libera #monero-dev

13 May 2022

Official repo: github.com/monero-project/monero || Don’t ask to ask, just ask:) || Monero GUI dev is at #monero-gui || Monero website dev is at #monero-site || Non-dev questions -> #monero, please || No fork help, you’re on your own

• 01:27
  dsanpang
  Hello, do you have any staff?
• 01:35
  plowsof[m]
  hello dsanpang what is the nature of your issue? This is the Monero Developer channel. I will direct you to the correct place where one of our trained staff members can assist you further.
• 01:37
  dsanpang
  I saved the address on monero wallets, but the address changed when I clicked the save button.
• 01:39
  plowsof[m]
  I'm sorry to hear that, i will contact you via direct message to further diagnose the issue, and get you up and running!
• 01:40
  dsanpang
  thanks
• 06:50
  sech1
  I can't build the latest master: paste.debian.net/hidden/9f69596b
• 06:50
  sech1
  Does anyone know what's wrong?
• 07:24
  moneromooo
  sech1: I'd try removing all CMakeCache.txt from build, and make again. Looks like it might be confused between in-tree and system unbound.
• 07:36
  sech1
  It was a build done right after a fresh git clone
• 07:38
  sech1
  I tried to VMs (Ubuntu 20.04 and Ubuntu 18.04), same result
• 07:38
  sech1
  *two VMs
• 07:50
  moneromooo
  Then I'd suggest checking what cmake thinks it found for unbound, then the link line, and see if those are consistent.
• 07:50
  moneromooo
  Double check cmake var names (ie, X_LIBRARY vs X_LIBRARIES) in case it changed recently.
• 07:54
  sech1
  Those two VMs are quite messy, I installed and uninstalled a lot of software there. I'll try with a fresh VM later today.
• 10:47
  sech1
  moneromooo this is on a fresh Ubuntu 20.04 VM after following the build steps from Github: paste.debian.net/hidden/b16cf76c
• 11:04
  moneromooo
  It's found a static libunbound, so it seems it needs linking against libevent, which is automatic for shared libs. I see depends configures without libevent. Your OS did the opposite I guess.
• 11:05
  moneromooo
  So either build your own and have cmake use that, or use a shared lib, or add a cmake find_library for libevent (or libevent2) and add at the end of UNBOUND_LIBRARY.
• 11:06
  moneromooo
  I remember doing a similar thing for libzmq, which can sometimes need... a couple other libs. If detected, I append them.
• 11:06
  » moneromooo looks
• 11:07
  moneromooo
  norm and pgm. So git grep PGM_LIB should get you all you need to replijcate this for libevent.
• 11:08
  moneromooo
  You might end up linking unnecessarily to it, but with static libs it's fine, it just won't pick symbols it doesn't need.
• 11:08
  moneromooo
  Hopefully that's all there is.
• 11:15
  sech1
  release-v0.17 built just fine
• 11:15
  sech1
  interesting
• 11:17
  moneromooo
  I think selsta switch libunbound at some point not too long ago. Might not be on the branch.
• 11:18
  sech1
  there's no UNBOUND_LIBRARY in CMakeLists.txt
• 11:20
  moneromooo
  It's "IES" here.
• 11:20
  moneromooo
  (in common)
• 11:21
  moneromooo
  (which might be wrong, there was at least one problem before with Y/IES mixup)
• 11:21
  sech1
  found in src/common/CMakeLists.txt
• 11:25
  sech1
  ok, it seems to work. Now to `nettle_yarrow256_seed`
• 11:31
  sech1
  okay, I give up: paste.debian.net/hidden/d01f52b1
• 11:32
  sech1
  I added "find_library(EVENT_LIBRARY event)" and "find_library(NETTLE_LIBRARY nettle)", then changed to "${UNBOUND_LIBRARIES} ${EVENT_LIBRARY} ${NETTLE_LIBRARY}" in CMakeLists.txt
• 11:32
  sech1
  it found both, but it still give linker errors
• 11:38
  selsta
  Are you trying to build a static binary?
• 11:40
  selsta
  if yes, try to compile unbound like this: github.com/monero-project/monero/bl…trib/depends/packages/unbound.mk#L9
• 11:41
