[bitcoin-dev] BIP149 timeout-- why so far in the future?

[bitcoin-dev] BIP149 timeout-- why so far in the future?

[Martijn Meijering]

Jorge Timón wrote:

"My preference would be a bip149 proposal that could be merged and
released now, but some people complain that would require more
testing, because *if you deploy bip149 and then sw gets activated pre
nov15, then you want bip149 nodes to use the old service bit for
segwit*, not the new one (you would use that one if it activates post
nov15, so that pre-bip149 nodes don't get confused)."
(emphasis added)

Why not just make sure BIP 149 will never activate unless BIP 141 has
expired unsuccessfully? If BIP 141 should unexpectly activate, then
BIP 149 nodes would notice and act as pre-SegWit nodes indefinitely,
but remain in consensus with BIP 141 nodes.

It might be slightly less convenient for BIP 149 users to upgrade
again, but then at least we could start deploying BIP 149 sooner.

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Jorge Timón]

On Sun, Jun 11, 2017 at 3:44 PM, Martijn Meijering via bitcoin-dev
<bitcoin-dev@lists.linuxfoundation.org> wrote:
> Jorge Timón wrote:
> Why not just make sure BIP 149 will never activate unless BIP 141 has
> expired unsuccessfully?

Right, that would be part of it, as well as not removing the BIP141
deployment with bip9.
See https://github.com/jtimon/bitcoin/commit/62efd741740f5c75c43d78358d6318941e6d3c04

> If BIP 141 should unexpectly activate, then
> BIP 149 nodes would notice and act as pre-SegWit nodes indefinitely,
> but remain in consensus with BIP 141 nodes.
>
> It might be slightly less convenient for BIP 149 users to upgrade
> again, but then at least we could start deploying BIP 149 sooner.

No, if segwit activates pre nov15, bip149 nodse can detect and
interpret that just fine.
The problem if it activates post nov15, then you need a separate
service bit in the p2p network, for pre-BIP149 will think sw hasn't
activated while post-BIP149 would know it has activated.

If you release it only after nov15, you don't need to test
compatibility between the two for neither of this two cases.
Or do you? Actually you only save testing the easier case of pre-nov15
activation.


> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Jorge Timón]

The current proposal assumes that bip149 would only be merged and
released after nov15, so there's not time in one day.

My preference would be a bip149 proposal that could be merged and
released now, but some people complain that would require more
testing, because if you deploy bip149 and then sw gets activated pre
nov15, then you want bip149 nodes to use the old service bit for
segwit, not the new one (you would use that one if it activates post
nov15, so that pre-bip149 nodes don't get confused).

I was slowly modifying shaolinfry's code to try to code that, but I'm
currently not working on it because there doesn't seem there's a lot
of interest in releasing bip149 before nov15...

https://github.com/jtimon/bitcoin/commits/b15-shaolinfry-bip149


On Sun, Jun 11, 2017 at 7:48 AM, Ryan Grant via bitcoin-dev
<bitcoin-dev@lists.linuxfoundation.org> wrote:
> Is there any reason that BIP149 activation on November 16th would
> cause a problem?
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Ryan Grant]

Is there any reason that BIP149 activation on November 16th would
cause a problem?

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Rusty Russell]

Matt Corallo <lf-lists@mattcorallo.com> writes:
> A more important consideration than segwit's timeout is when code can be
> released, which will no doubt be several months after SegWit's current
> timeout.

I was assuming it would be included in the next point release.

Cheers,
Rusty.

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Matt Corallo]

A more important consideration than segwit's timeout is when code can be
released, which will no doubt be several months after SegWit's current
timeout.

Greg's proposed 6 months seems much more reasonable to me, assuming its
still many months after the formal release of code implementing it.

Matt

On 05/24/17 04:26, Rusty Russell via bitcoin-dev wrote:
> Gregory Maxwell via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org>
> writes:
>> Based on how fast we saw segwit adoption, why is the BIP149 timeout so
>> far in the future?
>>
>> It seems to me that it could be six months after release and hit the
>> kind of density required to make a stable transition.
> 
> Agreed, I would suggest 16th December, 2017 (otherwise, it should be
> 16th January 2018; during EOY holidays seems a bad idea).
> 
> This means this whole debacle has delayed segwit exactly 1 (2) month(s)
> beyond what we'd have if it used BIP8 in the first place.
> 
> Cheers,
> Rusty.
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
> 

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[shaolinfry]

[-- Attachment #1: Type: text/plain, Size: 1150 bytes --]

I agree the date can be brought forward. FWIW, I originally set the date far out enough that people wouldn't immediately fixate on the date and rather look at the meat of the proposal instead.

Given that we saw around 70% of nodes upgrade to BIP141 in around 5/6 months, I dont see any reason why we cant reduce the date to being 6 months or less from Nov. Given people are starving for segwit to the point of running BIP148, there is good evidence the community will upgrade in record time to BIP149.

Sent from [ProtonMail](https://protonmail.com), Swiss-based encrypted email.

-------- Original Message --------

Based on how fast we saw segwit adoption, why is the BIP149 timeout so
far in the future?

It seems to me that it could be six months after release and hit the
kind of density required to make a stable transition.

(If it were a different proposal and not segwit where we already have
seen what network penetration looks like-- that would be another
matter.)
_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

[-- Attachment #2: Type: text/html, Size: 1711 bytes --]

Re: [bitcoin-dev] BIP149 timeout-- why so far in the future?

[Rusty Russell]

Gregory Maxwell via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org>
writes:
> Based on how fast we saw segwit adoption, why is the BIP149 timeout so
> far in the future?
>
> It seems to me that it could be six months after release and hit the
> kind of density required to make a stable transition.

Agreed, I would suggest 16th December, 2017 (otherwise, it should be
16th January 2018; during EOY holidays seems a bad idea).

This means this whole debacle has delayed segwit exactly 1 (2) month(s)
beyond what we'd have if it used BIP8 in the first place.

Cheers,
Rusty.

[bitcoin-dev] BIP149 timeout-- why so far in the future?

[Gregory Maxwell]

Based on how fast we saw segwit adoption, why is the BIP149 timeout so
far in the future?

It seems to me that it could be six months after release and hit the
kind of density required to make a stable transition.

(If it were a different proposal and not segwit where we already have
seen what network penetration looks like-- that would be another
matter.)