You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pig.apache.org by "Olga Natkovich (JIRA)" <ji...@apache.org> on 2008/05/19 19:59:56 UTC

[jira] Created: (PIG-243) Make pig work on Windows

Make pig work on Windows
------------------------

                 Key: PIG-243
                 URL: https://issues.apache.org/jira/browse/PIG-243
             Project: Pig
          Issue Type: Bug
            Reporter: Olga Natkovich


Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/30/08 10:23 AM:
----------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing '\r' (0x0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing '\r' |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Pi Song (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12598128#action_12598128 ] 

Pi Song commented on PIG-243:
-----------------------------

Like Alan said, our developers don't use Windows so it's difficult to test. Is it possible to also run tests on windows as nightly build?

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Daniel Dai updated PIG-243:
---------------------------

    Attachment: PIG_243.patch

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/28/08 12:54 AM:
------------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing need further work. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/28/08 12:52 AM:
------------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing need further work. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi, |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Patch for PIG under cygwin
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jianyong Dai updated PIG-243:
-----------------------------

    Status: Patch Available  (was: Open)

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600813#action_12600813 ] 

Olga Natkovich commented on PIG-243:
------------------------------------

Thanks, Daniel. Changes look good. Once concern I have is with changes to PigStorage since it adds a little bit of overhead to each record processing. However, since I assume that String caches the length, the overhead should be minimal. Could you run a quick test that reads a million lines with and without the extra processing and see what the time difference is. Thanks.

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/28/08 9:03 AM:
-----------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Daniel Dai updated PIG-243:
---------------------------

    Attachment: cygpath.patch

Fix the remaining 4 unit test error under cygwin

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: cygpath.patch, PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12601242#action_12601242 ] 

Olga Natkovich commented on PIG-243:
------------------------------------

Patch committed. Thanks Daniel for contributing!

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/28/08 12:53 AM:
------------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing need further work. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing need further work. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi, |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " |
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Jianyong Dai (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jianyong Dai updated PIG-243:
-----------------------------

    Attachment: PIG_243.patch

Patch for PIG under cygwin

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600905#action_12600905 ] 

Daniel Dai commented on PIG-243:
--------------------------------

I have tested reading one million records. Here is the result:
* On Linux, where trimming not taking place:
** Before patch: 10.53s
** After patch: 10.60s
* On Windows, where trimming branch taken:
** Before patch: 13.65s
** After patch: 13.69s

Seems that the overhead is neglectable. 

Another solution is to change the testcase rather than core code to fix this test failure. However, I think move it to core is better. Failing to read a file in dos format under cygwin sounds not reasonable. How do you think? 

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Pi Song (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12601104#action_12601104 ] 

Pi Song commented on PIG-243:
-----------------------------

+1

BTW, there is another guy called TextLoader that needs '\r' removed as well.


> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/30/08 11:26 AM:
----------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing '\r' (0x0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing '\r' |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing '\r') is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing '\r' (0x0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing '\r' |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12608231#action_12608231 ] 

Olga Natkovich commented on PIG-243:
------------------------------------

Thanks Daniel. Running tests now.

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: cygpath.patch, PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Daniel Dai updated PIG-243:
---------------------------

    Attachment:     (was: PIG_243.patch)

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 6/24/08 9:56 AM:
---------------------------------------------------------

Attached patch fix -most- all problems under cygwin. -4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work.- Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing '\r' (0x0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing '\r' |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |
| Pig streaming put the current directory into PATH. Current directory is obtained using "System.getProperty("user.dir")", it is a Windows style path under cygwin. However, Cygwin only accept Unix style path in PATH environment variable. So put the Windows path obtained from "System.getProperty" directly into PATH will not work | TestStreaming | Change Windows path to Unix path under Cygwin. Using Cygwin utility "cygpath"|
Only two change are in core code. One is PigStorage.java (to fix tailing '\r'). The other is ExecutableManager (Add cygpath). All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing '\r' (0x0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing '\r' |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing '\r') is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: cygpath.patch, PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12601213#action_12601213 ] 

Daniel Dai commented on PIG-243:
--------------------------------

You are right. We should change TextLoader as well. I have updated patch to include this. Thank you!

