bolt7: dns support by m-schmoock · Pull Request #4829 · ElementsProject/lightning

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock · 2021-09-29T15:38:26Z

This implements RFC lightning/bolts#911 to allow annoucements of DNS hostnames.
The feature can only be used with EXPERIMENTAL.
The bold quotes are disabled in the topmost commit to please CI.

Adds:

related fixes and changes (i.e. announce up to two addresses of the same type)
feature itself
and tests

m-schmoock · 2021-10-09T08:44:28Z

I think this is ready for review now.
Still not 100% happy with the way iffy way I changed opt_add_addr_withtype in lightningd/options.c, maybe you see a better way.

cdecker · 2021-10-09T13:33:39Z

Note: the BOLT checker complains about my not yet merged changes on BOLT7. Not sure how we solve this chicken and egg when implementing upcoming protocol changes.
* `5`: DNS hostname; data = `[byte:len][len*byte:hostname][u16:port]` (length up to 258)

Yeah, there is a special syntax that specifies a separate branch of the lightning-rfc repo to check against, though I'm not remembering how that worked rn

m-schmoock · 2021-10-09T15:52:58Z

Thx for the tip, looks like we can set the BOLTVERSION in the Makefile for that.
Update: I cant create branches in the RFC repo and had other issues with it so I went back to removing the BOLT quotes for now just to get past CI.

m-schmoock · 2021-11-17T10:26:36Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR. I can even reproduce this on my local machine. So I guess there is some condition in master branch that causes it...

ZmnSCPxj · 2021-11-17T10:30:07Z

Currently CI fails on wallet/test/run-db.c:209 which is very likely not related to this PR.

Kicked CI, just in case this is just test flakiness.

m-schmoock · 2021-11-17T12:13:09Z

Kicked CI, just in case this is just test flakiness.

@ZmnSCPxj The problem remains in wallet/test/run-db.c:209

… sync Incase the error handling happening after the quted line is non-critical: ``` return tal_fmt(NULL, "Unable to parse address '%s': %s", arg, err_msg); ``` we should not expand the proposed_listen_announce array without adding a proposed_wireaddr. So we move the expand of proposed_listen_announce to the location where we also expand the proposed_wireaddr. Changelog-None

Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell · 2021-11-28T00:34:46Z

Ack 930d25b

Rebased onto master, fixed trivial conflicts.

m-schmoock · 2022-02-23T08:09:02Z

Seems like rust-lightning is going to add support for this as the first other implementation: lightningdevkit/rust-lightning#1329

m-schmoock force-pushed the bolt7/dns branch 7 times, most recently from 30e9c00 to 38d9ea0Compare October 7, 2021 11:01

m-schmoock force-pushed the bolt7/dns branch 4 times, most recently from 1193fae to 6e1c37aCompare October 8, 2021 13:39

m-schmoock commented Oct 8, 2021
View reviewed changes

Comment threadconnectd/connectd.c Outdated

m-schmoock force-pushed the bolt7/dns branch 2 times, most recently from 63b35de to b7d098fCompare October 9, 2021 08:40

m-schmoock marked this pull request as ready for review October 9, 2021 08:41

m-schmoock requested a review from cdecker as a code owner October 9, 2021 08:41

m-schmoock requested a review from rustyrussell October 9, 2021 08:42

m-schmoock force-pushed the bolt7/dns branch 3 times, most recently from 964179d to d2ccfefCompare October 9, 2021 09:03

m-schmoock force-pushed the bolt7/dns branch from a6be613 to 2631f4cCompare October 9, 2021 15:52

m-schmoock force-pushed the bolt7/dns branch 6 times, most recently from bd48566 to 0b68251Compare October 13, 2021 14:21

m-schmoock force-pushed the bolt7/dns branch from de679ab to bb3fcf3Compare November 15, 2021 19:28

cdecker requested a review from rustyrussell November 15, 2021 20:10

rustyrussell requested changes Nov 17, 2021
View reviewed changes

Comment threadconnectd/connectd.c
Comment threadconnectd/connectd.c Outdated
Comment threadconnectd/connectd.c
Comment threadconnectd/connectd.c Outdated

m-schmoock force-pushed the bolt7/dns branch 3 times, most recently from 9598e74 to 1edb314Compare November 17, 2021 10:13

m-schmoock force-pushed the bolt7/dns branch 2 times, most recently from 195cbc6 to 1c83cfdCompare November 23, 2021 14:50

m-schmoock requested a review from rustyrussell November 23, 2021 15:43

rustyrussell force-pushed the bolt7/dns branch from 1c83cfd to bbd2ac9Compare November 28, 2021 00:28

m-schmoock added 9 commits November 28, 2021 10:59

options: make --disable-dns an early arg
185f28c

chore: gitignore gdb_history
4bd6573

options: fix respect --disable-dns when parsing wireaddress
dbfee2b
Changelog-Fixed: Options: Respect --always-use-proxy AND --disable-dns when parsing wireaddresses to listen on.

wireaddr: adds helper is_ipaddr, is_toraddr and is_dnsaddr
f99aa1d

bolt7: allow announcement of ADDR_TYPE_DNS
55825cb

connectd: resolve ADDR_TYPE_DNS
5cfcc6c
This will resolve ADDR_TYPE_DNS wireaddr by expanding connect->addrs with one new wireaddr ADDR_INTERNAL_WIREADDR per DNS result and calling recursion

pytest: test connecting to a DNS only announced node
6bc3f57

chore: use EXPERIMENTAL for BOLT7 DNSElementsProject#911
930d25b
Changelog-EXPERIMENTAL: Ability to announce DNS addresses

rustyrussell force-pushed the bolt7/dns branch from bbd2ac9 to 930d25bCompare November 28, 2021 00:34

rustyrussell approved these changes Nov 29, 2021
View reviewed changes

rustyrussell merged commit a3ea9fd into ElementsProject:masterNov 29, 2021

m-schmoock mentioned this pull request Nov 30, 2021
Send remote address in init ACINQ/eclair#1973
Merged

m-schmoock deleted the bolt7/dns branch November 30, 2021 17:50

remyers mentioned this pull request Apr 12, 2022
Support DNS hostnames in node announcements ACINQ/eclair#2234
Merged
3 tasks