| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-8017392 | 6u65 | Poonam Bajaj Parhar | P3 | Closed | Fixed | b01 |
| JDK-8006351 | 6u60 | Robert Mckenna | P3 | Closed | Fixed | b01 |
| JDK-8005462 | 6u35 | Poonam Bajaj Parhar | P3 | Closed | Fixed | b52 |
The intent of the bug fix for
6523983: (process) Raise Windows pipe buffer size (win)
was that every system had at least a 4095 least common denominator
pipe buffer size, for compatibility and performance.
A testlet was added to test/java/lang/ProcessBuilder/Basic.java
to ensure that 4095 bytes could be written between subprocesses without deadlock.
However, at least some Windows systems (Windows 2000) inexplicably
appears to need an extra 24 bytes of slop to avoid deadlock.
The size of the pipe buffer as passed to CreateProcess should be raised by 24 bytes.
6523983: (process) Raise Windows pipe buffer size (win)
was that every system had at least a 4095 least common denominator
pipe buffer size, for compatibility and performance.
A testlet was added to test/java/lang/ProcessBuilder/Basic.java
to ensure that 4095 bytes could be written between subprocesses without deadlock.
However, at least some Windows systems (Windows 2000) inexplicably
appears to need an extra 24 bytes of slop to avoid deadlock.
The size of the pipe buffer as passed to CreateProcess should be raised by 24 bytes.
- backported by
-
JDK-8005462 (process) Raise Windows pipe buffer size an extra 24 bytes (win)
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
- relates to
-
JDK-6523983 (process) Raise Windows pipe buffer size (win)
-
- Resolved
-