It's Oracle Java SE 8. Our customers will expect us to support that as long as the minimum requirement of our product is Java 8. We also implicitly thought that "HttpClient 5.4 requires Java 1.8 or newer" means that Oracle JREs of that version are supported. If this is not the case, we will have to delay updating to version 5.4 until we'll have removed support for Java 8 from our product.
Bernhard -----Original Message----- From: Oleg Kalnichevski <ol...@apache.org> Sent: Tuesday, October 8, 2024 11:08 To: HttpClient User Discussion <httpclient-users@hc.apache.org>; Bernhard Fey <bernhard....@realobjects.com> Subject: Re: Exception in version 5.4 with Java 8, potential bug On 08/10/2024 10:44, Bernhard Fey wrote: > We have reproduced the issue on Oracle Java SE 8, in my case "Oracle > Corporation Java HotSpot(TM) 64-Bit Server VM 1.8.0_411-b09". > We have not tried OpenJDK. Sorry for missing this detail in my initial mail. > > Bernhard > > Bernhard HttpClient expects the TLS half-close policy to be supported by the JRE for some of its features such as out of sequence response detection to work correctly. If we blindly ignore UnsupportedOperationException here things may still break elsewhere. Is there any particular reason for using a JRE with an incomplete TLS support? Oleg PS: please make sure you are subscribed to the list to receive response messages. > -----Original Message----- > From: Oleg Kalnichevski <ol...@apache.org> > Sent: Tuesday, October 8, 2024 10:31 > To: HttpClient User Discussion <httpclient-users@hc.apache.org> > Subject: Re: Exception in version 5.4 with Java 8, potential bug > > On Tue, 2024-10-08 at 08:09 +0000, Bernhard Fey wrote: >> Hello, I hope this is the right place to report potential bugs. >> >> When updating to version 5.4 our Java 8 tests fail with the following >> exception: >> >> java.lang.UnsupportedOperationException: method is not supported >> because of the TLS half-close policy >> at >> sun.security.ssl.SSLSocketImpl.shutdownInput(SSLSocketImpl.java:765) >> at >> sun.security.ssl.SSLSocketImpl.shutdownInput(SSLSocketImpl.java:743) >> at >> org.apache.hc.core5.http.impl.io.BHttpConnectionBase.close(BHttpConne >> ctionBase.java:255) >> at >> org.apache.hc.core5.http.impl.io.DefaultBHttpClientConnection.close(D >> efaultBHttpClientConnection.java:71) >> at >> org.apache.hc.client5.http.impl.io.DefaultManagedHttpClientConnection >> .close(DefaultManagedHttpClientConnection.java:176) >> at >> org.apache.hc.core5.pool.PoolEntry.discardConnection(PoolEntry.java:1 >> 80) >> at >> org.apache.hc.core5.pool.StrictConnPool$PerRoutePool.shutdown(StrictC >> onnPool.java:839) >> at >> org.apache.hc.core5.pool.StrictConnPool.close(StrictConnPool.java:142 >> ) >> at >> org.apache.hc.client5.http.impl.io.PoolingHttpClientConnectionManager >> .close(PoolingHttpClientConnectionManager.java:277) >> at >> org.apache.hc.client5.http.impl.classic.InternalHttpClient.close(Inte >> rnalHttpClient.java:198) >> at >> org.apache.hc.client5.http.impl.classic.InternalHttpClient.close(Inte >> rnalHttpClient.java:188) >> at J8HttpsErrorTest.main(J8HttpsErrorTest.java:21) >> >> It seems to be a bug, but please let me know if it is a mistake on our >> side. >> >> I have put together the following test class, which fails in Java 8 >> with the above exception, but finishes correctly with Java 11 (tested >> on Windows 10, but the Java 8 exception also occurs on Linux in our >> integration): >> > > ``` > Amazon.com Inc. OpenJDK 64-Bit Server VM 25.342-b07 Sucessfully loaded: > <!doctype html><html itemscope="" > itemtype="http://schema.org/WebPage" lang="de-CH"><head><meta > content="text/html; charset=UTF-8" http-equiv="Content-Type"><meta > content="/images/branding/googleg/1x/googleg_standard_color_128dp.png" > itemprop="image"><title>Google</title><script > nonce="sTJzMJQONf2G0BTpBOe_Tw">(function(){var _g={kEI:'9ewEZ_e- > I4vShbIPnNiauAk',kEXPI:'0,3700316,633,435,447881,648,90132,2872,2891,43 > 028,30022,16105,203622,6700,126319,8155,23350,22436,9779,12414,50243,76 > 209,15816,1804,21012,26070,1632,13496,15783,12989,8789,5208515,9465,102 > 3,5991408,2841102,880,1,38,6,29,2,69,14,3,8,1,3,9,68,4,11,3,7440095,205 > 39789,1007,15665,43887,3,1603,3,2124363,230293 >>>> > <<< > ``` > > The test case passes for me in my local environment. > > I used this JRE to test > > ``` > openjdk version "1.8.0_342" > OpenJDK Runtime Environment Corretto-8.342.07.4 (build 1.8.0_342-b07) OpenJDK > 64-Bit Server VM Corretto-8.342.07.4 (build 25.342-b07, mixed > mode) > ``` > > Oleg > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org > For additional commands, e-mail: httpclient-users-h...@hc.apache.org > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org > For additional commands, e-mail: httpclient-users-h...@hc.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: httpclient-users-unsubscr...@hc.apache.org For additional commands, e-mail: httpclient-users-h...@hc.apache.org