patch for intermittent Ant failures

[Vadim Nasardinov <vadimn@xxxxxxxxxx>]

I ran into a problem with unit tests in CAROL and was about to submit
a new bug report, when I noticed that this problem had been already
reported about two years ago.  So, I just posted a comment on the old
ticket:

http://forge.objectweb.org/tracker/index.php?func=detail&aid=100160&group_id=26&atid=100026

Unfortunately, the line wrapping is all messed up and I couldn't
figure out a way to upload the patch as a separate file.  I'm
attaching the bug report and the patch to this message.

Thanks,
Vadim

The attached patch seems to fix a non-deterministic build failure.

The build failure (included at the bottom of this file) occurs within
the "launch.test.double" target in
carol/test/conform/generic/carol_double.xml.

This target launches a number of process simultaneously within the
<parallel> task.  One of these processes does this:'


      <sequential>
        <sleep seconds="10"/>
        <delete dir="${report.dir}" />
        <mkdir dir="${report.dir}" />
        <mkdir dir="${report.dir}/html" />

        ... the rest omitted for brevity ...
      </sequential>

Another process does this:

      <sequential>
        <sleep seconds="10"/>
        <mkdir dir="${report.dir}" />
        <mkdir dir="${report.dir}/html" />
        <junit printsummary="yes" 

         ... snipped for brevity ...
      </sequential>

Clearly, this is a race condition.  Consider the following sequence of
events:

   1. Process #2 creates ${report.dir}.
   2. Process #2 attempts to create ${report.dir}/html.  This involves
      creating all non-existent parent directories.  In this
      particular, case the parent directory exists.  So, process #2
      obtains a file handle for the parent directory in preparation
      for creating the requested subdirectory.
   3. Process #1 deletes ${report.dir}.
   4. Process #2 is now holding an invalid file handle for a directory
      that no longer exists.  It attempts to create the html/
      subdirectory and fails.

As far as I can tell, the <delete> task is unnecessary.


Vadim



P.S. Here's the build failure:

==============================================================================
launch.test.double:
     [java] ManageableRegistry started on port 1099
     [java] Initial Naming Context:
     [java] 
IOR:000000000000002b49444c3a6f6d672e6f72672f436f734e616d696e672f4e616d696e67436f6e746578744578743a312e30000000000001000000000000007c000102000000000a3132372e302e302e310007d000000035afabcb0000000020e2e2004c0000000100000000000000010000000d544e616d65536572766963650000000000000004000000000a0000000000000100000001000000200000000000010001000000020501000100010020000101090000000100010100
     [java] TransientNameServer: setting port for initial object references 
to: 2000
     [java] Ready.
    [mkdir] Created dir: 
/var/vadim/scratch/ow/carol/output/dist/test/results/multi_suniiop_sunjrmp1.2
    [mkdir] Created dir: 
/var/vadim/scratch/ow/carol/output/dist/test/results/multi_suniiop_sunjrmp1.2/html
    [junit] Running 
org.objectweb.carol.jtests.conform.basic.clients.MultiProtocolTests
    [junit] Tests run: 4, Failures: 0, Errors: 0, Time elapsed: 0.726 sec
     [java] Timeout: killed the sub-process
     [java]     at org.apache.tools.ant.taskdefs.Java.fork(Java.java:749)
     [java]     at 
org.apache.tools.ant.taskdefs.Java.executeJava(Java.java:170)
     [java]     at org.apache.tools.ant.taskdefs.Java.execute(Java.java:83)
     [java]     at 
org.apache.tools.ant.UnknownElement.execute(UnknownElement.java:275)
     [java]     at org.apache.tools.ant.Task.perform(Task.java:364)
     [java]     at 
org.apache.tools.ant.taskdefs.Sequential.execute(Sequential.java:64)
     [java]     at 
org.apache.tools.ant.UnknownElement.execute(UnknownElement.java:275)
     [java]     at org.apache.tools.ant.Task.perform(Task.java:364)
     [java]     at 
org.apache.tools.ant.taskdefs.Parallel$TaskRunnable.run(Parallel.java:417)
     [java]     at java.lang.Thread.run(Thread.java:534)
     [java] Timeout: killed the sub-process
     [java]     at org.apache.tools.ant.taskdefs.Java.fork(Java.java:749)
     [java]     at 
org.apache.tools.ant.taskdefs.Java.executeJava(Java.java:170)
     [java]     at org.apache.tools.ant.taskdefs.Java.execute(Java.java:83)
     [java]     at 
org.apache.tools.ant.UnknownElement.execute(UnknownElement.java:275)
     [java]     at org.apache.tools.ant.Task.perform(Task.java:364)
     [java]     at 
org.apache.tools.ant.taskdefs.Parallel$TaskRunnable.run(Parallel.java:417)
     [java]     at java.lang.Thread.run(Thread.java:534)
     [java] Timeout: killed the sub-process
     [java]     at org.apache.tools.ant.taskdefs.Java.fork(Java.java:749)
     [java]     at 
org.apache.tools.ant.taskdefs.Java.executeJava(Java.java:170)
     [java]     at org.apache.tools.ant.taskdefs.Java.execute(Java.java:83)
     [java]     at 
org.apache.tools.ant.UnknownElement.execute(UnknownElement.java:275)
     [java]     at org.apache.tools.ant.Task.perform(Task.java:364)
     [java]     at 
org.apache.tools.ant.taskdefs.Parallel$TaskRunnable.run(Parallel.java:417)
     [java]     at java.lang.Thread.run(Thread.java:534)

BUILD FAILED
/var/vadim/scratch/ow/carol/build.xml:225: The following error occurred while 
executing this line:
/var/vadim/scratch/ow/carol/test/build.xml:183: The following error occurred 
while executing this line:
/var/vadim/scratch/ow/carol/test/conform/carol_all_tests.xml:8: The following 
error occurred while executing this line:
/var/vadim/scratch/ow/carol/test/conform/carol_all_tests.xml:32: The 
following error occurred while executing this line:
/var/vadim/scratch/ow/carol/test/conform/carol_all_tests.xml:184: The 
following error occurred while executing this line:
/var/vadim/scratch/ow/carol/test/conform/generic/carol_double.xml:111: The 
following error occurred while executing this line:
/var/vadim/scratch/ow/carol/test/conform/generic/carol_double.xml:190: 
Directory 
/var/vadim/scratch/ow/carol/output/dist/test/results/multi_suniiop_sunjrmp1.2/html
 creation was not successful for an unknown reason

Total time: 5 minutes 53 seconds

Index: test/conform/generic/carol_double.xml
===================================================================
RCS file: /cvsroot/carol/carol/test/conform/generic/carol_double.xml,v
retrieving revision 1.11
diff -u -r1.11 carol_double.xml
--- test/conform/generic/carol_double.xml	9 Feb 2004 11:00:24 -0000	1.11
+++ test/conform/generic/carol_double.xml	17 Dec 2004 22:17:53 -0000
@@ -167,7 +167,6 @@
 
       <sequential>
 	<sleep seconds="10"/>
-	<delete dir="${report.dir}" />
 	<mkdir dir="${report.dir}" />
 	<mkdir dir="${report.dir}/html" />
 	<junit printsummary="yes"