Re: [PATCH] SSL guide amendments (patch against 2.0 branch)
Site updated. Mike On Apr 15, 2004, at 5:09 PM, Oleg Kalnichevski wrote: Patch committed to CVS HEAD & 2.0 Oleg On Thu, 2004-04-15 at 17:20, Eric Johnson wrote: Oleg, Yes, of course. Sorry, I should have done that the first time. Attached (I hope). -Eric. Kalnichevski, Oleg wrote: Hi Eric Many thanks for taking time to correct my writing. All corrections make sense to me. (BTW, no need to be over-diplomatic. I am perfectly aware that my English has its limits, especially if I just type away. Just correct it. There'll be no questions asked). Do you still have the current CVS snapshot at your disposal? If yes, could you please recreate the patch with all those corrections, if that's not too much of a hassle? Oleg -Original Message- From: Eric Johnson [mailto:[EMAIL PROTECTED] Sent: Thursday, April 15, 2004 15:18 To: Commons HttpClient Project Cc: Daniel C. Amadei Subject: Re: [PATCH] SSL guide amendments (patch against 2.0 branch) Oleg, A few suggested edits I'm not a great editor myself (I frequently miss bevies of typos when my spouse asks me to review her writing), but since nobody else responded, I figured I would. Hopefully, my edits make sense. Oleg Kalnichevski wrote: __ ? sslguide2.patch Index: sslguide.xml === RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v retrieving revision 1.2.2.1 diff -u -r1.2.2.1 sslguide.xml --- sslguide.xml 21 Aug 2003 16:07:31 - 1.2.2.1 +++ sslguide.xml 15 Apr 2004 15:18:40 - @@ -240,6 +240,43 @@ + + + JSSE prior to Java 1.4 incorrectly reports socket timeout. + + + Prior to Java 1.4, in Sun's JSSE implementation, a read operation that has timed out incorrect + reports end of stream condition instead of throwing java.io.InterruptedIOException as expected. + HttpClient responds to this exception by assuming that the connection was dropped and throws a recoverable + HTTP exception: Error in parsing the status line from the response: unable to find line starting with "HTTP". + It should instead report "java.io.InterruptedIOException: Read timed out". If you see the "unable to find + line..." message when working with an older version of JDK and JSSE, it can be caused by the timeout + waiting for data and not by a problem with the connection. + + + Work-around: One possible solution is to increase the timeout value as the server is + taking too long to start sending the response. Alternatively you may choose to upgrade to Java 1.4 or + above which does not exhibit this problem. + + + The problem has been discovered and reported by Daniel C. Amadei. + + + + + + HttpClient does not work with IBM JSSE shipped with IBM Websphere Application Platform + + + Several releases of the IBM JSSE exhibit a bug that cause HttpClient to fail while detecting the size + of the socket send buffer (java.net.Socket.getSendBufferSize method throws java.net.SocketException: + "Socket closed" exception). + + + Solution: Make sure that you have all the latest fix packs applied. HttpClient users + have reported that IBM Websphere Application Server versions 4.0.6, 5.0.2.2, 5.1 do not exhibit the problem. + + __ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [PATCH] SSL guide amendments (patch against 2.0 branch)
Patch committed to CVS HEAD & 2.0 Oleg On Thu, 2004-04-15 at 17:20, Eric Johnson wrote: > Oleg, > > Yes, of course. Sorry, I should have done that the first time. > > Attached (I hope). > > -Eric. > > Kalnichevski, Oleg wrote: > > >Hi Eric > >Many thanks for taking time to correct my writing. All corrections make sense to > >me. (BTW, no need to be over-diplomatic. I am perfectly aware that my English has > >its limits, especially if I just type away. Just correct it. There'll be no > >questions asked). > > > >Do you still have the current CVS snapshot at your disposal? If yes, could you > >please recreate the patch with all those corrections, if that's not too much of a > >hassle? > > > >Oleg > > > >-Original Message- > >From: Eric Johnson [mailto:[EMAIL PROTECTED] > >Sent: Thursday, April 15, 2004 15:18 > >To: Commons HttpClient Project > >Cc: Daniel C. Amadei > >Subject: Re: [PATCH] SSL guide amendments (patch against 2.0 branch) > > > > > >Oleg, > > > >A few suggested edits I'm not a great editor myself (I frequently > >miss bevies of typos when my spouse asks me to review her writing), but > >since nobody else responded, I figured I would. > > > >Hopefully, my edits make sense. > > > >Oleg Kalnichevski wrote: > > > > > > > > __ > ? sslguide2.patch > Index: sslguide.xml > === > RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v > retrieving revision 1.2.2.1 > diff -u -r1.2.2.1 sslguide.xml > --- sslguide.xml 21 Aug 2003 16:07:31 - 1.2.2.1 > +++ sslguide.xml 15 Apr 2004 15:18:40 - > @@ -240,6 +240,43 @@ > > > > + > + > + JSSE prior to Java 1.4 incorrectly reports socket timeout. > + > + > + Prior to Java 1.4, in Sun's JSSE implementation, a read operation that