db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rick Hillegas (JIRA)" <derby-...@db.apache.org>
Subject [jira] Commented: (DERBY-1079) Build javadoc under jdk 1.6
Date Wed, 12 Apr 2006 17:41:20 GMT
    [ http://issues.apache.org/jira/browse/DERBY-1079?page=comments#action_12374223 ] 

Rick Hillegas commented on DERBY-1079:
--------------------------------------

As a first step, I would like to check in the latest version of JavaCC (4.0). This involves
making the following change to trunk/java/tools/org/apache/derby/impl/tools/build.xml. For
some reason, two generated grammars (ij and mtGrammar) are currently compiled against the
jdk1.3 libraries; the new JavaCC generates code which invokes a RuntimeException constructor
not available in 1.3. This change moves that grammar compilation to the corresponding jdk14
build target:

<       <exclude name="${derby.dir}/impl/tools/ij/ij.java"/>
<       <exclude name="${derby.dir}/impl/tools/ij/mtGrammar.java"/>
78,79d75
<       <include name="${derby.dir}/impl/tools/ij/ij.java"/>
<       <include name="${derby.dir}/impl/tools/ij/mtGrammar.java"/>

The new JavaCC does not generate code using the "enum" keyword for variable names. These illegal
variable names choked the jdk1.6 javadoc tool. Using the new JavaCC, I am able to build javadoc
under 1.6.

I will run derbyall before proceeding with this patch. Does anyone know why I should not make
this change and upgrade our checked-in version of JavaCC to the latest 4.0 version?

> Build javadoc  under jdk 1.6
> ----------------------------
>
>          Key: DERBY-1079
>          URL: http://issues.apache.org/jira/browse/DERBY-1079
>      Project: Derby
>         Type: Bug

>   Components: Build tools
>     Versions: 10.2.0.0
>     Reporter: Rick Hillegas
>     Assignee: Rick Hillegas
>      Fix For: 10.2.0.0

>
> We would like to build the javadoc under 1.6 so that all of the classes (including the
JDBC 3 and JDBC 4 support) end up in the same directory tree. This is particularly important
for the published API which we expose to end-users.
> Right now you can do the following:
> 1) Build the whole codeline using the 1.4 compiler for most classes and the 1.6 compiler
for the JDBC4 support.
> 2) Build javadoc in a 1.4 environment (with JAVA_HOME set to 1.4). This runs step (1)
if it has not already happened. This javadoc excludes the JDBC4 support because generics-laden
JDBC4 signatures choke the 1.4 compiler.
> 3) Build the javadoc in a 1.6 environment (with JAVA_HOME set to 1.6). This will fail
if you have not run step (1); this is because you can't build Derby in a 1.6 environment yet.
This also generates a number of warnings because the 1.6 javadoc tool objects to code generated
by the JAVACC grammar tool; JAVACC turns out code with loop variables distastefully named
"enum". In addition, today, the 1.6 javadoc excludes the JDBC4 support.
> We would like to end up with the following situation:
> a) If your ant.properties points at a 1.6 installation, then the javadoc targets will
use the 1.6 javadoc tool and will include Derby's JDBC4 support. This will work regardless
of whether you have already built the class tree. If you have not already built the class
tree, then we will compile it under scenario (1) above.
> b) If, however, your ant.properties does not point at a 1.6 installation, then the javadocs
target will continue to use the 1.4 javadoc tool to build only the classes built today. The
JDBC4 support will be excluded from this javadoc.
> c) As part of releasing 10.2, we will build the javadoc under scenario (a).
> d) Once 1.6 exits beta and becomes a production vm, the community can debate when we
want to fix DERBY-1078 and require 1.6 in the build environment.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira


Mime
View raw message