tuscany-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thomas Gentsch ...@e-tge.de>
Subject Re: Porting Tuscany C++ to C++11 and some code cleanup
Date Sun, 25 Nov 2012 21:31:02 GMT
On Sun, 2012-11-25 at 10:38 -0800, Jean-Sebastien Delfino wrote:
> On Fri, Nov 23, 2012 at 9:02 AM, Thomas Gentsch <tg@e-tge.de> wrote:
>         
>         Hi JS + all,
>         
>         we are using Tuscany C++ in one of our projects, hence we/I
>         have some
>         personal interest and also would like to contribute, if
>         possible. 
>         
>         Tbh, I don't know details about C++11 (will read a bit) nor
>         the other
>         things you mention.
> 
> 
> Great! Papers from the C++11 standard work groups are available on
> open-std.org there [1]. The final ISO standard is not free but the
> last draft [2] is free.  If you just want a high level overview
> there's one on Wikipedia [3]. Hope that helps.

Shortly looked through it, definitely interesting. Of course that makes
me worry a bit how far compilers support this already ... particularly
as partially maybe (as in our case) older OS versions are to be
supported.

>         One thing though which appears quite important to me is that
>         the various
>         patches we/I have done (most/all (?) submitted to Jira) and
>         other people
>         are incorporated into the general code so that they are in the
>         next
>         release if there will ever be one ... and of course this is
>         easier
>         before any refactorings.
> 
> 
> I found one JIRA with a patch from you [4], which looks like a change
> to the SDO + Axiom support in the SCA-Native-M3 release.

I quickly looked into Jira and found at least these:

https://issues.apache.org/jira/browse/TUSCANY-3841
https://issues.apache.org/jira/browse/TUSCANY-3715
https://issues.apache.org/jira/browse/TUSCANY-3597

https://issues.apache.org/jira/browse/TUSCANY-1552


> 
> The C++ trunk has already gone through some refactoring and quite a
> few changes over time since M3... There's no SDO anymore

Oh!!! Actually I always believed Tuscany is mainly about SDO!
Well, then I have to correct myself - we are using Tuscany C++ SDO. I
wasn't aware of this at all (haven't looked into it for a while though)

> (like the Java trunk, which I believe dropped SDO as well), the
> composite files use the new SCDL from OASIS, the runtime integrates
> with Apache HTTPD, etc. There's still an integration with Axis2C but
> I've not looked at it for a very long time.
> 
> 
> So, if you're still using SDO you'll probably want to stick to the M3
> level, but I can help apply your patch to a maintenance branch off M3
> at some point if that helps.

Well, if SDO is dead anyway then there will not be any release anymore I
suppose, so maybe no need to bother.

>         
>         The other thing: Are the planned changes going to modify the
>         API
>         interface or is it just internally?
> 
> 
> The C++11 changes only apply to the runtime internals, but again if
> you're using the SDO-based API from M3, that already changed a long
> time ago.

OK, then I should look into the alternatives? Is there any replacement?
Many thx + regards,
  tge

>         
>         Rgds,
>           tge
>         
>         
> 
> 
> [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2011/
> [2] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2011/n3242.pdf
> [3] http://en.wikipedia.org/wiki/C%2B%2B11
> [4] https://issues.apache.org/jira/browse/TUSCANY-3598




Mime
View raw message