has > timed out incorrect > + reports end of stream condition instead of throwing > java.io.InterruptedIOException as expected. > + HttpClient responds to this exception by assuming that the connection was > dropped and throws a recoverable > + HTTP exception: Error in parsing the status line from the response: unable > to find line starting with "HTTP". > + It should instead report "java.io.InterruptedIOException: Read timed out". > If you see the "unable to find > + line..." message when working with an older version of JDK and JSSE, it > can be caused by the timeout > + waiting for data and not by a problem with the connection. > + > + > + Work-around: One possible solution is to increase the > timeout value as the server is > + taking too long to start sending the response. Alternatively you may > choose to upgrade to Java 1.4 or > + above which does not exhibit this problem. > + > + > + The problem has been discovered and reported by Daniel C. Amadei. > + > + > + > + > + > + HttpClient does not work with IBM JSSE shipped with IBM Websphere > Application Platform > + > + > + Several releases of the IBM JSSE exhibit a bug that cause HttpClient to > fail while detecting the size > + of the socket send buffer (java.net.Socket.getSendBufferSize method throws > java.net.SocketException: > + "Socket closed" exception). > + > + > + Solution: Make sure that you have all the latest fix > packs applied. HttpClient users > + have reported that IBM Websphere Application Server versions 4.0.6, > 5.0.2.2, 5.1 do not exhibit the problem. > + > + > > > > > > __ > - > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [PATCH] SSL guide amendments (patch against 2.0 branch)
Oleg, Yes, of course. Sorry, I should have done that the first time. Attached (I hope). -Eric. Kalnichevski, Oleg wrote: Hi Eric Many thanks for taking time to correct my writing. All corrections make sense to me. (BTW, no need to be over-diplomatic. I am perfectly aware that my English has its limits, especially if I just type away. Just correct it. There'll be no questions asked). Do you still have the current CVS snapshot at your disposal? If yes, could you please recreate the patch with all those corrections, if that's not too much of a hassle? Oleg -Original Message- From: Eric Johnson [mailto:[EMAIL PROTECTED] Sent: Thursday, April 15, 2004 15:18 To: Commons HttpClient Project Cc: Daniel C. Amadei Subject: Re: [PATCH] SSL guide amendments (patch against 2.0 branch) Oleg, A few suggested edits I'm not a great editor myself (I frequently miss bevies of typos when my spouse asks me to review her writing), but since nobody else responded, I figured I would. Hopefully, my edits make sense. Oleg Kalnichevski wrote: ? sslguide2.patch Index: sslguide.xml === RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v retrieving revision 1.2.2.1 diff -u -r1.2.2.1 sslguide.xml --- sslguide.xml21 Aug 2003 16:07:31 - 1.2.2.1 +++ sslguide.xml15 Apr 2004 15:18:40 - @@ -240,6 +240,43 @@ + + + JSSE prior to Java 1.4 incorrectly reports socket timeout. + + + Prior to Java 1.4, in Sun's JSSE implementation, a read operation that has timed out incorrect + reports end of stream condition instead of throwing java.io.InterruptedIOException as expected. + HttpClient responds to this exception by assuming that the connection was dropped and throws a recoverable + HTTP exception: Error in parsing the status line from the response: unable to find line starting with "HTTP". + It should instead report "java.io.InterruptedIOException: Read timed out". If you see the "unable to find + line..." message when working with an older version of JDK and JSSE, it can be caused by the timeout + waiting for data and not by a problem with the connection. + + + Work-around: One possible solution is to increase the timeout value as the server is + taking too long to start sending the response. Alternatively you may choose to upgrade to Java 1.4 or + above which does not exhibit this problem. + + + The problem has been discovered and reported by Daniel C. Amadei. + + + + + + HttpClient does not work with IBM JSSE shipped with IBM Websphere Application Platform + + + Several releases of the IBM JSSE exhibit a bug that cause HttpClient to fail while detecting the size + of the socket send buffer (java.net.Socket.getSendBufferSize method throws java.net.SocketException: + "Socket closed" exception). + + + Solution: Make sure that you have all the latest fix packs applied. HttpClient users + have reported that IBM Websphere Application Server versions 4.0.6, 5.0.2.2, 5.1 do not exhibit the problem. + + - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: [PATCH] SSL guide amendments (patch against 2.0 branch)