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600925#action_12600925 ] 

Olga Natkovich commented on PIG-243:
------------------------------------

Yes, I think this is fine. I will test and commit the patch sometimes tomorrow or Monday.

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600384#action_12600384 ] 

daijy edited comment on PIG-243 at 5/29/08 2:36 PM:
---------------------------------------------------------

Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\ \", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 

      was (Author: daijy):
    Attached patch fix most problems under cygwin. 4 errors remaining in TestStreaming for map-reduce testing under cygwin need further work. Patch is tested in both Linux and Cygwin. Here is a list of detected problem and solution:
||Problem||Affected Testcase||Solution||
|Windows path delimit "\" interpreted as escape in java|TestMapReduceResultRecycling, TestMapReduce, TestOrderBy, TestBuiltin, TestFilterOpNumeric, TestBinaryStorage, TestFilterOpString, TestEvalPipeline, TestPigFile, TestReversibleLoadStore, TestAlgebraicEval, TestLargeFile, TestCombiner, TestPi |replace "\" with "\\", see [http://issues.apache.org/jira/browse/PIG-152]|
| Runtime.exec(String[]) fail for perl -e if perl statement include quote character for unknown reason |TestParamSubPreproc, TestStreaming|append a escape character \ to quote character " under Windows|
| There is a tailing CR (ascii: 0d) character for line read from text file under cygwin|TestBuiltin, TestEvalPipeline, TestPigSplit, TestStreaming|remove tailing 0x0d |
| Bug: forget to close output stream, following deletion operation fail|TestBinaryStorage|Fix bug |

Only one change to PigStorage.java (to fix tailing 0x0d) is inside core code. All other changes are in testcase. 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Updated: (PIG-243) Make pig work on Windows

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Olga Natkovich updated PIG-243:
-------------------------------

    Resolution: Fixed
        Status: Resolved  (was: Patch Available)

Patch committed. Thanks Daniel for contributing.

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: cygpath.patch, PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Pi Song (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12601602#action_12601602 ] 

Pi Song commented on PIG-243:
-----------------------------

So far, we have ensured the data internalize/externalize bit. Is there anything else in the engine that can cause trouble on Windows? 
I couldn't think of any.

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Commented: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600931#action_12600931 ] 

Daniel Dai commented on PIG-243:
--------------------------------

Thank you!

> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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


[jira] Issue Comment Edited: (PIG-243) Make pig work on Windows

Posted by "Daniel Dai (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/PIG-243?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12600905#action_12600905 ] 

daijy edited comment on PIG-243 at 5/29/08 1:19 PM:
---------------------------------------------------------

I have tested reading one million records. Here is the result:
* On Linux, where trimming not taking place:
** Before patch: 10.53s
** After patch: 10.60s
* On Windows, where trimming branch taken:
** Before patch: 13.65s
** After patch: 13.69s

Seems that the overhead is neglectable. 

Another solution is to change the testcase rather than core code to fix this test failure. However, I think move it to core is better. Failing to read a file in dos format (or need extra processing) under cygwin sounds not reasonable. How do you think? 

      was (Author: daijy):
    I have tested reading one million records. Here is the result:
* On Linux, where trimming not taking place:
** Before patch: 10.53s
** After patch: 10.60s
* On Windows, where trimming branch taken:
** Before patch: 13.65s
** After patch: 13.69s

Seems that the overhead is neglectable. 

Another solution is to change the testcase rather than core code to fix this test failure. However, I think move it to core is better. Failing to read a file in dos format under cygwin sounds not reasonable. How do you think? 
  
> Make pig work on Windows
> ------------------------
>
>                 Key: PIG-243
>                 URL: https://issues.apache.org/jira/browse/PIG-243
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
>         Attachments: PIG_243.patch
>
>
> Currently a large number of unit tests is failing on Windows. We need to fix that.

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