-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
R4R: Test that localnet-start makes at least 10 blocks #2018
Conversation
Codecov Report
@@ Coverage Diff @@
## develop #2018 +/- ##
========================================
Coverage 64.86% 64.86%
========================================
Files 115 115
Lines 6862 6862
========================================
Hits 4451 4451
Misses 2127 2127
Partials 284 284 |
26bd636
to
c48bf4e
Compare
Looks like the docker container isn't printing out the |
If dep already sees its scratch directory (.vendor-new), dep ensure fails. This rm -rf's that directory so make get_vendor_deps doesn't fail.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great idea! Notes:
- Can we not rely on
docker pull
? Depends on an external repository (security threat) & often slow. Would prefer a local image with proper layer caching. - Do we need to use Docker at all?
@cwgoes I was trying to reuse the existing Its really just 2 files. If we just run the script in circle, then we could reduce that to just the script. I think the hard part here is figuring out the docker networking. I'll try both approaches. |
afaik, the |
I think using docker is fine here, it's part of a test. If putting the docker file in the sdk is doable, let's do that. I'd prefer to get this working with docker pull, and we make an issue for fixing that dep later. It's not a huge security concern because worst case is that one of our circle ci tests passes when it shouldn't have. (It's running in its own container and persists nothing) |
So after doing a bit of testing here and playing around with circle's cool
This is kinda weird because the step before builds the linux binary. When I look where its supposed to be ( Running it in the Now I'm running into this fun issue which looks like a
Also this is going to need a rebase. Would love some feedback on the approach. |
related: tendermint/tendermint#1532 |
I need an extra set of 👀 on this from @greg-szabo |
Alright! ht to @greg-szabo for helping me find that |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Agreed with the concept, still a few concerns about implementation, see comments.
ifndef GCC | ||
$(error "gcc not installed for ledger support, please install or set LEDGER_ENABLED to false in the Makefile") | ||
endif | ||
ifeq ($(UNAME_S),OpenBSD) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this change intended to be in this PR?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
NO :*(
|
||
BUG FIXES | ||
* \#1988 Make us compile on OpenBSD (disable ledger) [#1988] (https://github.com/cosmos/cosmos-sdk/issues/1988) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this change intended to be in this PR?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
NO :(
@@ -38,6 +38,7 @@ BREAKING CHANGES | |||
* [types] sdk.NewCoin now takes sdk.Int, sdk.NewInt64Coin takes int64 | |||
* [cli] #1551: Officially removed `--name` from CLI commands | |||
* [cli] Genesis/key creation (`init`) now supports user-provided key passwords | |||
* [cli] unsafe_reset_all, show_validator, and show_node_id have been renamed to unsafe-reset-all, show-validator, and show-node-id |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this change intended to be in this PR?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No :( this is going to need another PR.
@@ -73,8 +74,10 @@ IMPROVEMENTS | |||
* [x/stake] \#1815 Sped up the processing of `EditValidator` txs. | |||
* [server] \#1930 Transactions indexer indexes all tags by default. | |||
* [x/stake] \#2000 Added tests for new staking endpoints | |||
* [tools] Make get_vendor_deps deletes `.vendor-new` directories, in case scratch files are present. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this change intended to be in this PR?
make get_vendor_deps | ||
make build-linux | ||
make localnet-start | ||
docker run -it --network=host jackzampolin/gaiad-poll:latest ./poll.sh 40 5 10 localhost |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I still am not a fan of this. You're right that it's just in CI, which we don't give a trusted role (presently), but it's hard to debug - now if something breaks where do I go to figure out what's in your Docker image?
Tried going to https://hub.docker.com/r/jackzampolin/gaiad-poll/ but that doesn't appear to have a Dockerfile
Can we just run the poll script natively in CI?
Closing this in favor of #2057 |
With Make 4.3+, the empty whitespace does not seem to work as originally intended. This causes build tags to be "netgo ledger," on Ubuntu 22.04 and other systems that include the newer Make version. The build tags were intended as "netgo,ledger" which can be observed on Make 4.2 (shipped with Ubuntu 20.04). This change swaps out the `+=` use in favor of an explicit `:=`. Ref: https://www.gnu.org/software/make/manual/html_node/Appending.html Closes cosmos/gaia#2017. Co-authored-by: Marius Poke <marius.poke@posteo.de> Co-authored-by: Milan Mulji <98309852+mmulji-ic@users.noreply.github.com>
This PR adds the check to the Circle CI script to ensure that the
make localnet-start
command starts a testnet that can produce at least 10 blocks. cc @ValarDragon