ASCIIZ PTR accepts STRPTR

[Egbert Zijlema]

Hi fellows,

Consider this as an odd question, if you want, but I've encountered some 'strange' behaviour.
In TYPE MSGBOXPARAMS (file: Win32api.inc) some items have been declared AS ASCIIZ PTR.
For instance: lpszText (the message) and lpszCaption (the caption).
However, when I declare (in my program) message and caption AS ASCIIZ strings and pass them
as STRPTR(szMessage) and STRPTR(szCaption) respectively, the following compiler error occurs:
"Dynamic string variable expected".
Okay, I now use dynamic strings instead and, indeed, the program works. But is'nt this controversial?
It puzzles me since a day or two. Does someone (Lance or Tom, perhaps?) have an explanation for this?
And, finally, what should I do to enable the use of genuine ASCIIZ strings?

Thanks very much!

------------------
mailto:[email protected][email protected]</A>
www.basicguru.com/zijlema/

[Lance Edmonds]

STRPTR() returns a pointer to the dynamic string data, but since the argument you tried to use was an ASCIIZ, you received a compile-timer error.

So, if your want to use an ASCIIZ string, you just need to use VARPTR() to return a pointer instead of STRPTR().

IOW, you could do either of the following depending on whether you are using a dynamic string or an ASCIIZ/Fixed-length:

DIM A$, B AS ASCIIZ * %BufLen
udt.param = STRPTR(A$)
udt.param = VARPTR(b)

------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Lance Edmonds]

One more note:

Internally, dynamic strings have a trailing CHR$(0) appended so that they can be passed to API functions that expect nul-terminated strings. This should explain why you can pass a STRPTR() where an ASCIIZ (pointer) was expected.

------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Egbert Zijlema]

Thank you Lance!
As far as I understand your message, it does not really matter whether I use STRPTR(DyNamicString) or VARPTR(szAsciizString)?

BTW: Since today the 'front page' of this site pops up with yellow lightbulbs only, as if every forum contained new postings.
In the Forum pages themselves I don't see red folder symbols, despite the fact there are new postings. Any idea?

------------------
mailto:[email protected][email protected]</A>
www.basicguru.com/zijlema/

[Eric Pearson]

The forums are working normally for me. Using Netscape 4.77.

-- Eric

[Bern Ertl]

You might have a cookie problem (filter)?


------------------
Bernard Ertl

[Egbert Zijlema]

Originally posted by Egbert Zijlema:
BTW: Since today the 'front page' of this site pops up with yellow lightbulbs only, as if every forum contained new postings.
In the Forum pages themselves I don't see red folder symbols, despite the fact there are new postings. Any idea?

Forget it, folks. It indeed was a cookie problem. For one or another unknown reason "BypassCookie=" was set to "true".
I did'nt do it myself, so one final question: how can those accidents occur?

------------------
mailto:[email protected][email protected]</A>
www.basicguru.com/zijlema/

[This message has been edited by Egbert Zijlema (edited August 23, 2001).]

[Lance Edmonds]

As far as I understand your message, it does not really matter whether I use STRPTR(DyNamicString) or VARPTR(szAsciizString)?

Yes, but the dynamic string must have content (not empty/null)...


------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Egbert Zijlema]

originally posted by lance edmonds:
yes, but the dynamic string must have content (not empty/null)...

it is'nt. see: http://www.powerbasic.com/support/pb...ad.php?t=23111

------------------
mailto:[email protected]egbert.zijl[email protected]</a>
www.basicguru.com/zijlema/

[this message has been edited by egbert zijlema (edited august 23, 2001).]

[Lance Edmonds]

In your case, you are setting a member of a UDT destined for an API call, so the API itself will very likely test for a null-pointer and react accordingly.

However, when dealing with exported PowerBASIC Subs/Functions that receive pointers (for example, in custom control libraries, etc), it may be wise to test that the pointer is non-zero before using it, but it will depend on how the pointer is going to be used. For example, if it is getting passed to an API that accepts null-pointers then there is no need to test it your code, but if the target memory is to be directly accessed... testing will help avoid a GPF.


------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Kev Peel]

You should never pass an empty dynamic string to an API function that expects an ASCIIZ string. I figured this out after my program was displaying strange "random" text in various windows, and it was all due to using empty dynamic strings (and it didn't cause any GPF either).

Regards,

------------------
Kev G Peel
KGP Software, Bridgwater, UK.
mailto:[email protected][email protected]</A>

[Lance Edmonds]

More by good luck than good design I would suggest...

------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Kev Peel]

------------------
Kev G Peel
KGP Software, Bridgwater, UK.
mailto:[email protected][email protected]</A>

[Tom Hanlin]

Depends on the API. In general, you should only pass a null ASCIIZ pointer
(or empty dynamic string, which is the same thing) if the function is
documented to accept a null pointer. Otherwise, it may just go boom.
The Windows APIs do a bare minimum of error checking (frequently less).

------------------
Tom Hanlin
PowerBASIC Staff