Gilles Filippini authored 1 year ago and pini committed 1 year ago

Use image paritytech/ci-linux:production for 'labels' and 'quality' stages.
We don't need to force a specific Debian release as this image is not used
for the 'build' stage anymore.

Gilles Filippini authored 1 year ago and pini committed 1 year ago

The build runs in two stages:
1. Cross compilation for target architecture on native arch image
2. Installation into target architecture image

Currently it requires a third stage to workaround an issue in podman/buildah:
https://github.com/containers/buildah/issues/4742

Configuring a gitlab-runner with podman:
1. Follow this documentation:
   https://docs.gitlab.com/runner/executors/docker.html#use-podman-to-run-docker-commands
2. Install the `crun` and `qemu-user-static` packages

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* fix cargo files

* fix slashing strategy

* refactore remove_member_from_blacklist

* add missing test

* add license

* refac tests

adds comments
remove some repetitions
reorder tests (offence test are below)
rename tests

* feat(pallet_grandpa) add offences handler

* feat(pallet_authority_members) add offences execution

* feat(pallet_offences) add pallet offences

* fix gtest runtime type name

* add embeded client spec and genesis in binary

(this re-buids chainspecs at each run)

* refac chainspecs generation

* add what is needed to build gtest chainspecs

* fix smith certification validity

* get values of parameters

* tuxmain review

* add checks and improve formatting

* improve genesis parsing

adds info

* fix json with new format

* fix warnings

* new gtest genesis format

* get build working with gtest feature

`cargo build --features gtest --no-default-features`

* update lib.rs

* update cargo.toml

* add readme for runtimes

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* fix(runtime) fix benchmarks for g1 and gtest

* feat(runtime) enforce gdev-benchmark for benchmark commands

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* Add extrinsic_weights benchmark

* Fix block_weights benchmark

* Revert "run benchmarks"

This reverts commit 2c775131.

* run benchmarks

commit only section which changed

* fix build benchmark

Hugo Trentesaux authored 1 year ago and Pascal Engélibert committed 1 year ago

* review tuxmain

* fix rebase

* rename force_add_cert to do_add_cert_checked

* remove origin from force_add_cert

* remove duplicate UD handling

* claim membership - validating identity

ensure claim membership does not prevent validating identity

* clippy

* add acquired membership handler

and new integration test scenario

* run benchmarks

* refactor claim_membership

and expire membership instead of revoking
when too few certs

* re-enable claim_membership

* membership expiry no identity removal

membership expiry no more triggers identity removal

* control by test

* at membership expiry : back to pending membership

* reveal identity removal on membership expiry

* refac test

* refac membership renewal

add expiry unscheduling
do not verify expiration block expiring memberships
add and refac tests
add integration test

* wip

Hugo Trentesaux authored 1 year ago and Pascal Engélibert committed 1 year ago

* fix certification period on renewal

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* feat: add pallet session benchmark

* fix(pallet_collective): fix parameters for benchmark

* fix(pallet_treasury): fix parameters for benchmark

* fix(pallet_im_online): fix parameters for benchmark

Fixed the pallet_im_online parameters to be pass benchmarks. MaxKeys needs to be at least 1000 because of https://github.com/paritytech/substrate/blob/e94cb0dafd4f30ff29512c1c00ec513ada7d2b5d/frame/im-online/src/benchmarking.rs#L35

* feat(runtimes): use our benchmarks for substrate pallets

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* fix after rebase

* feat(runtimes): use our benchmarks for pallet authority_members

* feat(pallet_authority_members): add weights info

* feat: add pallet authority_members benchmark

paritytech/ci-linux:production was recently switched from buster to bullseye,

Gilles Filippini authored 1 year ago and Hugo Trentesaux committed 1 year ago

It makes more sense to set the release LTO profile into the related
Cargo.toml section than into the Dockerfile only.

Using LTO 'thin' profile because according to the documentation it is
a good compromise with no or little performance loss.

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* fix after rebase

* add comment warning about hardcoded smith metadata

* feat(runtimes): use our benchmarks for pallet membership

* feat(pallet_membership): add weights info

* feat: add pallet membership benchmark

Benjamin Gallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

* feat(runtimes): use our benchmarks for pallet oneshot_account

* feat(pallet_oneshot_account): add weights info

* fix(pallet_oneshot_account): remove unecessary where clause

* fix not enough for transfer

* cucumber: add identity validation

* wip add revocation test

* fix same bug in cucumber test

* fix(#109): handle certification renewal

* test(#109): reveal certification renewal issue

duplicate functions
renamed smiths -> smith

bgallois authored 1 year ago and Hugo Trentesaux committed 1 year ago

Added a genesis builder for benchmarking with pre-configured accounts that can be used during the benchmarking and circumvent limitation imposed by pallets coupling.

* clippy

* review tuxmain

* refac: simplify syntax

* ref: reorder keys to limit confusion

* fix: session keys instead of authority keys

* refac: move stuff and add comments