[chuck-dev] Patches: compiler warnings and test programs

[Rasmus Kaj]

>>>>> "GW" == Ge Wang <gewang at CS.Princeton.EDU> writes:

 GW> Greetings!

Hi!

 GW> Thanks for the patches.  We have applied the larger one and also dealt
 GW> with what we could find of the rest of the compiler warnings
 GW> (-Wall).

Great news!

 GW> If you have v2 checked out, please update, make clean, and give
 GW> it a try and see if things work, including on 64-bit systems.
 GW> Please email me any errors or warnings.

I certainly will.  But for the moment I'm on vacation.  I can read and
write a little bit of mail, but won't be able to test it the coming
week.  I will take a look at it the week after that, though.

 GW> Thank you again for the patches.  We haven't applied the first one yet
 GW> because we haven't finalized what <<<foo>>> should output exactly, or
 GW> if there will be variants of <<< >>>.

 GW> Presently, <<<foo>>> prints the value of foo and its type.
 GW> <<<foo, bar> prints foo and bar without type.  So on for <<<>>>
 GW> with more arguments.  Some issues are (1) should <<<foo>>> print
 GW> the type? (2) should we make a new operator based on <<<>>> for
 GW> printing foo to arbitrary file handles (stdout, stderr, files,
 GW> etc.)?  We will likely bring back chucking to file handles (foo
 GW> => stdout), so <<<>>> won't be the only option.  Thoughts?

Myself, I never understood why "<<<foo>>>" should be used instead of
"foo => stdout" in the first place, and I would certainly be
interested in "chucking" data to arbitrary file descriptors (sockets
and named pipes comes to mind) ...

Given the existence of "foo => stdout", I would consider "<<<foo>>>"
beeing more of a debugging facility, and so thinks it seems reasonable
to include the type.

-- 
Rasmus Kaj --+-- rasmus at kaj.se --+-- http://www.stacken.kth.se/~kaj/
Money is the root of all evil, and man needs roots