Status December 19th, 2012

[Adam Domurad]

Last status report of 2012... exciting isn't it

This week:

patches:
   [rfc][icedtea-web] C++ unit-tests for memory leak w/ simple 
leak-detecting operator-new, along with fix
   Re: Regression in icedtea-web - parent classsloader have been stopped 
to be invoked
   [rfc][icedtea-web] Rewrite of MethodOverloadResolver
   [rfc][icedtea-web] Minor addition to C++ testing; cleanup

reviews:
   Re: [rfc][icedtea-web] @Remote and automated manual testcases
   Re: [rfc][icedtea-web] Reproducer of BeansStatement behaviour x2
   Re: [rfc][icedtea-web] a new reproducer LiveConnect "function 
parameters" tests
   Re: [rfc] icedtea-web minor fix for two unittests x2

misc:
   Took a small detour from looking into the deadlock but will get back 
into it
   'Small refactoring' of MethodOverloadResolver lead to finding many 
issues, spent some time making sure that it matched the (thankfully 
somewhat loose) spec exactly with in-depth unit tests.
   Looked into failure of ITW failing when used with ASAN. It's not 
clear to me whether ASAN works with dynamically loaded shared objects.
       - Wrote some additional C++ unit tests so that ASAN behaviour 
could be tested 'in a vacuum' and everything seems fine with static linking
       - Without ASAN browser -- very weird segfault with ASAN+plugin, 
with ASAN browser, fails 'globals check' that seems to pass when not in 
browser environment
       It would be nice to get this working as it'd complement runs of 
the test suite.
   Lots of patch reviewing

Next week:
   Pick up work again on deadlock
   Look into expiring-signatures issue