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

super3 / llmjob / 44580646907
100%

Build:
DEFAULT BRANCH: main
Ran 19 Nov 2026 05:47PM UTC
Jobs 1
Files 28
Run time 1min
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

19 Nov 2026 05:45PM UTC coverage: 100.0%. Remained the same
44580646907

push

github

web-flow
Release v0.7.7 (#310)

* Start the working PR after v0.7.6

The founder's rule (18 Nov): a new working PR opens as soon as a PR is merged, and
bug fixes and performance gains land on it as they are verified.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gS41t1UsdRNxyzdFjGjj6

* Let a request keep the node mining while the LLM answers

The founder asked for it on 18 Nov: by default a prompt pauses mining on
its card, but a caller should be able to pick low priority and have
mining carry on, with a slower answer.

The option is `low_priority: true` on /v1/chat/completions and on the
local and LAN gate (port 8000). It is a boolean, not the jobs table's
integer `priority`, because that one orders the server's queue and is
never set by callers. This one changes only what the node does once it
has the job. Anything but the literal true is the default, so mining
pauses as before.

The server stores it on the job and hands it to the node on poll. On the
node, the job worker and the gate report each request to LlmFleet, which
marks the work on a card low priority only when every request we sent
that card is low. A default request that joins low-priority work ends it
at once, so overlapping requests still pause mining. Requests we did not
route (the in-app chat, a program calling llama-server directly) count
as default. promptPause treats low-priority work as idle.

The CLI's co-running mode has no port 8000 gate, so there only cluster
jobs can be low priority. Pricing and rewards are unchanged; nothing
ties them to this.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gS41t1UsdRNxyzdFjGjj6

* Site: explain Kryptex IDs, and one-line curl commands for Hive Shell

Issue #308. Pasting a krx… ID into the HiveOS flight-sheet box got the
generic "doesn't look like a Pearl address" hint, which read as the page
doing nothing. Kryptex accounts stop wor... (continued)

965 of 965 branches covered (100.0%)

Branch coverage included in aggregate %.

1 of 1 new or added line in 1 file covered. (100.0%)

1432 of 1432 relevant lines covered (100.0%)

39.4 hits per line

Jobs
ID Job ID Ran Files Coverage
1 44580646907.1 19 Nov 2026 05:47PM UTC 28
100.0
GitHub Action Run
Source Files on build 44580646907
  • Tree
  • List 28
  • Changed 3
  • Source Changed 3
  • Coverage Changed 3
Coverage ∆ File Lines Relevant Covered Missed Hits/Line Branch Hits Branch Misses
  • Back to Repo
  • Github Actions Build #44580646907
  • a3e17664 on github
  • Prev Build on main (#44449532699)
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