTION__. It is NOT
for diagnostic
purposes. (it was never meant to survive so many releases.)
> whereas cxx_initialize_diagnostics seems to initialize a different
> pretty-printer for C++.
This is where we allocate a C++ pretty printer attached to the diagnostic
context passed in as a parameter. It is this pretty printer that is used by
the diagnostic machinery.
>
> What is the difference between the two?
>
> Cheers,
>
> Manuel.
