db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From V Narayanan <V.Naraya...@Sun.COM>
Subject Re: IBM Java 5.0 class file for test harness
Date Thu, 15 Dec 2005 08:27:58 GMT
Thank you for pointing this out Dag. You have pointed out the following

1) JDBC4.0 compilation target does not echo anything about the number of
files printed
2) Upon inserting a bug and setting proceed=false the compilation
proceeds instead of failing

I have fixed this in addition to a small change documented below

a) Removing the -quiet option
b) Setting failonerror depending on the value of the proceed variable
c) Removing the echo task since if the exec task fails we would want 	
    the user to know the reason for failure which otherwise wouldnt be

printed if echo occurred after the exec task
d) I am also removing the part of the build which checks os and sets
executable to ant or ant.bat and am instead using vmlauncher attribute
of the exec task and am setting it to false.

I am attaching the patch to this email. Would you want me to raise a
JIRA issue for this and attach it to that?


Dag H. Wanvik wrote:
> Hi,
>>>>>>"V" == V Narayanan <V.Narayanan@Sun.COM> wrote:
> V> can you please confirm that the compilation was done with jdk16
> V> variable in ant.properties pointing to mustang home. The
> V> ClientPooledConnection40.java gets compiled only with this variable
> V> included.  thanx V.Narayanan
> I trid to compile with Mustang and it worked for me. I notice,
> however, that the jdbc4 target did not echo anything about number of
> files compiled, as does the ordinary compilation. Then, when I
> inserted a bug in a file (CallableStatement40.java), the build
> proceeded in spite of my setting proceed=false in ant.propeties. The
> build eventually reported BUILD SUCCEEDED, although the JDBC4 build
> failed. This seems to be a bug, unless something is wrong with my
> setup?
> Thanks,
> Dag

View raw message