Re: Setting breakpoints in HTTPAPI callbacks

Thanks, Charles. RDI/SEP is one of the "bigger debug hammers" I'm considering, but there are some setup issues involved there which is why I just want to make sure I'm not misunderstanding the green screen process.

As to whether the call looks strange, I didn't write it and cannot comment except to note that this is working code.  We're just trying to get into the callback procedure at runtime.  But I think it just means that there is no StartProc.


On 7/30/2020 10:25 AM, Charles Wilt wrote:

IIRC, I've debugged "incoming" callbacks using RDi and SEP without issue...

Can't recall using STRDBG & STRSRVJOB...

OTOH, that call looks funny given the PR, why are they passing *NULL for
peStartProc? Not sure that's valid...

D http_url_post_xml...
D PR 10I 0
D peURL 32767A varying const options(*varsize)
D pePostData * value
D pePostDataLen 10I 0 value
D peStartProc * value procptr
D peEndProc * value procptr
D peUsrDta * value
D peTimeout 10I 0 value options(*nopass)
/if defined(HTTP_ORIG_SHORTFIELD)
D peUserAgent 64A const options(*nopass:*omit)
D peContentType 64A const options(*nopass:*omit)
D peSOAPAction 64A const options(*nopass:*omit)
/else
D peUserAgent 16384A varying const
D options(*nopass:*omit)
D peContentType 16384A varying const
D options(*nopass:*omit)
D peSOAPAction 16384A varying const
D options(*nopass:*omit)

Charles

On Thu, Jul 30, 2020 at 8:13 AM Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx>
wrote:

Cross-posted to RPG400 and WEB400 lists.

A colleague is using Scott Klement's HTTPAPI routines and is running
into an issue. They have the following code:

rc = http_url_post_xml( URL
: %addr(SOAP) + 2
: %len(SOAP)
: *null
: %paddr(Incoming)
: %addr(call_status)
: HTTP_TIMEOUT
: HTTP_USERAGENT
: 'text/xml'
: %trim(WSWSUR));

The program runs in a never-ending batch job. They execute STRSRVJOB on
the job and STRDBG on the program, and then set a breakpoint on the
first line of the "Incoming" procedure. It looks like the procedure is
being executed because it writes to a file and records are indeed being
written, but the breakpoint is never hit.

Before we dig a lot deeper or start using bigger debug hammers, is this
an issue anyone else has seen? I was wondering if perhaps there are
multi-threading concerns with the HTTP APIs that might complicate
service jobs.

Thanks as always!