Skip to content

docs: document Gall's five undocumented vane scries - #276

Open
sigilante wants to merge 1 commit into
urbit:masterfrom
sigilante:docs/pr6c-gall-scries
Open

docs: document Gall's five undocumented vane scries#276
sigilante wants to merge 1 commit into
urbit:masterfrom
sigilante:docs/pr6c-gall-scries

Conversation

@sigilante

Copy link
Copy Markdown
Contributor

Eleventh PR from the audit against urbit/urbit@08026c84b2. The Gall vane scry arm (sys/vane/gall.hoon:2897-3160) dispatches 15 cares; gall/scry.md documented 10.

Companion PRs: #266, #267, #268, #269, #270, #271, #272, #273, #274, #275.

The five missing cares

Care Meaning Produces Source
%b blocked moves (map term (qeu blocked-move)) gall.hoon:2936
%g flubbed apps (jug ship app=term) gall.hoon:2984
%h halted apps (jug app=term [ship =duct]) gall.hoon:2993
%i flub ducts (map ship duct) gall.hoon:3001
%y live agents (map term yoke) gall.hoon:3138

Types and descriptions come from the +$ state definition (gall.hoon:78-90) and its own field comments. %y filters the yoke map to live agents only.

%g, %h and %i are the observability surface for the Ames/Gall backpressure protocol documented in #269: %g is which agents have refused a $plea, %h is which flows are halted because an agent is missing or suspended, and %i is the duct used to send %flub notifications per ship.

Verification note — please read

Unlike the other PRs in this series, these five were not verified by live scry. Every attempt to type a vane-scry path wedged the dojo's line editor: the $ path element did not survive send-keys or paste-buffer, and repeated attempts left the input line unusable until the ship was restarted.

Rather than keep fighting the terminal — or, worse, write a plausible-looking example product I had not actually observed — the entries are documented from the dispatch conditions and return marks in gall.hoon, which are unambiguous, and the example blocks show only the invocation, not a fabricated output. Every other PR in this series contains real captured output; these deliberately do not, and I would rather say so than have a reviewer assume otherwise.

%y is exercised indirectly and does work: the +vats generator scries it (.^(yokes ... %gy ...)) on every run, and +vats was verified live in #268.

Avoided a cross-PR link

An earlier draft pointed %g at [%halt](../ames/tasks.md#halt), but that anchor is added by #269 and does not exist on master. If these merged in the wrong order the link would break, so the reference is prose instead.

Test-merged against the ten open companion PRs — all clean.

🤖 Generated with Claude Code

Against urbit/docs.urbit.org@08026c84b2. The Gall vane scry arm (sys/vane/gall.hoon:2897-3160)
dispatches 15 cares; gall/scry.md documented 10. Adding the missing five, in
alphabetical position:

  %b  blocked moves        (map term (qeu blocked-move))   gall.hoon:2936
  %g  flubbed apps         (jug ship app=term)             gall.hoon:2984
  %h  halted apps          (jug app=term [ship =duct])     gall.hoon:2993
  %i  flub ducts           (map ship duct)                 gall.hoon:3001
  %y  live agents          (map term yoke)                 gall.hoon:3138

Types and descriptions are taken from the `+$ state` definition
(gall.hoon:78-90) and its own field comments. %y filters the yoke map to live
agents only.

%g, %h and %i are the observability surface for the Ames/Gall backpressure
protocol: %g is which agents have refused a $plea, %h is which flows are halted
because an agent is missing or suspended, and %i is the duct used to send %flub
notifications per ship.

VERIFICATION NOTE, stated plainly: unlike the other PRs in this series, these
five were NOT verified by live scry. Each attempt to type a vane-scry path ended
up wedging the dojo's line editor -- the `$` path element did not survive
send-keys or paste-buffer, and repeated attempts left the input line unusable
until the ship was restarted. Rather than keep fighting the terminal or, worse,
write a plausible-looking example I had not actually run, the entries are
documented from the dispatch conditions and return marks in gall.hoon, which are
unambiguous, and the example blocks show only the invocation, not a fabricated
product.

%y is exercised indirectly and does work: the +vats generator scries it
(`.^(yokes ... %gy ...)`) on every run, and +vats was verified live in an
earlier PR in this series.

Deliberately avoided a cross-PR link: an earlier draft pointed %g at
[%halt](../ames/tasks.md#halt), but that anchor is added by the Ames
backpressure PR and does not exist on master. If these merged in the wrong order
the link would break, so the reference is prose instead.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant