Topic: Bitcoin Core to Release SegWit in November - page 2. (Read 2281 times)

Who makes the decision of applying these new rules?

Awesome technology. Yes we need some modification and new technology in bitcoin to make it go on the same track with the modern technology and trends. I think if money investment in bitcoin happens then we will have price rise in bitcoin, i am not sure about the price.

Very interesting. Lets see what game theory says. Chinese miners seem to start evaluating on chain scaling again:

https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%3F__biz%3DMzIxNTA0NDQzMA%3D%3D%26mid%3D2651797972%26idx%3D1%26sn%3Da1dea9858a16b519258dee91176d2a8b%26chksm%3D8c65c1f5bb1248e382745b80b421bcc3fbbeb91cac20ae74c020315363d36756ad2f056325ec%26mpshare%3D1%26scene%3D1%26srcid%3D1022MmwyTZv0RtFUNVZwztBb%26from%3Dgroupmessage%26isappinstalled%3D0%23wechat_redirect&edit-text=&act=url

This thread escalated quickly into a word war between core and the big blockers. While I do support both movements and could see the advatages of both, I do like to see Segwit and then the Lightning Network be implemented first. I just think avoiding a hard fork for as long as possible as a safety measure must be the path to follow.

I am excited for 0.13.1, let's all hope everything all goes as planned and as smoothly as possible. From a speculator's standpoint, the potential for the Lightning Network is more promising and exciting for me.

No, I never said that. The 1% are legitimate bugs that are all low severity bugs. No severe or critical bugs have been found on a release, only on testing versions, if at all. Usually they are caught during the pull request code review time before being merged into the master branch.

Do you really think that they are rushing? Have you not seen the massive deadline slip? Core releases are always late. They are always released later than they plan to.

No, I wasn't avoiding your question, I was misunderstanding what you were saying. I do not think it is a bad thing for segwit to activate as soon as possible because I know that miners have been writing and testing segwit implementations for a few months now. I know that several core devs are also helping them write and test segwit so that it will be ready by the start time. However, I don't think it will activate until early 2017 at the earliest because there are still some miners who actively block it.

The latest date for the retarget period to have segwit active before Christmas is actually November 27th. November 24th is more than 4 weeks before Christmas. The actual earliest possible activation date is December 17th. That means that miners would need to start signalling on November 19th, 4 days after 0.13.1's release. After that it would be December 31st.

What makes a transaction on testnet a "false transaction"? That it has no monetary value? It is all technically the same. Furthermore, doing any of that testing on mainnet before activation would not even work because Segwit hasn't activated on mainnet.

All the tests that can be done "onchain" cannot actually be done on the mainnet blockchain without activation. All the tests that do not rely on the network can and already have been done (i.e. import/export keys, create and sign transactions, create witness addresses).

Most of them are implementing segwit in their own software, so there is no concern about that with running Bitcoin Core. Furthermore, Bitcoin Core uses a deterministic build process so the binaries are always the same. They can review the code changes since the last release and build it themselves and see if the hashes match. It's fairly simple to do, and part of the release process is to have multiple people do the deterministic build to check that all the hashes match.

so ignore the 1% because rush rush rush.. got ya

and i said for a activation pre christmas requires the retarget period to begin at the latest november 24th.. thus only giving 9 days to go from 0-95% to then be constant 95 for the period.
remember you replied to me about people who wanted a pre-christmas activation. so i was showing the requirements needed to get it.
which i see you avoided answering as a bad thing. but subtly suggesting people just just run it and trust it works

even without signalling support. pools can run a node on the side to ensure it can import/export keys, ensure theres no trojans, send tx's to see if they get rejected. again before flagging.
even without making a block. (imagine it usernode to usernode). use actual bitcoin transactions in combination with the segwit signing procedure to see if it reveals bugs when using real bitcoin transactions. as oppose to the false transactions on testnet. this can be done without activation.
to trule see how it interacts with 'legacy' nodes within the real blockchain (im talking tx relay not block creation)

there are many other tests that can be done onchain without having to actually activate it. but still do things onchain
i do love how you are brushing possible tests that can be done purely to push the rush rush rush rhetoric


even without signalling support. pools can run a node to ensure it can import/export keys, ensure theres no trojans, send tx's to see if they get rejected. etc etc etc again before flagging.

