linuxdcpp-team team mailing list archive

[klondike <franxisco1988@xxxxxxxxx>]

> 4) look for "signal" through adchpp/ClientManager.h, adchpp/ClientManager.cpp,
> swig/adchpp.i then copy-paste. :)
Not there yet, let me recover for the lan party and I'll fix this one. For now I reverted these changes since they seem to be behing the segfaults we got on Thursday.


> 5) the 2 dev.build calls should be switched.
Fixed

> 6) in swig/bloom.i, i don't think getBM should return a shared_ptr. that
> implies that the calling script maintains a shared ownership of the bloom
> plugin along with the plugin manager, which doesn't make sense. the plugin
> manager should remain in full control of the lifetimes of the plugins it
> manages. so make it return BloomManager* or BloomManager& with
> getCurrentCore(l)->getPluginManager().getPlugin("BloomManager").get().
Maybe we should change the getPlugin() API then, I can't see any other legitimate use for this that would return a shared_ptr here. The main problem is that if you are not returning a shared_ptr and the plugin gets removed then you may end up dereferencing invalid memory.
-- 
https://code.launchpad.net/~klondike/adchpp/adchpp/+merge/150163
Your team Dcplusplus-team is subscribed to branch lp:adchpp.