Marc-André Lureau <marcandre.lur...@redhat.com> writes:
> We would like to use a same QObject type to represent numbers, whether
> they are int, uint, or floats. Getters will allow some compatibility
> between the various types if the number fits other representations.
>
> Add a few more tests while at it.
>
> Signed-off-by: Marc-André Lureau <marcandre.lur...@redhat.com>
[...]
> diff --git a/blockdev.c b/blockdev.c
> index 892d768574..154c95d402 100644
> --- a/blockdev.c
> +++ b/blockdev.c
> @@ -334,9 +334,11 @@ static bool parse_stats_intervals(BlockAcctStats *stats,
> QList *intervals,
> break;
> }
>
> - case QTYPE_QINT: {
> - int64_t length = qint_get_int(qobject_to_qint(entry->value));
> - if (length > 0 && length <= UINT_MAX) {
> + case QTYPE_QNUM: {
> + int64_t length;
> +
> + if (qnum_get_try_int(qobject_to_qnum(entry->value), &length) &&
> + length > 0 && length <= UINT_MAX) {
> block_acct_add_interval(stats, (unsigned) length);
> } else {
> error_setg(errp, "Invalid interval length: %" PRId64,
> length);
Less churn and a bit more legible:
+ case QTYPE_QNUM: {
+ int64_t length = qnum_get_int(qobject_to_qint(entry->value));
+
+ if (length > 0 && length <= UINT_MAX) {
Aborts when entry->value doesn't fit into int64_t, but that's no worse
than before.
[...]
> diff --git a/tests/test-qobject-input-visitor.c
> b/tests/test-qobject-input-visitor.c
> index 3a6ce2226c..9e3d4aa7e5 100644
> --- a/tests/test-qobject-input-visitor.c
> +++ b/tests/test-qobject-input-visitor.c
> @@ -164,9 +164,9 @@ static void
> test_visitor_in_int_overflow(TestInputVisitorData *data,
> Error *err = NULL;
> Visitor *v;
>
> - /* this will overflow a Qint/int64, so should be deserialized into
> - * a QFloat/double field instead, leading to an error if we pass it
> - * to visit_type_int. confirm this.
> + /* this will overflow a QNUM_I64, so should be deserialized into a
Wing both ends and start with a capital letter, please:
+ /*
+ * This will overflow a QNUM_I64, so should be deserialized into a
> + * QNUM_DOUBLE field instead, leading to an error if we pass it to
> + * visit_type_int. confirm this.
Humor me:
* visit_type_int(). Confirm this.
> */
> v = visitor_input_test_init(data, "%f", DBL_MAX);
>
[...]
With these minor touch-ups:
Reviewed-by: Markus Armbruster <arm...@redhat.com>