  selsta
  you also have to compile expat and ldns
• 11:42
  selsta
  like here monero-project/monero-gui #3696/commits/87b151802375c7d7e2ae5432dc36dc0372b2507a
• 11:46
  selsta
  you can also just use `make depends target=x86_64-linux-gnu` (that will compile all the dependencies for you)
• 11:55
  sech1
  Not necessarily static
• 11:55
  sech1
  but `make release` give the same linker errors
• 11:56
  sech1
  I just need any working binary
• 12:13
  selsta
  hmm we build every commit on ubuntu 18.04 and 20.04 without issues, with all dependencies installed from apt
• 12:16
  selsta
  `make depends target=x86_64-linux-gnu` will give you a working binary for now
• 12:17
  selsta
  `sudo apt -y install build-essential libtool cmake autotools-dev automake pkg-config bsdmainutils curl git ca-certificates gperf python3-zmq libdbus-1-dev libharfbuzz-dev`
• 12:18
  selsta
  install these dependencies before running `make depends`
• 12:18
  sech1
  I'll try whatever .github/workflows/build.yml does
• 12:18
  sech1
  Because I followed Readme.MD instructions, maybe they are incomplete
• 12:43
  moneromooo
  Make sure you nuke CMakeCache.txt, cmake can be an ass when you change that kind of stuff.
• 12:44
  moneromooo
  Then link with VERBOSE=1 to check the link line, compare with what it should be (ie, what follows -lunbound or similar).
• 12:46
  sech1
  Huh. I decided to build it on my server and it just built fine. The only difference is that the VM had Ubuntu Desktop 20.04 and this is Ubuntu Server 20.04
• 12:48
  sech1
  Whatever. I can at least spin up the testnet node now. Testnet hardfork in 3 days?
• 12:52
  sech1
  Found it - block 1982800, around 23:00 CEST on May 16th
• 13:50
  selsta
  Ubuntu Desktop and Server should be irrelevant. Maybe you have some old custom compiled unbound version somewhere that gets picked up by cmake?
• 13:55
  sech1
  The second VM I tried was a fresh install
• 13:56
  sech1
  I certainly don't remember compiling Monero or unbound before on that server
• 14:09
  sech1
  hmm, this testnet syncing looks weird: p2pool.io/u/e581ce7a18d3e3b5/weird_sync.png
• 14:42
  selsta
  does it fix itself after some time?
• 14:44
  selsta
  it seems to sync fine, just the target height looks off
• 14:44
  sech1
  It keeps syncing
• 15:32
  gingeropolous
  probably some weird testnet node using a fixed diff that someone forgot about
• 15:33
  gingeropolous
  for the record i got the xmrchain node running testnet on master and mining, so there'll be some help moving the chain forward
• 16:12
  r4v3r23[m]
  will 8046 make it in the HF?
• 16:13
  r4v3r23[m]
  saw it in the merge queue but it didnt get included in the last batch
• 16:13
  selsta
  .merges
• 16:13
  xmr-pr
  8046 8312
• 16:14
  r4v3r23[m]
  thx
• 16:14
  selsta
  moneromooo: 8046 needs a rebase
• 17:06
  moneromooo
  selsta: done
• 19:07
  luigi1112
  are we branching monday?
• 20:10
  sethforprivacy
  Once the tag is up I'll have 2 testnet nodes up and mining on a thread.
• 20:14
  selsta
  we won't have a tag
• 20:14
  selsta
  tag will be done shortly before release
• 20:34
  sethforprivacy
  branch = tag 🙂
• 20:34
  sethforprivacy
  Should have been more specific
• 20:35
  sethforprivacy
  If we really will only branch and not tag then I will update the phrasing in the checklist to reflect that, I thought we did actually tag before testnet launch
• 21:05
  hyc
  selsta is correct. the tag comes late. the branch comes first
• 21:05
  hyc
  and we apply any last minute fixes to the branch, prior to tagging
• 21:06
  hyc
  the point of forking testnet first is to give time to look for bugs on the branch before tagging the official release
• 21:06
  selsta
  yep, tagging before testing is pointless
• 23:15
  sethforprivacy
  Thanks for the clarifications!