Review Request for 8027481: jdeps to handle classes with the same package name ...

[Mandy Chung]

On 10/29/13 2:19 PM, Daniel Fuchs wrote:
> Hi Mandy,
>
> The changes look good to me - though I am a novice in this area:
> so I could have missed a few things.
>
> As a general comment I would say that this code could benefit
> from some more commenting - if you don't know what it's supposed
> to do then it is a bit difficult to reverse-engineer :-)
>

Thanks for the review, Daniel.  I have added some more comments and 
updated the webrev per the questions you asked at IM.

Mandy

> best regards,
>
> (not a reviewer)
>
> -- daniel
>
>
> On 10/29/13 8:30 PM, Mandy Chung wrote:
>> Webrev at:
>> http://cr.openjdk.java.net/~mchung/jdk8/webrevs/8027481/webrev.00/
>>
>> https://bugs.openjdk.java.net/browse/JDK-8027481
>>
>> This patch fixes a couple of small jdeps issues:
>>
>> 1) jdeps doesn't handle the same package name being used in more than
>> one JAR files.  It currently incorrectly shows the dependencies for
>> example if two JAR files have classes in the unnamed package.
>>
>> The fix is to change the Analyzer class to maintain the target name
>> together the Archive where the dependence comes from (the new
>> ArchiveDeps.Dep class in this patch).
>>
>> 2) javax.crypto.* packages are in compact1 but shown as "JDK internal
>> API" because they are not in rt.jar and thus not included in ct.sym. The
>> simple fix is to special it in jdeps rather than changing ct.sym to work
>> with multiple jars (which will impact javac)
>>
>> 3) -verbose prints all class dependencies including those within the
>> same JAR file. Having the summary.dot file to show package-level summary
>> would be more useful for analysis when -verbose is specified.
>>
>> I also added an experimental option to label the edges in the
>> summary.dot with the dependencies to make it easier to view the graph.
>> It's useful when it has managable amount of dependencies displayed in
>> the graph and not sure how well the dot tool can render. So I keep it
>> experimental for now.
>>
>> thanks
>> Mandy
>