Eric Blake <ebl...@redhat.com> writes: > On 12/03/2015 10:13 AM, Markus Armbruster wrote: >> Eric Blake <ebl...@redhat.com> writes: >> >>> On 12/03/2015 09:37 AM, Markus Armbruster wrote: >>>> prop_get_fdt() misuses the visitor API: when fdt is null, it doesn't >>>> visit anything. object_property_get_qobject() happily >>>> object_property_get_qobject(). Amazingly, the latter survives the >>> >>> Something got lost or otherwise corrupted in that sentence. Were you >>> trying to say one function happily calls another? If so, which of the >>> two "object_property_get_qobject()" strings should be changed, to what? >> >> No idea what happened. Correction: insert "calls" after "happily": >> >> prop_get_fdt() misuses the visitor API: when fdt is null, it doesn't >> visit anything. object_property_get_qobject() happily calls >> object_property_get_qobject(). >> > That still reads "A() happily calls A()" - are we talking about > recursion here? I still wonder if you meant a second function name, > and/or mention of the fact that we are calling a function with NULL > rather than a QObject?
I'm incapable of proof-reading anything I wrote myself %-} prop_get_fdt() misuses the visitor API: when fdt is null, it doesn't visit anything. object_property_get_qobject() happily calls qmp_output_get_qobject() then. Amazingly, the latter survives the misuse. Turns out we've papered over it long before prop_get_fdt() existed, in commit 1d10b44.