On 06/08/2015 11:45 PM, Junio C Hamano wrote:
Karthik Nayak <karthik.188@xxxxxxxxx> writes:
+/*
+ * API for filtering a set of refs. The refs are provided and iterated
+ * over using the for_each_ref_fn(). The refs are stored into and filtered
+ * based on the ref_filter_cbdata structure.
+ */
+int filter_refs(int (for_each_ref_fn)(each_ref_fn, void *), struct ref_filter_cbdata *data)
+{
+ return for_each_ref_fn(ref_filter_handler, data);
+}
I do not think it is such a good idea to allow API callers to
specify for-each-ref-fn directly. See my message in an earlier
review.
>
I did read your previous message. I misunderstood some things.
I also think ref_filter_cbdata is an implementation detail of
filter_refs and may not have to be exposed to the API callers.
It probably is more sensible for them to pass
- an array of refs to receive filtered results (your ref_array thing)
- the criteria to use when filtering (your ref_filter thing)
This could be done.
>
as two separate parameters to this function, together with other
parameters that lets you (meaning the implementation of filter_refs())
to decide which for-each-ref iterator to call, e.g. do you want to
use raw iteration? do you want to iterate only over refs/heads? etc.
In other words, the caller of this API should not have to know that
you (meaning the implementation of filter_refs()) are internally
using for_each_ref() API.
I'm a little confused about this, how exactly do you propose we go about
doing something like this? I mean, usually the user of the API
knows what exactly they want, like in tag.c, branch.c and for-each-ref.c
But I'm not sure what you mean by "parameters that lets you (meaning the
implementation of filter_refs()) to decide which for-each-ref iterator
to call". A small example maybe? Thanks!
--
Regards,
Karthik
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html