freepooma-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [pooma-dev] goofy Domain constructors


From: Julian C. Cummings
Subject: RE: [pooma-dev] goofy Domain constructors
Date: Mon, 2 Apr 2001 15:55:18 -0700

OK, guys.  Thanks for the comments.  I agree
that adding "this->" all over the place is a 
nuisance and is unwarranted in this case.  As
Scott points out, our coding style eliminates 
concerns of name collisions.  I can strip this
change out of my changes to the Domain sources.
Beyond this, is the change from enums to static
typed data OK?

The HP compiler only marks these things as 
"future" errors, so they do not inhibit the
compilation.  The messages refer to a specific 
section of the ANSI standard, so I will have a 
look at that.  If HP is indeed wrong in their
interpretation of the standard, I should let 
them know.  Meanwhile, I think there is a flag
to turn off their Koenig name lookup, so that
the flood of "future" errors will go away.

BTW, the HP compiler did point out some minor 
bugs that I fixed yesterday.  And I think the 
elimination of enums and the associated casting
of types cleans things up a bit.  So it hasn't 
been a completely worthless exercise attempting 
to run Pooma II through their compiler.  I hope
aCC will eventually be able to handle Pooma, so
we can use it on our parallel machines here.

-- Julian C.


-----Original Message-----
From: Mark Mitchell [mailto:address@hidden Behalf Of Mark
Mitchell
Sent: Monday, April 02, 2001 10:18 AM
To: address@hidden
Cc: address@hidden; address@hidden
Subject: Re: [pooma-dev] goofy Domain constructors


>>>>> "Scott" == Scott Haney <address@hidden> writes:

    >> However, your changes still aren't a bad idea; the standard
    >> says that if there was a global `i', the code in B would use
    >> that *rather* than the base class version.

    Scott> I think these changes are a bad idea.

    Scott> First, as Mark points out, data members should be
    Scott> visible. Moreover, we add the suffix _m to data members

Given this naming convention, I agree with Scott.

    Scott> with a global. I don't not want to see a bunch of
    this-> that show up because of an HP compiler bug.

Right.  

Note that I'd feel differently if Blanca used aCC -- but as far as I
know they don't.

--
Mark Mitchell                   address@hidden
CodeSourcery, LLC               http://www.codesourcery.com

reply via email to

[Prev in Thread] Current Thread [Next in Thread]