RE: HTTPAPI get_chunk_size returned -1

[Greg Wilburn <gwilburn@xxxxxxxxxxxxxxxxxxxxxxx>]

The update seems to have fixed our problem... we've been using the updated version for about a week and have not seen the issue.

Thanks Scott!

-----Original Message-----
From: RPG400-L [mailto:rpg400-l-bounces@xxxxxxxxxxxxxxxxxx] On Behalf Of Greg Wilburn
Sent: Wednesday, July 08, 2020 9:57 AM
To: RPG programming on IBM i <rpg400-l@xxxxxxxxxxxxxxxxxx>
Subject: RE: HTTPAPI get_chunk_size returned -1

Thanks Scott & Brad... This is a TLS/SSL connection (like everything nowadays).

The debug/trace is where I found the error.  I can upgrade to see if that fixes it first.  It is very random and seems to occur more often under heavier communications loads.

-----Original Message-----
From: RPG400-L [mailto:rpg400-l-bounces@xxxxxxxxxxxxxxxxxx] On Behalf Of Scott Klement
Sent: Tuesday, July 07, 2020 8:31 PM
To: rpg400-l@xxxxxxxxxxxxxxxxxx
Subject: Re: HTTPAPI get_chunk_size returned -1

Greg,

If this is during a TLS/SSL session, then please upgrade your HTTPAPI to 
the current version, there is a fix that has to do with TLS/SSL 
handshaking problems that could potentially cause this.

If this is non-TLS/non-SSL then I need a full debug/trace file to 
understand it better.   From what little you've provided, Brad Stone has 
it right...  a bad chunk size was sent from the server. But, there's a 
lot more detail in the debug/trace file, and some of it could confirm 
whether it's true, etc.

-SK


On 7/7/2020 12:40 PM, Greg Wilburn wrote:
> HTTPAPI Ver 1.39 released 2018-03-09
> NTLM Ver 1.4.0 released 2014-12-22
> OS/400 Ver V7R3M0
>
> <Partial/incomplete JSON document here>
>
> comm_blockread returned 3336
> get_chunk_size(): entered
> get_chunk_size returned -1
> http_close(): entered
>
>
> I'm getting this error (in the debug log) when receiving a JSON payload from a Web Service API.
>
> It seems to occur randomly.  I just want to be sure the problem is not on my side.  To me it appears that the API stops communicating.  But I'm being told that the issue cannot be reproduced.
>
> Thanks,
> Greg