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

Re: MCH0601 in SendDoc at 4523 (v1.24)




This looks like a good fix to me...


On 9/16/2013 3:33 PM, Thomas Raddatz wrote:
Hi Scott,

What do you think about this issue? Any comments?

The problem only happens, when we send about 5000 requests in about 90 minutes to a Terracotta cache server. It does not occur on other request to e.g. web services. But take that just as a side note.

Anyway, the problem went away after I changed HTTPAPIR4 as described in my initial posting.

I am just curious whether or not you agree to my fix and if you want to adopt it to your code base.

Regards,

Thomas.

Am 28.08.2013 09:27, schrieb thomas.raddatz@xxxxxxxxxxx:
    Hi Scott,



Today I got a MCH0601 error in procedure SendDoc() at statement 4523. I debugged the problem and I found out that SendDoc() does not check the
    return value of comm_BlockWrite().



    In order to reproduce the error you may:



    a) Add a breakpoint to the first line of SendDoc().



    b) Call your test program.



c) Shutdown your (test) web server or wait for a timeout (broken pipe).



    d) Let the program continue.



Now wwLeft is incremented whereas wwPos is decremented and it is just a
    question of time until MCH0601 is sent. For now I fixed the problem
    like this:



         c                   eval      wwSent = comm_BlockWrite( peComm
         c                                                     : wwPos
         c                                                     : wwChunk
c : peTimeout
    )
          *  Monitor return code of comm_BlockWrite()
         c                   if        wwSent < 0
         c                   return    -1
         c                   endif
         c                   eval      wwLeft = wwLeft - wwSent

    Do you agree with me?



    Regards,



    Thomas.



-----------------------------------------------------------------------
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
-----------------------------------------------------------------------