Re: [Bitcoin-development] Missing fRelayTxes in version message

[Mike Hearn <mike@plan99.net>]

[-- Attachment #1: Type: text/plain, Size: 8730 bytes --]

The protocol version was bumped when Bloom filtering was added so there's
not much point bumping it again - you have to handle the old clients no
matter what. Nobody brought this up as an issue when the BIP or code was
first written and as you can see from main.cpp, it was done this way to be
consistent with how other version fields are handled:

        if (!vRecv.empty())
            vRecv >> addrFrom >> nNonce;
        if (!vRecv.empty())
            vRecv >> pfrom->strSubVer;
        if (!vRecv.empty())
            vRecv >> pfrom->nStartingHeight;
        if (!vRecv.empty())
            vRecv >> pfrom->fRelayTxes; // set to true after we get the
first filter* message

The existence of the nStartingHeight field for instance depends on the
message length and not anything else.

Anyway, are you really asking for the protocol to be changed to work around
an issue specific to how you wrote your parsing code? This is the first
time anyone has suggested this minor detail is a problem. It doesn't
present any issues for the C++ code or bitcoinj where message objects know
their own length at parse time.



On Thu, Jun 20, 2013 at 8:20 AM, Turkey Breast <turkeybreast@yahoo.com>wrote:

> I never said that Bitcoin message field lengths should always be the same.
> But before this change they certainly were constant per protocol version.
> All I'm saying is that optional lengths shouldn't be used (a field exists
> or not) and for every field change, the protocol version should be upgraded.
>
> Now that fRelayTxes is part of the protocol, the version number should be
> upgraded to reflect this fact.
>
>   ------------------------------
>  *From:* Mike Hearn <mike@plan99.net>
> *To:* Paul Lyon <pmlyon@hotmail.ca>
> *Cc:* Turkey Breast <turkeybreast@yahoo.com>; "
> bitcoin-development@lists.sourceforge.net" <
> bitcoin-development@lists.sourceforge.net>
> *Sent:* Wednesday, June 19, 2013 3:20 PM
>
> *Subject:* Re: [Bitcoin-development] Missing fRelayTxes in version message
>
> If you want to criticise the Bitcoin protocol for sloppyness, the variable
> length of some messages isn't where I'd start.
>
> Note that ping has the same issue, its length has changed over time to
> include the nonce.
>
> If your parser can't handle that kind of thing, you need to fix it. The
> protocol has always worked that way.
>
>
>
> On Wed, Jun 19, 2013 at 3:03 PM, Paul Lyon <pmlyon@hotmail.ca> wrote:
>
>  I’m also running into this exact same issue with my parser, now I
> understand why the relay field behavior I was seeing doesn’t match the wiki.
>
> So to parse a version message, you can’t rely on the protocol version? You
> have to know how long the payload is, and then parse the message
> accordingly? I agree with Turkey Breast, this seems a bit sloppy to me.
>
> Paul
>
> P.S. I’ve never used a dev mailing list before and I want to get involved
> with the Bitcoin dev community, so let me know if I’m horribly violating
> any mailing list etiquette. 😊
>
> *From:* Mike Hearn
> *Sent:* ‎Wednesday‎, ‎June‎ ‎19‎, ‎2013 ‎7‎:‎43‎ ‎AM
> *To:* Turkey Breast
> *Cc:* bitcoin-development@lists.sourceforge.net
>
> Bitcoin-Qt on master does send it now although it doesn't affect anything,
> but as old pre-filtering versions will continue to exist, you'll always
> have to be able to deserialize version messages without it.
>
> Bitcoin version messages have always had variable length, look at how the
> code is written in main.cpp. If you didn't experience issues until now all
> it means is that no sufficiently old nodes were talking to yours.
>
> The standard does not say it should appear. Read it again - BIP 37 says
> about the new version message field:
> If false then broadcast transactions will not be announced until a
> filter{load,add,clear} command is received. *If missing or true*, no
> change in protocol behaviour occurs.
>
>
> On Wed, Jun 19, 2013 at 12:33 PM, Turkey Breast <turkeybreast@yahoo.com>wrote:
>
> It's a problem if you work with iterators to deserialize the byte stream.
> Even failing that, it's just sloppy programming. What happens in the future
> when new fields are added to the version message? It's not a big deal to
> say that this protocol version has X number of fields, that (higher)
> protocol version message has X + N number of fields. Deterministic number
> of fields per protocol version is sensical and how Bitcoin has been for a
> long time.
>
> And yes, it was a problem for me that caused a lot of confusion why this
> byte didn't exist in many version messages despite the standard saying it
> should and the code in bitcoind indicating it should. Nowhere was this
> written. It doesn't help other implementations to have an unclear behaviour
> that depends on some magic from one implementation.
>
>   ------------------------------
>  *From:* Mike Hearn <mike@plan99.net>
> *To:* Turkey Breast <turkeybreast@yahoo.com>
> *Cc:* "bitcoin-development@lists.sourceforge.net" <
> bitcoin-development@lists.sourceforge.net>
> *Sent:* Wednesday, June 19, 2013 11:39 AM
>
> *Subject:* Re: [Bitcoin-development] Missing fRelayTxes in version message
>
> It has to be optional because old clients don't send it, obviously.
>
> Why is this even an issue? There's no problem with variable length
> messages in any codebase that I'm aware of. Is this solving some actual
> problem?
>
>
> On Wed, Jun 19, 2013 at 12:30 AM, Turkey Breast <turkeybreast@yahoo.com>wrote:
>
> That's me. I never said to make all messages fixed length. I said to make
> a fixed number of fields per protocol. So given a protocol version number,
> you know the number of fields in a message. This is not only easier for
> parsing messages, but just good practice. I don't see why a 1 byte flag
> needs to be optional anyway.
>
>   ------------------------------
>  *From:* Mike Hearn <mike@plan99.net>
> *To:* Turkey Breast <turkeybreast@yahoo.com>
> *Cc:* Bitcoin Dev <bitcoin-development@lists.sourceforge.net>
> *Sent:* Tuesday, June 18, 2013 9:48 PM
> *Subject:* Re: [Bitcoin-development] Missing fRelayTxes in version message
>
> It's not a bug (although there was recently a change to make bitcoind/qt
> always send this field anyway).
>
> I don't know where Amir is going with BIP 60. Version messages have always
> been variable length. There's nothing inherent in the Bitcoin protocol that
> says all messages are fixed length, indeed, tx messages are allowed to have
> arbitrary data appended after them that gets relayed.
>
>
> On Tue, Jun 18, 2013 at 7:45 PM, Turkey Breast <turkeybreast@yahoo.com>wrote:
>
> See this BIP. I'm not sure if this is a bug or what, but it would be good
> if messages always had a fixed number of fields per protocol version.
>
> https://en.bitcoin.it/wiki/BIP_0060#Code_Updates
>
> This BIP details everything that needs to be done and proposes a protocol
> upgrade.
>
>
> ------------------------------------------------------------------------------
> This SF.net <http://sf.net/> email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
>
>

[-- Attachment #2: Type: text/html, Size: 17746 bytes --]