Topics: updates for December and four new topics #318

harding · 2020-01-13T21:40:09Z

Updates and new topics, plus a couple minor QA changes.

bitschmidty

@harding thanks for giving us something new each month in the topics section! (and for keeping references up to date each month!) Great work

tested changes to schema checker catch errors (including topic date)
tested schema checker outputs failing files
confirm that default topic date is indeed ridiculous
went through new topic links(mentions) in December and sanity checked destination url (I noticed several "bad anchor links" and pointed them out individually, not to be annoying but because I was not sure if they were best addressed with a single fix or individually)
reviewed each of the 4 new topics

_topics/en/watchtowers.md

_topics/en/bech32.md

_topics/en/assumeutxo.md

_topics/en/miniscript.md

_topics/en/signet.md

_topics/en/bip70-payment-protocol.md

_topics/en/eclipse-attacks.md

_topics/en/op_codeseparator.md

harding · 2020-01-15T18:56:02Z

I noticed several "bad anchor links" and pointed them out individually, not to be annoying but because I was not sure if they were best addressed with a single fix or individually

Whoops, these all appear to be bugs in the content of Newsletter 78. I'll fix them in a separate PR presently. Great catch, thanks!

- Makefile: report which files failed - Topics schema: detect non-date strings

harding · 2020-01-15T23:58:38Z

Forced pushed rebased on master after #321 was merged and verified this fixes all the bad anchors. I'll address @bitschmidty's other feedback tomorrow. Thanks!

jonatack

Valuable additions to the topics @harding! Looked at the content without checking the links and jekyll mechanics.

jonatack · 2020-01-16T12:09:14Z

_topics/en/anchor-outputs.md

+extended_summary: |
+  Each time the balance changes in an LN channel, a *commitment
+  transaction* is created and signed by the participating parties.  The
+  transaction is only broadcast if one party decides they want to


nit: can optionally remove "they want"

_topics/en/anchor-outputs.md

jonatack · 2020-01-16T12:16:30Z

_topics/en/bip70-payment-protocol.md

+      using Bluetooth or NFC to pay even if they didn't currently have
+      an Internet connection
+
+  It's main disadvantage was that it required spenders support the SSL


s/It's/Its/

jonatack · 2020-01-16T12:18:01Z

_topics/en/bip70-payment-protocol.md

+  in Bitcoin and a complex Public Key Infrastructure (PKI) system.  In
+  practice, this meant software needed to include a library such as
+  OpenSSL as a dependency.  Bugs in that library could then be used to
+  exploit Bitcoin programs, such as the OpenSSL [heartbleed][]


Could be wrong but it seems to me the comma before "such" should be removed here.

I think it's a non-restrictive appositive and so is properly set off by a comma.

jonatack · 2020-01-16T12:56:46Z

_topics/en/op_codeseparator.md

+  a pre-release version of Bitcoin [may][pre todd1] have [supported][pre
+  todd2] this capability.
+
+  In [Bitcoin 0.3.7][] (July 2010), a fix to the [1 OP_RETURN][] bug was


Awesome history section.

Officially, it seems there were two consecutive CVEs fixed on 31 July and 1 Aug 2010:

[CRITICAL] Disabled OP_RETURN - v0.3.5 (Jul 28th, 2010)

[CRITICAL] Separate scriptSig & scriptPubKey eval - v0.3.7 (Aug 1st, 2010)

Did some sleuthing into this last June (with @jnewbery's help) and it turned out the OP_RETURN bug was fixed in bitcoin/bitcoin@a75560d which has the confusing commit message "reverted makefile.unix wx-config -- version 0.3.6".

In that commit, the Bitcoin version appears to have been bumped from 0.3.3 to 0.3.6 here.

I made a presentation on the OP_RETURN issue last June (and would be happy to correct it if needed):

https://drive.google.com/file/d/1nhxsJiSZh_Nhn0Oql9yCfAuryAy8Ed3W/view

Yeah, a75560d makes an executed OP_RETURN terminate the script in failure and so fixes the specific 1 OP_RETURN bug, but I don't think it fixes the general class of fault. For example, imagine your scriptPubKey was:

<pubkey> OP_CHECKSIGVERIFY OP_IF ... OP_ENDIF

Now imagine an attacker creates the following scriptSig:

OP_TRUE OP_FALSE OP_IF

If those scripts are concatenated like in Bitcoin 0.1, then the attacker's un-executed IF wraps the CHECKSIG and the other IF statement, allowing the attacker to spend the money without providing a signature. (I think that works, but I'm not exactly sure of Script's block-matching rules within unexecuted blocks.)

So I think we're ok linking to 6ff5f718b, but I'll see if I can clarify to mention that there were several commits involved in the fix.

jonatack · 2020-01-16T13:41:02Z

_topics/en/op_codeseparator.md

+  In [Bitcoin 0.3.7][] (July 2010), a fix to the [1 OP_RETURN][] bug was
+  deployed as a stealth [hard fork][lukejr says sipa confirms hf] that
+  changed how the scripts were evaluated, including discontinuing the
+  use of the virtual `OP_CODESEPARATOR`.  However, the opcode itself


Separate scriptSig & scriptPubKey eval (OP_CODESEPARATOR) bug fixed in commit bitcoin/bitcoin@73aa262 "fixed segfault in bignum.h, additional security limits, refactoring"

Yeah, that's identical to the commit this PR already links to 6ff5f718b6a67797b2b3bab8905d607ad216ee21 (all, or almost all, the svn commits imported into git have duplicate commits, probably the result of some bug in the import script or someone running the import twice).

jonatack · 2020-01-16T13:30:42Z

_topics/en/op_codeseparator.md

+  part of the proposal may be changed.
+
+  Additionally, Bitcoin Core 0.16.1 and later will not relay or mine any
+  transaction that uses `OP_CODESEPARATOR` in a legacy script.


