777 freezes and FPS loss

[Hooray]

You are correct, I understand what it is intended to do, i.e. it is supposed to encapsulate the name of the subsystem as part of the subsystem instance, so that there aren't all sorts of "foo" lookups all over the place in sg/fg - however, I haven't checked what else it might be doing.

For instance, if there is still old code using the old lookup method, it may be possible (in theory) that multiple instances of a subsystem may get allocated, using different handles.

For that, we need to look at the patch in question - reverting is unlikely to work, because this is stuff at the simgear/sgsubsystem level, i.e. low-level stuff that other code depends on.

Technically, it using a new template method apparently to make the pointer acquisition more compact, i.e. a single place for handling this in SG, instead of having this all over the place in sg and fg

[sanhozay]

As I like nothing more than a good bisect on a Sunday morning, I was going to see if I could find the commit that breaks the particle system. Sadly someone already did it .

https://sourceforge.net/p/flightgear/ma ... /34808426/

Hmm, commit 154464 ... haven't we met somewhere before.

[Hooray]

The particle system being broken is not necessarily a bad thing though, because it had a number of other issues, even back when it was still "working" otherwise: http://wiki.flightgear.org/Particle_system#Known_Issues

[legoboyvdlp]

As I like nothing more than a good bisect on a Sunday morning

Who doesn't?!
On the other hand, glad you found it, both of you.

[biloute974]

Hi all,

it seem that FP loss disappear with an update to 2016.2.0 version. Thanks to all devs for this great job.

[Hooray]

and like Hyde said, it was unrelated to the Canvas system

[biloute974]

Exactly