On Sun, Sep 21, 2025 at 11:39 AM Bill Blough <[email protected]>
wrote:

> Hi Robert,
>
> I've had some time to look at this.
>
> On Sat, Sep 13, 2025 at 11:17:47PM -0400, Bill Blough wrote:
> > On Wed, Sep 10, 2025 at 02:02:58PM -1000, robertlazarski wrote:
> > > I was able to run the json tests successfully today.
> > >
> > > I have some time again for Axis2/C and I made some commits today -
> please
> > > review, I got some help from modern AI tools.
> > >
> > >
> https://github.com/apache/axis-axis2-c-core/commit/580af19055cf77b47602d8f5443784efb7555bdd
> > >
>
> The changes look fine to me.
>
> However, I did update the PKG_CHECK_MODULES to require the latest json-c
> version.
>
> Some of the enum values added to json_tokener_error_to_str are only
> available in the latest version, and trying to compile with an older
> version gives a compile error.
>
> By specifying the version via the pkgconf check it will give a more
> user-friendly message about the required version not being available,
> rather than a compiler error about undeclared variables.
>
> > > Anyways, I have some time in the next week and any guidance on what
> needs
> > > to happen before a 2.0.0 release would be helpful.
> >
> <snip>
> >
> > I can go through and triage again to see what we can pare down to a
> > reasonable release target, but that probably won't happen until next
> > weekend.
> >
>
> I've looked through the issues again.  It was a fairly quick look, and I
> didn't delve too deeply into specifics.  I untagged a few things that
> were tagged 1.7.0 that I no longer think are good candidates for the
> next release, and also tagged the JSON related issues, since they
> weren't included. I also fixed one really easy issue, so that's
> something, at least.  But after the add/remove we're still at 84 (for
> now).
>
> The majority of the issues I had tagged for 1.7.0 are memory-related:
> crashes, leaks, invalid dereferences, etc.  In a perfect world, these
> would all be fixed before the next release. But I also suspect some
> aren't fixable in their current state.
>
> Also, it's unclear to me how much pressure you're getting from the board
> with regard to the c side of things, but I imagine it's non-zero. And
> I'm curious if there are any expectations around a time frame for a
> release.  If either of these are true, then I suspect a significant
> number of those issues will need to be deferred in the interest of
> cutting a release sooner rather than later.
>
> Regardless, I suspect the next thing I should do is take a deeper dive
> into the open issues and start trying to reproduce them.  I'd say
> anything currently tagged 1.7.0 that is also reproducible with the
> current codebase could be a potential candidate for 2.0.  Anything not
> reproducible with the information given should be, at minimum, excluded
> from 2.0.  Depending on the circumstances, those items could be either
> closed or deferred for additional investigation at a later time.
>
> The caveat there is that at a few hours a week, it's likely to take me a
> significant amount of time to get through everything, so if there is
> some sort of target date, then we'll need to come up with a different
> approach.
>
>
> Best regards,
> Bill
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>
I just made a huge commit that replaces Axis2/Java codegen with a native
Axis2/C codgen implementation - with modern AI tools doing the heavy
lifting, though I know a lot about the use cases from being an Axis
committer since 2006.

I added extensive test cases and closed a bunch of codegen related issues -
but not all of them, there are a bunch more to go that I plan on reviewing.

In other news, I completed the http/2 Java transport - primarily
benefitting JSON, less so for SOAP and XML  - and the next step for me is
to port the entire http/2 requirements from Axis2/Java into Axis2/C.

Things are moving so quickly, from my standpoint I could be ready to
release in a week or two. If we need more time, we are all volunteers so
feel free.

We need 3 votes of course, so please review and let us know about any
outstanding work for 2.0.0.

Reply via email to