Weekly musing #3 - Mr Bitcoin is hungry
Bitcoin Unlimited
I raised BUIP051, am going to take a close look at porting those changes over from XT during the coming week or two.
Happily, the proposal to spend some money for a better (paid) Travis build service has met with a supportive reception at BU, so I think we are soon going to have the possibility to run longer test suites (up to 120 minutes each - maybe need to break the test suite down into several batches which individually fit in under 2 hrs, or just execute randomly and time out just below 2 hrs).
Parallel execution of regression tests might be helpful. I'm planning to add that feature to BU's rpc-tests.py (Core has had that feature for a while).
Until then, I've pull-requested some minor test runner improvements I made in MVF-BU, to allow execution of single tests with specific arguments, and allow one-off test runs of disabled tests. This should make it a little easier for testers to run disabled tests - now they don't have to edit the rpc-tests.py script code to temporarily enable a test.
Some nice changes from Classic have been ported to BU by kyuupichan: the improved command line parameter handling which prevents invalid parameters from being silently ignored at startup.
BTCfork
Finally got around to merging the latest Bitcoin Unlimited dev branch changes back into MVF-BU. Turned out to be something like 1200+ commits - who said BU isn't developmentally active?
Unfortunately, Travis is not yet completely happy with my merge - it seems some build environment changes or file movements have broken the test framework. My own 32-bit test system also needs to upgrade to a C++11 compiler to handle the upstream compiler requirement changes. All in all, these seems like minor things that can be cleaned up fairly quickly - over the course of the next few days. After that, MVF-BU will be back on track for further development.
Extension blocks or bigger regular blocks
After Gregory Maxwell's recent post on bitcoin-dev, the clamor for UASF SegWit seems to have died down a bit - maybe it's just on the back burner.
Bitcoin Unlimited seems to be within good reach of regular big blocks.
Extension Blocks looks like another promising idea, but quite a complex undertaking in comparison to straightforward big blocks.
Maybe Mr Bitcoin will place an order for both, and see which one the kitchen can deliver first?
Footnotes
[1] Original Japanese Silky Tofu image by DryPot (CC-BY-SA-3.0) at
https://en.wikipedia.org/wiki/Tofu#/media/File:Japanese_SilkyTofu_%28Kinugoshi_Tofu%29.JPG>