expecting pools to just change from legacy based on "trust" and not doing their own due diligence is a flaw in security.

its better to tell people to expect a spring 2017 .. rather then go with people shouting "activate pre-christmas"

I agree, they should test it very thoroughly. However, it is not necessary for them to have to wait for another couple of months while they test it because they are not going to start implementing and testing at the time of release. They started implementing and testing months ago. The spec has been public for a long time, discussions and reference implementations have been available for a long time. No respectable dev (for both miners and wallets) is going to sit on his ass and wait for an official release of a new thing before going to start work on stuff that will use it. All of that is already implemented or being implemented and being tested with testnet.

Also, you keep saying "test it on chain" "test it on mainnet". Do you realize that it is impossible to test a consensus change on the mainnet without actually deploying it? The only way to test it is to use a test network, either testnet, or a private test network. These test networks simulate the mainnet as best as possible. So while it's not mainnet, it's as good as it gets to being mainnet.

translate
I hope that pools review next months release. tests it onchain and ensures it does everything it should do as standard and able to do what it should BEFORE mining towards segwit activation when confident
though devs put a lot of effort into coding it, so it would suck if pools rush it because peer pressure. but rather done due to individual confidence in the code they are using based on their own review, and again not based on "trust" by others.

Where have I ever claimed that they are perfect? All I have said is that the software has been extensively tested and any more testing is unnecessary and just stalling.

Also, what is Core 0.132? It surely isn't 0.13.2 because that isn't a version that doesn't exist. 0.13.1 is the version with segwit.

The extensive testing has primarily been done with builds of the master branch of Bitcoin Core. You'll see it as version 0.13.99. 0.13.1 is made from backports of stuff that was already merged into master.

Again, I have never said that everything is perfect.

99% of the issues are all due to user error. The vast majority of those being unconfirmed transactions because people do not set proper fees. There have been very few legitimate bugs and those bugs are of low severity, just minor nuisances. Those bugs are also fixed very quickly when reported.

I have never called it perfect nor utopia. All of us developers know that review and testing is extremely important. That is part of the reason why segwit is behind the original timeline.

It only requires 95% over one retarget period, starting on November 15th.


What are you even asking? In order to signal segwit support, they have to stop using their legacy nodes and use segwit capable ones. If there is a "rollback" (which is impossible without a hard fork), then they have previous versions they can use unless they are so stupid that they don't use a VCS.

It is impossible to test on the mainnet without it being active. It is tested on the network most analogous to the mainnet which is the testnet.

I hope that everyone collaborates and starts mining towards segwit activation as soon as possible and we avoid any idiotic fights like the ViaBTC morons going against it. They put a lot of effort into codding it so it would suck if they delay it because of that.

yes gmaxwell and sipa can play with themselves. but to be truly USEFUL merchants need to be involved. you know where bitcoin is actually useful (hint:buy things).

and over them months many things have changed. and in all those months it has not interacted with 7 years of historic data or mutiple pools or thousands of users.
sandbox tests does not equal utopian perfection in the real world. miners will still want to review and test the public release it before implementing it as their full live main node.
afterall the code next month is not the same code as last month, they are still tweaking it.

if the fiat lovers think that segwt should run by christmas. then where is the screams of grace period they were crying about..
oh wait they need no grace because they want everyone to just run core..
like i said many times in the past, the network should remain diverse. giving time for those to review and adapt the code into their own implementations or atleast test it on bitcoins mainnet before waving their hands in the air


tested on testnet.. tested on testnet.. tested on testnet..
thats like saying testing it on Clams or other alts... meaningless
it needs testing on bitcoin main net before settling minds can then say its safe to then flag consensus.
so stop thinking it should be active by christmas and atleast settle for atleast spring next year (if people are doing the right thing and checking)

EG even windows gets alpha and beta tested. but even after retail release it still ends up with people having system crashes and patches needed.
never claim anything is utopia until is has had time to run on the platform it is meant for.