Great section. I'm trying to figure where commit bitcoin/bitcoin@9dabfe4 in PR bitcoin/bitcoin#11423 fits into all this.

Add constant scriptCode policy in non-segwit scripts This disables OP_CODESEPARATOR in non-segwit scripts (even in an unexecuted branch), and makes a positive FindAndDelete result invalid. This ensures that the scriptCode serialized in SignatureHash() is always the same as the script passing to the EvalScript.

I'm not sure I understand your comment, but jl2012's commit basically makes the change proposed in the consensus cleanup section of the topic post and for the same reason, although obviously Lau's commit is only a standardness change and not consensus.

jonatack · 2020-01-16T13:41:02Z

_topics/en/op_codeseparator.md

+  In [Bitcoin 0.3.7][] (July 2010), a fix to the [1 OP_RETURN][] bug was
+  deployed as a stealth [hard fork][lukejr says sipa confirms hf] that
+  changed how the scripts were evaluated, including discontinuing the
+  use of the virtual `OP_CODESEPARATOR`.  However, the opcode itself


Separate scriptSig & scriptPubKey eval (OP_CODESEPARATOR) bug fixed in commit bitcoin/bitcoin@73aa262 "fixed segfault in bignum.h, additional security limits, refactoring"

harding · 2020-01-16T14:49:38Z

Forced pushed edits for @bitschmidty and @jonatack feedback (thanks!)

bitschmidty · 2020-01-17T16:38:40Z

@harding all of my original comments have been marked resolved. thanks for the quick turnaround!

I will do another review after any remaining reviews.

_topics/en/covenants.md

_topics/en/erlay.md

_topics/en/miniscript.md

_topics/en/multipath-payments.md

_topics/en/bip70-payment-protocol.md

_topics/en/op_codeseparator.md

harding · 2020-01-27T20:55:22Z

Force pushed edits for @jnewbery feedback, thanks!

jnewbery · 2020-01-27T21:13:45Z

All my comments are addressed. Thanks Dave!

jnewbery added the topics Topics indexes label Jan 13, 2020

harding mentioned this pull request Jan 14, 2020

[Topics] Change date index layout, add stats, add internal page links #320

Merged

bitschmidty reviewed Jan 15, 2020

View reviewed changes

harding mentioned this pull request Jan 15, 2020

Eliminate duplicate anchors from all pages #321

Merged

harding added 2 commits January 15, 2020 18:51

Tests: improve schema reporting and date detection

c1a3766

- Makefile: report which files failed - Topics schema: detect non-date strings

Contrib/New Topic: make date ridiculous to ensure it gets updated

1acea20

harding force-pushed the 2020-01-topics-update branch from 7d19e62 to 52b908b Compare January 15, 2020 23:56

jonatack reviewed Jan 16, 2020

View reviewed changes

harding force-pushed the 2020-01-topics-update branch from 52b908b to e72eefd Compare January 16, 2020 14:49

jnewbery reviewed Jan 21, 2020

View reviewed changes

harding added 2 commits January 27, 2020 14:53

Topics: updates for December 2019

f493fed

New Topics: anchor outputs, BIP70, eclipse attacks, OP_CODESEPARATOR

d0cf26e

harding force-pushed the 2020-01-topics-update branch from e72eefd to d0cf26e Compare January 27, 2020 20:54

jnewbery merged commit f3f9b7a into bitcoinops:master Jan 27, 2020

jnewbery mentioned this pull request Jan 27, 2020

Newsletters: add 82 (2020-01-29) #328

Merged

2 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Topics: updates for December and four new topics #318

Topics: updates for December and four new topics #318

harding commented Jan 13, 2020

bitschmidty left a comment

harding commented Jan 15, 2020

harding commented Jan 15, 2020

jonatack left a comment

jonatack Jan 16, 2020

jonatack Jan 16, 2020

jonatack Jan 16, 2020

harding Jan 16, 2020

jonatack Jan 16, 2020

harding Jan 16, 2020

jonatack Jan 16, 2020 •

edited

Loading

harding Jan 16, 2020

jonatack Jan 16, 2020

harding Jan 16, 2020

jonatack Jan 27, 2020

jonatack Jan 16, 2020 •

edited

Loading

harding commented Jan 16, 2020

bitschmidty commented Jan 17, 2020

harding commented Jan 27, 2020

jnewbery commented Jan 27, 2020

Topics: updates for December and four new topics #318

Topics: updates for December and four new topics #318

Conversation

harding commented Jan 13, 2020

bitschmidty left a comment

Choose a reason for hiding this comment

harding commented Jan 15, 2020

harding commented Jan 15, 2020

jonatack left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jonatack Jan 16, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jonatack Jan 16, 2020 • edited Loading

Choose a reason for hiding this comment

harding commented Jan 16, 2020

bitschmidty commented Jan 17, 2020

harding commented Jan 27, 2020

jnewbery commented Jan 27, 2020

jonatack Jan 16, 2020 •

edited

Loading

jonatack Jan 16, 2020 •

edited

Loading