On Sat, 25 Apr 2009, Andi Jahja wrote:
> The following entries shows us how my idea was used in Harbour without
> giving credits because "Idea Cannot Be CopyRighted". I know for sure the
> codes are not mine but I can assure you that the "Idea" to implement it
> in xHarbour was mine. This idea was first strongly rejected by Harbour
> developers, but I found it now implemented in Harbour. I know I'm not
> entitled to claim nor I am looking for "rewards" because truthfully I do
> not expect it. If people can use my idea, I am a very happy man
> eventhough it was implemented with many exceptions/comments. :-)
Yes it is. You added code with similar functionality to xHarbour.
I do not like this extension and I do not like that you have made it.
As the result some 3-rd party projects begin to use :-( it and it was
a problem in porting their code to Harbour so I decided to add such
extension though I still think it's bad idea.
Unfortunately I was not able to use xHarbour code. Just simply
I created small test code you can find in the ChangeLog with few
different bad cases which should be resolved by code:
proc main()
? YEAR( DATE() ), MONTH( DATE() ), DAY( DATE() ), TEST()
return
#pragma begindump
#include "hbapi.h"
/* are comments stripped? If not then HB_FUNC_STATIC( DATE )
in comment can create troubles. */
// HB_FUNC_STATIC( YEAR ) just like this one in C++ comments
#define MYHB_FUNC_STATIC( fun ) #fun // to test whole word tokens
/* COMMENT */HB_FUNC_STATIC( TEST )
{
hb_retc( "HB_FUNC_STATIC( DAY )->" MYHB_FUNC_STATIC( MONTH ) );
}
#pragma enddump
And tried to compile it using xHarbour.
It immediately exploited the problems in xHarbour.
AFAIR xHarbour is not able to correctly detect any of danger situation
in the above test code.
So I decided to not copy it from xHarbour but implement it myself
with some basic support for tokens. This implementation works correctly
and I think that xHarbour can benefit from it. If not they you should
update xHarbour to pass correctly above test.
In this particular case I many times wrote that it's very bad extension
and I haven't changed my opinion. IMHO it's only the source of problem
and I presented it on this and other forums many times so I though that
direct reference to you as the author of this idea will not be pleasure
for you. Looks that I was wrong. Of course I'll update the ChangeLog entry
and add an message that the idea of parsing the code inside #pragma begindump
to detect static C function was 1-st time implemented by Andi Jahja in
xHarbour.
Anyhow here I haven't took any source code or algorithms from xHarbour
so it's not comparable situation but I think it's good practice to make
leave such information especially now when I know that you have nothing
against binding your name with with extension.
best regards,
Przemek
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
