Re: [aspectj-users] State of the AspectJ/Java editor?

[Andrew Eisenberg <andrew@xxxxxxxxxxxx>]

Hi Johan,

It is certainly possible to generate ITD-aware elements in the Java
model.  Currently, in AJDT, we only generate them when required and
they are discarded when no longer needed.

You will be required to have ajdt.core installed when you are doing
this functionality.  Loading (almost) any class in ajdt.core will
start the plugin, so there is no need to explicitly start it.

Take a look at the classes ITDAwareSourceTypeInfo and ITDInserter.
These are the two classes that add ITD awareness to elements in the
Java model.  You may not need to call them directly.

Also, take a look at ITDAwareLookupEnvironment and
ITDAwareNameEnvironment.  These classes allow you to ask questions
about types and compilation units in an ITD-aware way.

You can see how they are used in the class AJCompilationUnitProblemFinder.

Hope this helps,
--a


On Thu, Jul 9, 2009 at 9:59 AM, Johan Fabry<jfabry@xxxxxxxxxxxxx> wrote:
> Hi Andrew, all,
>
> I have a question related to this: I'm doing some code analysis, related to
> cross-cutting report plugin that I was working on previously. I'm talking to
> Eclipse using its standard API, getting information from all the
> ICompilationUnit elements that I can access. There's a problem however,
> related to these red squiggles on valid code, which admittedly may be due to
> my setup (see below). The problem is that if the base code references (eg.
> does a method call of) an ITD, the ICompilationUnit for that class is not
> available. I suppose this is because the code does not compile using
> standard Java.
>
> I am using some external package to access the ICompilationUnits, which
> actually has a headless eclipse running in the background and talking to
> that instance. I suppose this is an eclipse that does not have AJDT loaded.
> If I would hack the package to ensure that eclipse loads AJDT (which I would
> like to avoid at all costs), will this solve my problem? (BTW Is there a way
> to 'manually' load a plugin by talking to this headless eclipse?). If having
> AJDT loaded does not solve the problem, is there another way?
>
> Thanks in advance!
>
> On 07 Jul 2009, at 17:53, Andrew Eisenberg wrote:
>
>> Hi David,
>>
>> Our goal is to ensure that spurious AspectJ errors (ie- red squiggles
>> on valid code) are not seen at all (but this has to be weighed against
>> being to aggressive and ignoring actual errors).  We have taken care
>> to ensure that the common uses of AspectJ syntax will not show red
>> squiggles in the editor, but more advanced uses of aspectj (such as
>> generic ITDs, especially those on interfaces) have been less well
>> tested.
>>
>> There is no explicit documentation that describes what is working and
>> what is not, but there are a slew of test cases for this and the code
>> for this is well documented.  At this point, if you are seeing
>> something in the editor being marked as a problem when the compiler
>> thinks it isn't, please let us know and we will address it as soon as
>> we can.
>>
>> For the past little while, we have been waiting to hear from users
>> when they come across some of these spurious errors.  And we fix them
>> as they come in.  The best way to help me debug this is to send some
>> examples of failing code.  You can raise a bug if you like, or you can
>> send it through this mailing list.
>>
>> --a
>
> --
> Johan Fabry
> jfabry@xxxxxxxxxxxxx - http://dcc.uchile.cl/~jfabry
> PLEIAD Lab - Computer Science Department (DCC) - University of Chile
>
>
>
>