From: Sonali A. Kale (sonali@pspl.co.in)
Date: Thu Jan 30 2003 - 17:53:49 IST
Harold,
Thanks for the feedback. We are working on adding
streaming support for this parser and will definitely solve
the namespace related issues in the coming version.
Let me know if you have any other feedback to be
incorporated in the new version.
- Thanks,
Sonali
----- Original Message -----
From: "Harold Ravlin" <hr@sabik.ncsa.uiuc.edu>
To: <sonali@pspl.co.in>
Sent: Thursday, January 30, 2003 3:33 AM
Subject: Re: Status of C++ parser for VOTable?
> Sonali,
>
> Thanks for the source code. I was able to compile it and run the test
> program under gcc 3.1.1.
>
> However, a namespace issue came up when I tried to integrate it into an
> AIPS++ program.
>
> For instance, both the VOTableParser and AIPS++ define Bool, True and
> False:
>
> VOTableParser
> enum Bool {
> True = true,
> False = false,
> Null = 3 // Not NULL since it's value is same as false.
> };
>
> AIPS++
>
> typedef bool Bool;
> const Bool True = true;
> const Bool False = false;
>
> Have you considered using a namespace to minimize issues like this?
>
>
> There are also class name conflicts. For example:
>
> VOTableParser
>
> class Link {
>
> public:
> ...
>
> AIPS++
>
> template<class t> class Link {
> protected:
> t store;
> Link<t> *Next;
> Link<t> *Prev;
> public:
> ...
>
>
> (Last summer, when I wrote a test program to convert from VOTable files to
> AIPS++ tables, I used names like votLink, etc. to minimize possible
> conflicts).
>
>
> I'm also concerned about using filenames like 'error.h'. Wouldn't it be
> possible for the compiler to grab /usr/include/error.h rather than your's?
>
>
>
> Are there any plans to use namespaces with the VOTableParser?
>
>
> Thanks,
>
> Harold
>
This archive was generated by hypermail 2.1.4 : Thu Jan 30 2003 - 23:30:01 IST