RE: Re: QUSRTOOL and asyncIO.h

midrange-l-request@xxxxxxxxxxxx wrote:

>   6. Re: QUSRTOOL and asyncIO.h (Vernon Hamberg)
>
>The QATTINFO member for SCO refers to Openness Includes for 5763SS1 - 
>what is the last release that used that number - I think it's V3R1. 
>You'd need to get one of those, maybe V3R2.

Vern:

I saw that and figured I'd chance that it hadn't been updated. Unfortunately, I 
didn't have anything older than V5R1 to compare back-levels against. And the 
coding in the Apache module gave some hope that it wasn't totally superseded by 
new OS/400 function. Without seeing what the original include had, I couldn't 
tell what might be missing in recent OS/400.

>At any rate, the info also refers to C/400, which is long gone. The 
>trouble with QUSRTOOL is, QSYSINC has changed drastically but 
>QUSRTOOL has not been updated.

Although infrequent, there are some changes I've seen over time. It's perhaps 
always a new tool every once in a while rather than any updates/fixes to older 
ones. In this case, I didn't know how old this one was. It's possibly been 
around a long time but I never looked into it because the sites I was at never 
kept a C compiler around.

It's kind of disappointing now that C is (almost) always available but IBM 
examples are long obsolete.

>   A couple options - look in the current 
>QSYSINC for similar functions and defines (didn't find any when I 
>tried this one) or find similar functioning includes and convert to 
>names in that include. There is a SOCKET include, as you probably 
>know, in QSYSINC/SYS, and this would probably have the current functionality.

If nothing else comes up, I'll first create a bogus asyncIO.h and look at what 
references are missing as the module compiles. I should be able to reconstruct 
something step by step.

>Maybe a PMR is in order, but I would not hold my breath.

I don't recall much PMR/service support for QUSRTOOL. That's always bothered me 
since it provides a common IBM-controlled set of objects that would allow 
communication between IBM and developers. It ought reduce the number of times a 
developer would send their objects to IBM for support and speed up replies. 
With fewer examples, developers start with far more guesses than they should.


>BTW, just for fun, old manuals can still be found by digging around in
>
>http://publib.boulder.ibm.com/cgi-bin/bookmgr/LIBRARY
>
>I found some V3R1 stuff there, even.

Yeah, I try to keep those links around and browse through there from time to 
time. Good idea, though, in this case. Maybe an old manual has a related 
compiler listing buried somewhere. There were no hits via Google. But Google 
wouldn't necessarily index deep enough.

Thanks for the ideas.

Tom Liotta

>At 11:51 PM 10/15/2005, you wrote:
>
>>I was going to look over a QUSRTOOL example, the TSO sockets tool, 
>>and found that the compile of one of the C modules failed with this error:
>>
>>36 |#include <asyncIO.h>
>>*=SEVERE==========> CZM0296 #include file <asyncIO.h> not found.
>>
>>  QSYSINC has no asyncIO.h in V5R1, R2 or R3. And surprise -- 
>> InfoCenter has no search references to such an include, although 
>> there are a couple comments in a couple Apache source modules to asyncIO.
>>
>>Does no one compile things from QUSRTOOL anymore?
>>
>>  Anyway, does anyone out there have an example of asyncIO.h that 
>> might be appropriate?
>>
>>I suspect it doesn't exist for OS/400, but I might as well ask.