Hi Eric Many thanks for taking time to correct my writing. All corrections make sense to me. (BTW, no need to be over-diplomatic. I am perfectly aware that my English has its limits, especially if I just type away. Just correct it. There'll be no questions asked). Do you still have the current CVS snapshot at your disposal? If yes, could you please recreate the patch with all those corrections, if that's not too much of a hassle? Oleg -Original Message- From: Eric Johnson [mailto:[EMAIL PROTECTED] Sent: Thursday, April 15, 2004 15:18 To: Commons HttpClient Project Cc: Daniel C. Amadei Subject: Re: [PATCH] SSL guide amendments (patch against 2.0 branch) Oleg, A few suggested edits I'm not a great editor myself (I frequently miss bevies of typos when my spouse asks me to review her writing), but since nobody else responded, I figured I would. Hopefully, my edits make sense. Oleg Kalnichevski wrote: >Folks, > >Daniel C. Amadei has kindly contributed a paragraph on recently >discovered problem with bogus error messages caused by a bug in older >versions of Sun JSSE > >I have also long promised to write a few words regarding the known >problems with IBM JSSE > >Please let me know what you think. I would also like to kindly ask >someone of our resident native English speakers to proof-read the >writing > >Cheers, > >Oleg >PS: Daniel, I changed your original text somewhat. Please let me know if >you agree with the changes > > > > > >Index: sslguide.xml >=== >RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v >retrieving revision 1.2.2.1 >diff -u -r1.2.2.1 sslguide.xml >--- sslguide.xml 21 Aug 2003 16:07:31 - 1.2.2.1 >+++ sslguide.xml 14 Apr 2004 20:47:48 - >@@ -240,6 +240,45 @@ > > > >+ >+ >+ Socket timeout not correctly reported when using oder versions of >JSSE > "oder" --> "older" For that matter, how about "JSSE prior to Java 1.4 incorrectly reports socket timeout." >+ (prior to Java 1.4) >+ >+ >+ There is a bug in older versions of Sun JSSE that causes timed out read >operation to report end of > "Prior to Java 1.4, in Sun's JSSE implementation, a read operation that has timed out incorrect reports end of" >+ stream condition instead of throwing java.io.InterruptedIOException as >expected. As a result if read >+ operation on a secure (SSL) connection times out (SO_TIMEOUT is set to a >positive value) HttpClient > Replace starting from "As a result...", "HttpClient responds to this exception by assuming that the connection was dropped and throws a recoverable..." >+ mistakingly assumes that the connection was dropped and throws a >recoverable HTTP exception: Error in >+ parsing the status line from the response: unable to find line starting >with "HTTP" instead of >+ "java.io.InterruptedIOException: Read timed out" as expected. If you get >this message when working with > ...starting with "HTTP". It should instead report "java.io.InterruptedIOException: Read timed out". If you see the "unable to find line..." message when working with... >+ an older version of JDK and JSSE, it can be caused by the timeout waiting >for data and not by a problem >+ with the connection. >+ >+ >+ Solution: One possible solution is to increase the timeout >value as the server is > "Solution" -- more like "Work-around" >+ taking too long to start sending the response. Alternatively you may choose >to upgrade to Java 1.4 or >+ above which does not exhibit this problem. >+ >+ >+ The problem has been discovered and reported by Daniel C. Amadei. >+ >+ >+ >+ >+ >+ HttpClient does not work with IBM JSSE shipped with IBM Websphere >Application Platform >+ >+ >+ There is a bug in several releases of IBM JSSE that causes HttpClient to >fail while detecting the size > "Several releases of the IBM JSSE exhibit a bug that cause HttpClient..." >+ of the socket send buffer (java.net.Socket.getSendBufferSize method throws >java.net.SocketException: >+ "Socket closed" exception). >+ >+ >+ Solution: Make sure that you have all the latest fix packs >applied. IBM Websphere >+
Re: [PATCH] SSL guide amendments (patch against 2.0 branch)
Oleg, A few suggested edits I'm not a great editor myself (I frequently miss bevies of typos when my spouse asks me to review her writing), but since nobody else responded, I figured I would. Hopefully, my edits make sense. Oleg Kalnichevski wrote: Folks, Daniel C. Amadei has kindly contributed a paragraph on recently discovered problem with bogus error messages caused by a bug in older versions of Sun JSSE I have also long promised to write a few words regarding the known problems with IBM JSSE Please let me know what you think. I would also like to kindly ask someone of our resident native English speakers to proof-read the writing Cheers, Oleg PS: Daniel, I changed your original text somewhat. Please let me know if you agree with the changes Index: sslguide.xml === RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v retrieving revision 1.2.2.1 diff -u -r1.2.2.1 sslguide.xml --- sslguide.xml21 Aug 2003 16:07:31 - 1.2.2.1 +++ sslguide.xml14 Apr 2004 20:47:48 - @@ -240,6 +240,45 @@ + + + Socket timeout not correctly reported when using oder versions of JSSE "oder" --> "older" For that