Re: Help with PHP Extension

Hello Christoph,

You were absolutely correct! Thank you so much!

It looks like I misinterpreted the `static zval*` return type on
current data to actually require a return of a zval from the function.
After checking how it was done in `com_dotnet/com_iterator.c` and
initialising in the `php_sparsehashmap_iterator` with a zval for data,
and assigning the data to that rather than an in-scope variable, it
worked like a dream :)

Again, thank you for your help with this. Happy to finally see this
working after months of trying :)

Regards,

Ciaran Moore

On Sat, Sep 10, 2022 at 6:12 PM Christoph M. Becker <[email protected]> wrote:
>
> On 10.09.2022 at 18:04, Ciaran Moore wrote:
>
> > I'm not sure where else to turn with this. It says something about a
> > problem when you post a question on StackOverflow and it gets zero
> > responses in over a week.
> >
> > I have been learning extension development, and have managed to build
> > the extension to do what I want and how I want it, until I try to
> > implement the Iterator. I don't know what it is about the data being
> > returned but I keep getting segmentation faults on the return *zval
> > for current data.
> >
> > This is a link to the StackOverflow question I posted which goes into
> > more detail about what is happening when the SIGSEGV is thrown.
> >
> > https://stackoverflow.com/questions/73542769/segmentation-fault-zval-str-when-accessing-current-data-in-iterator
> >
> > This also includes a link to a Github Repo I created to give a full
> > view of the extension and not rely on just posting snippets here and
> > there on the post.
> >
> > Any help or guidance with this would be greatly appreciated.
>
> The problem is that `data` points to uninitialized memory.  You need to
> actually have a `zval` somewhere, which is usually in the
> `php_sparsehashmap_iterator_t` structure.  See, for instance, how that
> is done in ext/com_dotnet/com_iterator.c.  The `zdata` member is set on
> initialization and whenever the iterator is moved, and then simply
> returned on `get_current_data()`.
>
> --
> Christoph M. Becker
>