RE: [j-t-c] problem in ajp_process_callback (in jk_ajp_common.c)

[GOMEZ Henri]

I know it well, and take a serious look even
if it will be seriously reworked in jk2...

-
Henri Gomez ___[_]
EMAIL : [EMAIL PROTECTED](. .) 
PGP KEY : 697ECEDD...oOOo..(_)..oOOo...
PGP Fingerprint : 9DF8 1EA8 ED53 2F39 DC9B 904A 364F 80E6 



-Original Message-
From: Kevin Seguin [mailto:[EMAIL PROTECTED]]
Sent: Friday, December 07, 2001 12:16 AM
To: Tomcat-Dev (E-mail)
Subject: [j-t-c] problem in ajp_process_callback (in jk_ajp_common.c)


in ajp_process_callback, here is the block of code that 
handles sending the
next chunk of data from the webserver to tomcat:

--- snip ---
case JK_AJP13_GET_BODY_CHUNK:
{
   unsigned len = (unsigned)jk_b_get_int(msg);
jk_log(l, JK_LOG_DEBUG, received 
JK_AJP13_GET_BODY_CHUNK,
len=%d\n, len);

if(len  AJP13_MAX_SEND_BODY_SZ) {
len = AJP13_MAX_SEND_BODY_SZ;
}
if(len  ae-left_bytes_to_send) {
jk_log(l, JK_LOG_DEBUG, len  
ae-left_bytes_to_send
(%d  %d)\n,
   len, ae-left_bytes_to_send);
len = ae-left_bytes_to_send;
}
   if(len  0) {
   len = 0;
   }

   /* the right place to add file storage for upload */
   if ((len = ajp_read_into_msg_buff(ae, r, msg, 
len, l)) = 0)
{
   r-content_read += len;
   return JK_AJP13_HAS_RESPONSE;
   }  

   jk_log(l, JK_LOG_ERROR, Error ajp_process_callback -
ajp_read_into_msg_buff failed\n);
   return JK_INTERNAL_ERROR;   
}
   break;
--- end snip ---

in this line:

   if ((len = ajp_read_into_msg_buff(ae, r, msg, 
len, l)) = 0)
{

shouldn't pmsg (the post message) be read into, not msg?  i 
think you only
run into this situation when the posted data doesn't fit into the first
message to tomcat, or a handling servlet doesn't make use of 
content-length
and tries to read more bytes than are available.  i'm not sure 
though...
the code is a little hard to follow :)

I've written that piece of code some months ago but it was modified and 
now I'm not sure (I feel you could be rigth). This code turned to be 
a real spaghetti and should we reworked. I'll work on it as soon
as Costin will finish refactory

anyway, i was running into a problem with an infinite loop in 
the ajp layer.
when i made this change, the problem went away.

can someone who knows this code better than i take a look?  thanks in
advance.

I'm on it :)

--
To unsubscribe, e-mail:   mailto:[EMAIL PROTECTED]
For additional commands, e-mail: mailto:[EMAIL PROTECTED]

[j-t-c] problem in ajp_process_callback (in jk_ajp_common.c)

[Kevin Seguin]

in ajp_process_callback, here is the block of code that handles sending the
next chunk of data from the webserver to tomcat:

--- snip ---
case JK_AJP13_GET_BODY_CHUNK:
{
unsigned len = (unsigned)jk_b_get_int(msg);
jk_log(l, JK_LOG_DEBUG, received JK_AJP13_GET_BODY_CHUNK,
len=%d\n, len);

if(len  AJP13_MAX_SEND_BODY_SZ) {
len = AJP13_MAX_SEND_BODY_SZ;
}
if(len  ae-left_bytes_to_send) {
jk_log(l, JK_LOG_DEBUG, len  ae-left_bytes_to_send
(%d  %d)\n,
   len, ae-left_bytes_to_send);
len = ae-left_bytes_to_send;
}
if(len  0) {
len = 0;
}

/* the right place to add file storage for upload */
if ((len = ajp_read_into_msg_buff(ae, r, msg, len, l)) = 0)
{
r-content_read += len;
return JK_AJP13_HAS_RESPONSE;
}  

jk_log(l, JK_LOG_ERROR, Error ajp_process_callback -
ajp_read_into_msg_buff failed\n);
return JK_INTERNAL_ERROR;   
}
break;
--- end snip ---

in this line:

if ((len = ajp_read_into_msg_buff(ae, r, msg, len, l)) = 0)
{

shouldn't pmsg (the post message) be read into, not msg?  i think you only
run into this situation when the posted data doesn't fit into the first
message to tomcat, or a handling servlet doesn't make use of content-length
and tries to read more bytes than are available.  i'm not sure though...
the code is a little hard to follow :)

anyway, i was running into a problem with an infinite loop in the ajp layer.
when i made this change, the problem went away.

can someone who knows this code better than i take a look?  thanks in
advance.

--
To unsubscribe, e-mail:   mailto:[EMAIL PROTECTED]
For additional commands, e-mail: mailto:[EMAIL PROTECTED]