matter, how about "JSSE prior to Java 1.4 incorrectly reports socket timeout." + (prior to Java 1.4) + + + There is a bug in older versions of Sun JSSE that causes timed out read operation to report end of "Prior to Java 1.4, in Sun's JSSE implementation, a read operation that has timed out incorrect reports end of" + stream condition instead of throwing java.io.InterruptedIOException as expected. As a result if read + operation on a secure (SSL) connection times out (SO_TIMEOUT is set to a positive value) HttpClient Replace starting from "As a result...", "HttpClient responds to this exception by assuming that the connection was dropped and throws a recoverable..." + mistakingly assumes that the connection was dropped and throws a recoverable HTTP exception: Error in + parsing the status line from the response: unable to find line starting with "HTTP" instead of + "java.io.InterruptedIOException: Read timed out" as expected. If you get this message when working with ...starting with "HTTP". It should instead report "java.io.InterruptedIOException: Read timed out". If you see the "unable to find line..." message when working with... + an older version of JDK and JSSE, it can be caused by the timeout waiting for data and not by a problem + with the connection. + + + Solution: One possible solution is to increase the timeout value as the server is "Solution" -- more like "Work-around" + taking too long to start sending the response. Alternatively you may choose to upgrade to Java 1.4 or + above which does not exhibit this problem. + + + The problem has been discovered and reported by Daniel C. Amadei. + + + + + + HttpClient does not work with IBM JSSE shipped with IBM Websphere Application Platform + + + There is a bug in several releases of IBM JSSE that causes HttpClient to fail while detecting the size "Several releases of the IBM JSSE exhibit a bug that cause HttpClient..." + of the socket send buffer (java.net.Socket.getSendBufferSize method throws java.net.SocketException: + "Socket closed" exception). + + + Solution: Make sure that you have all the latest fix packs applied. IBM Websphere + Application Server versions 4.0.6, 5.0.2.2, 5.1 have been reported to not exhibit the problem. HttpClient users have reported that IBM Websphere ... , and 5.1 do not exhibit the problem." + + - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[PATCH] SSL guide amendments (patch against 2.0 branch)
Folks, Daniel C. Amadei has kindly contributed a paragraph on recently discovered problem with bogus error messages caused by a bug in older versions of Sun JSSE I have also long promised to write a few words regarding the known problems with IBM JSSE Please let me know what you think. I would also like to kindly ask someone of our resident native English speakers to proof-read the writing Cheers, Oleg PS: Daniel, I changed your original text somewhat. Please let me know if you agree with the changes Index: sslguide.xml === RCS file: /home/cvspublic/jakarta-commons/httpclient/xdocs/sslguide.xml,v retrieving revision 1.2.2.1 diff -u -r1.2.2.1 sslguide.xml --- sslguide.xml 21 Aug 2003 16:07:31 - 1.2.2.1 +++ sslguide.xml 14 Apr 2004 20:47:48 - @@ -240,6 +240,45 @@ + + + Socket timeout not correctly reported when using oder versions of JSSE + (prior to Java 1.4) + + + There is a bug in older versions of Sun JSSE that causes timed out read operation to report end of + stream condition instead of throwing java.io.InterruptedIOException as expected. As a result if read + operation on a secure (SSL) connection times out (SO_TIMEOUT is set to a positive value) HttpClient + mistakingly assumes that the connection was dropped and throws a recoverable HTTP exception: Error in + parsing the status line from the response: unable to find line starting with "HTTP" instead of + "java.io.InterruptedIOException: Read timed out" as expected. If you get this message when working with + an older version of JDK and JSSE, it can be caused by the timeout waiting for data and not by a problem + with the connection. + + + Solution: One possible solution is to increase the timeout value as the server is + taking too long to start sending the response. Alternatively you may choose to upgrade to Java 1.4 or + above which does not exhibit this problem. + + + The problem has been discovered and reported by Daniel C. Amadei. + + + + + + HttpClient does not work with IBM JSSE shipped with IBM Websphere Application Platform + + + There is a bug in several releases of IBM JSSE that causes HttpClient to fail while detecting the size + of the socket send buffer (java.net.Socket.getSendBufferSize method throws java.net.SocketException: + "Socket closed" exception). + + + Solution: Make sure that you have all the latest fix packs applied. IBM Websphere + Application Server versions 4.0.6, 5.0.2.2, 5.1 have been reported to not exhibit the problem. + + - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]