You are viewing a plain text version of this content. The canonical link for it is here.
Posted to proton@qpid.apache.org by "Rafael H. Schloming (JIRA)" <ji...@apache.org> on 2012/12/03 22:33:58 UTC

[jira] [Commented] (PROTON-159) port proton to C++

    [ https://issues.apache.org/jira/browse/PROTON-159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13509068#comment-13509068 ] 

Rafael H. Schloming commented on PROTON-159:
--------------------------------------------

So looking at this patch I'm a bit overwhelmed by how huge it is. I don't think it's feasible to actually review it and then apply it because I think in the time it would take to actually review it, trunk will move far enough that it won't apply anymore. For example since this patch was created, I've checked in changes to codec.c which I'm guessing will conflict since there are 76 diffs against it. Given that this patch combines both semantic changes (like the VLA stuff) with mechanical ones, I also don't think it's a very good idea to short-circuit the review process.

Why don't we break out the build system changes into a separate patch and get that in as a first step? I realize this will result in the BUILD_WITH_CXX=On version of the build spewing out lots of errors, but we should then be able to pull in patches that incrementally reduce the errors. I'd suggest maybe a second patch limited only to the mechanical code changes and then subsequent patches addressing the semantic changes, with that pattern possibly repeated on a per-file basis as dictated by the build order.

Other than the comment on the overall size and approach to getting these changes in, I did a random spot check on a few of the files to see the general shape of these changes. I'm a little concerned with the apparent randomness of some of the changes, e.g. changing loop counters from int to size_t or changing declarations to unsigned. It's not at all obvious to someone working with the C build why things like that would be necessary. I don't want to create a situation where you pretty much always need to build every change against both the C and C++ compiler in order to know if it is going to break the larger build.
                
> port proton to C++
> ------------------
>
>                 Key: PROTON-159
>                 URL: https://issues.apache.org/jira/browse/PROTON-159
>             Project: Qpid Proton
>          Issue Type: New Feature
>          Components: proton-c
>    Affects Versions: 0.3
>            Reporter: Cliff Jansen
>         Attachments: proton-159-0.diff, proton-159-0-partial.diff
>
>
> Make code compile in both C99 and C++, using the gnu toolchain.  This is a necessary first step
> towards a Microsoft Visual Studio port (where the compiler supports C++ but not C99).

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira