• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

hyperledger / fabric-x-committer / 35712748875
91%

Build:
DEFAULT BRANCH: main
Ran 22 Sep 2026 09:57AM UTC
Jobs 0
Files 0
Run time –
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

pending completion
35712748875

push

github

web-flow
[loadgen] Stop block preparation from capping the generator at small block sizes (#825)

#### Type of change

- Improvement (improvement to code, performance, etc)
- Test update

#### Description

- `workload.MapToOrdererBlock` now computes the block's data hash. The
hash covers the block's own
data, so unlike the number and the previous hash it does not depend on
the chain and need not be
computed in chain order — and it is the whole per-transaction cost of
preparing a block, which a
single block-preparing goroutine would otherwise pay for every block in
sequence.
- `prepare-in-place` on the mock orderer declares that blocks submitted
to it are its own, so it
writes their header into the block it was given and reuses the data hash
the submitter filled in
rather than deep-cloning and rehashing. Applied only after the genesis
block, which may be the
package-level default and would otherwise be rewritten for every orderer
in the process.
- The sidecar adapter turns it on, since it builds a block per batch and
never touches it again. What
  the committer receives is unchanged; only the cost of producing it is.
- Adds `BenchmarkOrdererBlockRate`, which measures the rate the adapter
is bounded by, at both block
sizes the evaluation used. It pins the buffer between submitter and
preparer to two blocks, because
`SubmitBlock` only hands the block over and a deep buffer measures the
channel write instead of the
work behind it. Its floor excludes the consenter signature, roughly 45
µs, which a deployment pays
  and this configuration does not exercise.

#### Additional details (Optional)

Preparation becomes independent of block size, which is the point — the
clone and the hash were the
only costs that scaled with it:

go test ./mock/ -run '^$' -bench '^BenchmarkOrdererBlockRate$'
-benchtime=30x

#### Related issues

- resolves #824
Source Files on build 35712748875
Detailed source file information is not available for this build.
  • Back to Repo
  • 26a693f6 on github
  • Prev Build on main (#35705135807)
  • Next Build on main (#35721614387)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc