Something for the release notes? Gary
On Tue, Feb 13, 2018 at 11:29 AM, <ol...@apache.org> wrote: > Repository: httpcomponents-core > Updated Branches: > refs/heads/4.4.x 7bfffc547 -> 11909bdcc > > > ExpandableBuffer breaks on large response > > Ticket: https://issues.apache.org/jira/browse/HTTPCORE-513 > This closes #58 > > > Project: http://git-wip-us.apache.org/repos/asf/httpcomponents-core/repo > Commit: http://git-wip-us.apache.org/repos/asf/httpcomponents-core/ > commit/11909bdc > Tree: http://git-wip-us.apache.org/repos/asf/httpcomponents-core/ > tree/11909bdc > Diff: http://git-wip-us.apache.org/repos/asf/httpcomponents-core/ > diff/11909bdc > > Branch: refs/heads/4.4.x > Commit: 11909bdcc33fbc29037338d870dd2f8a4e507ec7 > Parents: 7bfffc5 > Author: imoldovan-intacct <imoldo...@intacct.com> > Authored: Mon Feb 12 17:02:06 2018 +0200 > Committer: Oleg Kalnichevski <ol...@apache.org> > Committed: Tue Feb 13 19:27:40 2018 +0100 > > ---------------------------------------------------------------------- > .../org/apache/http/nio/util/ExpandableBuffer.java | 14 +++++++++++++- > 1 file changed, 13 insertions(+), 1 deletion(-) > ---------------------------------------------------------------------- > > > http://git-wip-us.apache.org/repos/asf/httpcomponents-core/ > blob/11909bdc/httpcore-nio/src/main/java/org/apache/http/ > nio/util/ExpandableBuffer.java > ---------------------------------------------------------------------- > diff --git > a/httpcore-nio/src/main/java/org/apache/http/nio/util/ExpandableBuffer.java > b/httpcore-nio/src/main/java/org/apache/http/nio/util/ > ExpandableBuffer.java > index 7e06da4..74b5765 100644 > --- a/httpcore-nio/src/main/java/org/apache/http/nio/util/ > ExpandableBuffer.java > +++ b/httpcore-nio/src/main/java/org/apache/http/nio/util/ > ExpandableBuffer.java > @@ -116,7 +116,19 @@ public class ExpandableBuffer implements BufferInfo, > org.apache.http.nio.util.Bu > protected void expand() { > int newcapacity = (this.buffer.capacity() + 1) << 1; > if (newcapacity < 0) { > - newcapacity = Integer.MAX_VALUE; > + final int vmBytes = Long.SIZE >> 3; > + final int javaBytes = 8; // this is to be checked when the > JVM version changes > + @SuppressWarnings("unused") // we really need the 8 if we're > going to make this foolproof > + final int headRoom = (vmBytes >= javaBytes) ? vmBytes : > javaBytes; > + // Reason: In GC the size of objects is passed as int (2 > bytes). > + // Then, the header size of the objects is added to the size. > + // Long has the longest header available. Object header seems > to be linked to it. > + // Details: I added a minimum of 8 just to be safe and > because 8 is used in > + // java.lang.Object.ArrayList: private static final int > MAX_ARRAY_SIZE = 2147483639. > + // > + // WARNING: This code assumes you are providing enough heap > room with -Xmx. > + // source of inspiration: https://bugs.openjdk.java.net/ > browse/JDK-8059914 > + newcapacity = Integer.MAX_VALUE - headRoom; > } > expandCapacity(newcapacity); > } > >
