+++ This bug was initially created as a clone of Bug #33870 +++ While fixing Bug #33870 I noticed that: - if you only run one of the dbus-test sub-tests by putting its name on the command line, the skipped tests still produce output when they check for memory leaks - the recursive marshalling test is #if 0, but at least the quicker early stages of it should be run - the arguments to find_field_for_modification are rather opaque - some of dbus-marshal-recursive's extern API is obscure, and only used to realign messages; making it static clarifies its scope
Created attachment 47931 [details] [review] [PATCH 1/4] dbus-test: only test for memleaks if we actually ran the test This reduces output spam if you're only running one part.
Created attachment 47932 [details] [review] [PATCH 2/4] Reinstate the recursive marshalling test but skip the later parts by default
Created attachment 47933 [details] [review] [PATCH 3/4] find_field_for_modification: document where the readers are left pointing
Created attachment 47934 [details] [review] [PATCH 4/4] dbus-marshal-recursive: make some internal functions static Having a disabled type writer, comparing type readers by position, and performing a single step of copying from a reader to a writer all seem pretty obscure, and were only used within dbus-marshal-recursive.c (to realign after insertion).
Created attachment 51333 [details] [review] [2/4 v2] Reinstate the recursive marshalling test but skip the later parts by default Updated version of Attachment #47932 [details], which conflicted with more recent changes to master.
Review++, having no hat though
(In reply to comment #6) > Review++, having no hat though Thanks, fixed in git for 1.5.10.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.