harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Li Jing Qin (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HARMONY-6259) [classlib][luni] PipedInputStream.read throw different info when IOException is raised
Date Tue, 14 Jul 2009 02:50:14 GMT

    [ https://issues.apache.org/jira/browse/HARMONY-6259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12730656#action_12730656
] 

Li Jing Qin commented on HARMONY-6259:
--------------------------------------

Thanks Tim.

!isClosed check is not redundant, the reason is:
1. PipedInputStream and PipedOutputStream are often on the different thread
2. isConnected is set to true by the PipedOutputStream.connect() and will
never set back to false. (At first this seems odd to me. That means input
stream can connect to output stream only once. I have tested on the RI and
it is correct unfortunately)
3. When PipedOutputStream close, it only change isClose to false

Since HARMONY-6259 and HARMONY-6260 are fixes on the same file, should I put
them together, create a new diff?




-- 
Yours sincerely,
Charles Lee


> [classlib][luni] PipedInputStream.read throw different info when IOException is raised
> --------------------------------------------------------------------------------------
>
>                 Key: HARMONY-6259
>                 URL: https://issues.apache.org/jira/browse/HARMONY-6259
>             Project: Harmony
>          Issue Type: Bug
>          Components: Classlib
>            Reporter: Li Jing Qin
>            Assignee: Tim Ellison
>             Fix For: 5.0M11
>
>         Attachments: HARMONY-6259.diff, HARMONY-6259.v2.diff
>
>
> When the write end is closed unexpectedly, RI returns "Write end dead". Harmony do not
differentiate the "Pipe is broken" and "Write end dead"

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message