Re: [hibachi-dev] Hibachi Plans Questions

[Todd Wallentine <tcw@xxxxxxx>]

Tom and Adam, thanks for the information.  You have answered all the
questions that we have at this time.  Our group will be meeting about
this again this week and I hope we make a decision on whether or not we
will use Hibachi as a base for our work or implement our own plugins
from scratch (I am anxious to get started on the implementation).

todd

ps
When I said grammar I meant the ANTLR grammar in the
org.eclipse.hibachi.antlr project (Ada.g3).  Whan I talked about the
model and AST I meant the code that holds the parsed Ada which is in
org.eclipse.hibachi.core project (in org.eclipse.hibachi.model and
.hibachi.internal packages).  Sorry for not being specific ... but you
did answer the question despite my ambiguous questions.

Tom GROSMAN wrote:
> Hello Todd,
> 
> Just to add to what Adam wrote- several months ago, Rod Chapman of Praxis
> mentioned to me that there was a beta version of a SPARK plugin for Eclipse.
> I suggest contacting him to determine the status of that effort. Please let
> us know what you find out.
> 
> There is no effort to include a compiler with Hibachi. This is not possible
> for several reasons. First, everything released under the Hibachi project
> must meet Eclipse's rigorous Intellectual Property rules. So for instance,
> even the GNAT compiler,which is under GPL, couldn't be distributed as part
> of Hibachi since GPL is not compatible with EPL. There are other reasons why
> a compiler is not included as part of Hibachi, but the one I listed is a
> deal breaker, so no need to go on. 
> 
> That said, there will probably be a release of an Eclipse based Ada
> development environment bundle including Hibachi, the GNAT toolchain and
> anything else needed to do Ada development with a single download. However,
> it will not be part of the Hibachi project. It would be available via a
> non-Eclipse download site. This is actually what Doug Schaefer (the CDT
> project lead) has done for C/C++ with Wascana (see
> http://wascana.sourceforge.net/).
> 
> Regards,
> 
> Tom
>> -----Original Message-----
>> From: Adam Haselhuhn [mailto:ahaselhu@xxxxxxxxxxxxx]
>> Sent: jeudi 21 février 2008 19:55
>> To: Tom GROSMAN (p); tom.grosman@xxxxxxxxx
>> Subject: Fw: [hibachi-dev] Hibachi Plans Questions
>>
>> Todd,
>>
>> I'm not entirely clear on what you mean by the "grammar".  Do you mean the
>> API?  If so, it is likely that it will change, but the plan is to keep
>> backwards compatibility as much as possible using deprecated methods.
>> Hopefully, there wouldn't be a large effort required for you to update to
>> the "official" API.
>>
>> Currently, the model does not correctly represent all Ada elements, in
>> particular generics handled in any special way.  There will be changes to
>> the model to support such elements that currently are not fully
>> represented.
>> Also, there is likely to be an new AST implementation to provide a better
>> API for tool builders.
>>
>> It is unlikely that there will be a "built-in" Ada compiler, if that is
>> what
>> you are referring to.  A free Ada compiler is available (GNAT), which
>> Hibachi supports, and it isn't in the interests of the contributors of the
>> Hibachi project to provide such a compiler.  However, it would be
>> acceptable
>> for a third party to implement such a compiler.  I am not aware of anyone
>> working on support for SPARK.
>>
>> That is correct.  Identifiers are resolved to model objects by generating
>> the identifier's fully-qualified name, locating the appropriate
>> compilation
>> unit, and finally searching for each component of the fully-qualified name
>> until the element is found.  I'm not sure exactly what the benefits would
>> be
>> in implementing a symbol table over the current system.  If there are
>> sufficient benefits, we would certainly consider adding it.
>>
>> Adam
>>
>>
>> ----- Original Message -----
>> From: "Todd Wallentine" <tcw@xxxxxxx>
>> To: <hibachi-dev@xxxxxxxxxxx>
>> Sent: Wednesday, February 20, 2008 3:56 PM
>> Subject: [hibachi-dev] Hibachi Plans Questions
>>
>>
>>> As I mentioned in a previous email to the group, my research group is
>>> looking at using Hibachi as a base for our new project (a SPARK
>>> development environment that uses our analysis tools).  Because of that,
>>> we hoped we could get some questions answered to help us make our
>>> decision.
>>>
>>> First, we like what we have seen so far but wonder what the
>>> current/future plans look like.  For example, is the grammar currently
>>> stable or are there plans to change it?  And are there any changes
>>> planned for the model (or AST)?  I think our concern here is evaluating
>>> how much effort it will require on our part to keep our extensions
>>> current with the mainline changes (e.g., if there are changes planned it
>>> will take more effort to stay current).
>>>
>>> Second, we wondered about the plans for the compiler and tool chain.
>>> Are there any plans for a full Ada compiler (instead of relying on the
>>> external tool-chains)?  And does anyone know of a group working on
>>> integrating a SPARK tool-chain or any SPARK features?  It would be great
>>> to coordinate with others that are interested in SPARK.
>>>
>>> Last, a technical question that came up during our meeting.  Based upon
>>> our (brief) review of the code, there does not seem to be a symbol table
>>> that is queried. Rather there are helper methods to resolve symbols in
>>> the AdaModelManager (in the org.eclipse.hibachi.model package).  Is this
>>> correct or are we missing something?  And if it is correct, is there
>>> interest in implementing a symbol table?
>>>
>>> Thanks in advance,
>>> Todd Wallentine
>>> _______________________________________________
>>> hibachi-dev mailing list
>>> hibachi-dev@xxxxxxxxxxx
> 
> No virus found in this outgoing message.
> Checked by AVG Free Edition. 
> Version: 7.5.516 / Virus Database: 269.20.9/1292 - Release Date: 21/02/2008
> 16:09
>  
> 
>