[Bug libstdc++/56430] In __airy: return-statement with a value, in function returning 'void'.

[Ed Smith-Rowland]

On 06/13/2013 04:30 AM, Paolo Carlini wrote:
> On 06/13/2013 02:38 AM, Paolo Carlini wrote:
>> If we really have to add a testcase - I'm not sure - please double 
>> check that it passes testing with -Wall, no unused vars.
> Patch as went in had still the testcase wrong, triggering at least 3 
> warnings with -Wall. All in all, I decided to also remove the 
> additional functions: it doesn't make sense to add *now* functions to 
> tr1, which otherwise is deeply in regression fixes only mode. And 
> certainly not under a completely unrelated PR.
>
> Paolo.
>
1. Fine. I get fixing just the PR and not conflating things.

2. How do you test with Wall?  Do you just test just the library with 
Wall or the whole build?
I've tried several 'make check-libstdc++ RUNTESTFLAGS="-Wall"', etc. and 
no dice.
Obviously I should have hit the test with wall.  Sorry.

3. I would like to implement TR29123 which adds the TR1 math functions 
to std.  I obviously would like to make cleanups to the algorithms there.
We could either copy the math libs to bits and maintain separate files 
for the same functions or we could keep the math headers in one place.  
If we don't add anything to the math it seems like splitting the 
implementation would be a waste.  There would be a lot of overlap.  
OTOH, cleaning up TR29123 in, say, bits and totally leaving tr1 alone 
would encourage people to switch up (especially with guidance to that 
effect).

4. I would like some way to add experimental support for new math 
functions that appear in std proposal papers (airy_ai, etc.).  I suppose 
tr2 or ext.  Any ideas which you'd prefer?