i do find it funny how achow101 proclaims 0.132 to be perfect and well tested. though not running live on mainnet..
i presume 0.12 and 0.131 are perfect too... where even after release.. even after running there are still issues
 even achow101 spends alot of time in the support category dealing with crashes and many issues. with 'legacy' versions.
kind of strange to even need a support section if everything is perfect.. right?

and now you see why its safer to review test and use it before hailing it utopia to garner 95% utility

ok lets put my waffle into a simple question.
knowing it requires 95% over a month. where start of month needs to be at 95%, throughout month, and end of month needs to be 95% (3831 of 4032blocks)
are you saying mining pools should drop their "legacy" (nodes as you call it) at the latest november 24th and use only core 0.132 from november 24th because its safe to all rush to run before christmas? honest answer please

Franky, do us all a favor and stop spreading FUD.

Not necessarily merchants. Users too can use segwit. The sender can take advantage of segwit even if the receiver is not segwit capable.

That's a load of bullshit. Segwit has been and still is being continuously tested on the testnet. It is not as if the software was just written and now is being released. It was written several months ago and been in continuous testing since then. Furthermore several of the Core devs have reached out to miners, merchants, and wallet devs to help them with implementing and testing segwit.

What are you saying? The vast majority of full nodes run some version of core. Do you mean that everyone has to have their own implementation? Also, the only other true alternative full node implementation is btcd and they have already completed implementing segwit.

Or they care about testing and have tested it themselves on testnet and are sure that the software is secure. I know a lot of the developers want it out ASAP and have extensively tested it on testnet and are absolutely sure about its security.

remember folks. its not just getting miners to allow tx's into blocks. its also merchants changing their wallets to have new keys to be able to "spend" funds if they too want to use segwit..

if its active by christmas then miners have not done any testing on the side.
if merchants have the nodes before christmas, then they too have not done any testing.
if 95% of "segwit" is purely core0.132 and not peoples own implementations of the code. there is no point having distribution apart from worries of power cuts and data loss.

seems the people that want it running before christmas are sounding like they care less about testing, care less about security and more about bitcoin price.

and ontop of that the majority that have this careless nature are mostly people that do not even run a full node in the first place
kind of funny that the "slow careful and smart testing mindset defending cores delays" turns into "before christmas, before christmas before christmas"

anyone wanting it active and in full use before christmas should be tagged as fiat loving bitcoin security hating nutters.
if those same people have previously shouted how any other proposal needs months of review and months of grace period. should be tagged as hypocritical fiat loving bitcoin security hating nutters

I cant wait for this release, i hope it sends bitcoin to the moon.  Although i would have to say if it make bitcoin work better i will be more impressed......

I would have actually wanted to see it sooner, so that the bugs.. if there were any, could be sorted out before the Xmas season started.

We do not want to be in a testing phase, when the holidays hits us and we start buying presents for Xmas. Oh, better late than never, or

so they say. 

No. Segwit is not active yet. Bitcoin Core 0.13.1 will be advertising the NODE_WITNESS service though in preparation for segwit activation. As of now, it means nothing.

segwit are already in service on Bitcoin network : https://bitcointalksearch.org/topic/m.16624848

He's talking to Bitcoin "Classic" programmer Tom Zander about his inability to grasp the safety of the the length of the prescribed fallow period during the soft fork (the 2016 blocks where non-miners upgrade to 12.2/13.1 after lock-in). Even Tom's (former?) lead developer Gavin Andersen disagreed with him, lol

I'm trying to find more info on a point Matt Corallo brought up.
Hopefully someone who knows a lot more about Bitcoin programming can explain this further-

technically.
if 92% of pools started accepting segwit transactions into blocks (early) all that happens is the 8% wont accept the tx. so lets say the next block was solved by the 8%.. it wont contain the tx and ur waiting for a pool in the 92% to solve a block. basically the same game theory as 8% reject tx's with low fees and 92% accept any tx. its just a waiting game for a pool that likes your tx to solve a block including it.

but pools should wait for 95%+ to ensure there is little orphan risk if a bug causes disagreement, thus an orphan across the network

what concerns me is if pools just throw 0.132 online without review.
what concerns me is if pools just throw 0.132 without taking the code and adapting it to their own diverse code base
what concerns me is if pools just throw 0.132 online and everyone is running just one brand.