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

Re: EXAMPLE18 fails in LIBHTTP 1.17 and 1.18



Take a look at the code in the CCSIDR4 member included with HTTPAPI, and 
the XML_Xlate() stuff.  Currently, there are 5 different translation 
tables set up in HTTPAPI, for the control data (2 tables, one for each 
direction) the POST data (which I think also has 2 directions, though I 
can't think why it'd ever need to), and the XML data (to support 
arbitrary input CCSIDs in HTTPXMLR4).

You'd need to add one more, and the routines needed to support it.  It'd 
trnaslate from CCSID 1200 or 13488 (depending on what the old versions 
of OS/400 support?) to the job's CCSID (special value of 0).

Then replace the %char() translations with a call to the new routine you 
add to CCSIDr4.


Scott Schollenberger wrote:
> Scott --
> 
> I'm not sure I could get you access to our V5R1 server, but if you could
> offer some suggested code changes using iconv() I would be happy to try
> them out and provide feedback on the process.  I was looking at your
> iconv() example in the midrange archives
> (http://archive.midrange.com/rpg400-l/200403/msg00085.html) for some
> direction.  I don't have any experience with iconv() so I wasn't quite
> sure how to approach the change in HTTPXMLR4 even after reading that
> post.
> 
> -----
> Scott A. Schollenberger
> Director, R&D
> Schools Division
>    
> 2011 Renaissance Blvd., Suite 100
> King of Prussia, PA 19406
> Phone: 610-239-9988 x305
> Fax: 610-239-9995
> Email: sschollenberger@xxxxxxxxxxxxxxxxxx
> 
> -----Original Message-----
> From: ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
> [mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx] On Behalf Of Scott
> Klement
> Sent: Friday, April 27, 2007 7:40 PM
> To: HTTPAPI and FTPAPI Projects
> Subject: Re: EXAMPLE18 fails in LIBHTTP 1.17 and 1.18
> 
> Hi Scott,
> 
> That APAR explains why we're getting the strange "Call stack not found" 
> message, but doesn't explain why the %char(%subst()) combination doesn't
> work to begin with.
> 
> I wish I had access to a V5R1 system to try this with.  I don't suppose
> anyone would be willing to give me access (including enough access to
> install HTTPAPI, EXPAT, and to experiment with running HTTPAPI and
> changing/compiling the code, etc...)?
> 
> I'm sure I could work around the problem by using iconv(), but seems
> like a lot of mucking around for something that's SUPPOSED to work in
> RPG...
> 
> 
> 
> Scott Schollenberger wrote:
>> Scott --
>>
>> This is the same problem.  I changed the name variable to xname as you
>> had suggested and was able to see the value assigned to xname and then
>> the failure occurs.
>>
>> newElem.name = %char( %subst( name:1:len));
>>
>> It appears something in the V5R1 RPG runtime support doesn't like this
>> statement now that name is a UCS-2 variable rather than just being a
>> character variable in 1.15.
>>
>> I found one APAR for the MSGCPF2479 that seemed related.  But it
> didn't
>> offer a solution or workaround for V5R1.
>>
>>
> http://www-912.ibm.com/n_dir/nas4apar.nsf/c79815e083182fec862564c00079d1
> 17/e6eb28d38195f35786256d6c003cb3a4?OpenDocument&Highlight=2,MSGCPF2479
> -----------------------------------------------------------------------
> This is the FTPAPI mailing list.  To unsubscribe, please go to:
> http://www.scottklement.com/mailman/listinfo/ftpapi
> -----------------------------------------------------------------------
> -----------------------------------------------------------------------
> This is the FTPAPI mailing list.  To unsubscribe, please go to:
> http://www.scottklement.com/mailman/listinfo/ftpapi
> -----------------------------------------------------------------------

-----------------------------------------------------------------------
This is the FTPAPI mailing list.  To unsubscribe, please go to:
http://www.scottklement.com/mailman/listinfo/ftpapi
-----------------------------------------------------------------------