dalten-collective / keep Goto Github PK

handle details around activating/deactivating before/after wrapping.

after performing a local backup, active agents are wiped.

seems the response is bad/has empties:

{
  "state": {
    "pending": [],
    "auto": [],
    "live": false,
    "saved": []
  },
  "diff": {
    "ship": "zod",
    "time": 1662066589
  },
  "type": "saved"
}

there are two kinds of pendings.

• pending backup (this is the one we want to show on "outstanding invites" as it currently exists
• pending restore (we need a new UI for this)

currently they're just all under "outstanding invites"

Put "waiting for restore..." directly under the ship in Live backup targets

Unclear if HTML5 history mode will work out of the box on a globbed distribution served from urbit.
Test this after globbing and installing locally. Go back to createWebHashHistory() in router.ts otherwise.

Note to self @vcavallo change the "backup complete!" message to show on successful backup. change to "backup requested" or something.

if backup requester has c gora from
backup provider, consider them to be whitelisted.

In my current setup, Gora is wrapped and live.
I've then installed Pick, wrapped it, committed with no errors. When I +dbug keep, it appears in the live list:

> :keep +dbug
>   [%0 kept={} live={%gora %pick}]

But when +dbug'ing Pick, the state does not indicate that it is live (no Keep state there)

After poking Pick to make it live (:pick &keep [%live %.y]), the keep state is identical:

> :keep +dbug
>   [%0 kept={} live={%gora %pick}]

But the Pick state then does show it is Keep-live:

[...]
  ]
  %0
  live=%.y
  last={}
  auto={}
  pending={}
]

Update: The same does not happen after a |nuke then |revive. The agent starts off as live: false in this case.

Diff

{
  "backups": [
    {
      "ship": "zod",
      "agent": "simple",
      "time": 1662036983
    }
  ],
  "agents": [
    "simple"
  ]
}

planets must be whitelisted before you'll take backups from them.

We should provide either documentation or include in the desk a directory of the files a wrapped agent will need to include.

Partial list accreting here:

• mar/keep.hoon
• mar/js.hoon, mar/json.hoon: #13
• lib/keep.hoon
• lib/sane.hoon
• sur/keep.hoon
• add keep to /+ imports
• wrap with %- agent:keep

• Merge Peat & Keep Desks
• Identify Provider Moon Name
  • Proposed: ~savler-baclyn-dozzod-dalten
  • Proposed: ~magped-magped-rabsef-bicrym
• Glob Front End
• Prepare Docket & Garden Tile Image

ake: failed
  bar-stack
~[~[//http-server/0vi.vrbqo/176/1] ~[/dill //term/1] ~[//behn]]
wire=/channel/subscription/1662753836-9920dc/2/~zod/simple
[%poke %wake]
%simple: on-fail with term %arvo-response
%simple: on-fail with term %arvo-response
error in %simple
%arvo-response
! /lib/default-agent/hoon:<[63 3].[63 5]>
!
  "unexpected system response %behn to %simple on wire /keep/behn/ship/~zod"
! /lib/default-agent/hoon:<[62 3].[63 5]>
! /lib/keep/hoon:<[284 35].[284 63]>
! /lib/keep/hoon:<[284 9].[284 64]>
! /lib/keep/hoon:<[283 7].[291 9]>
! /lib/keep/hoon:<[282 7].[291 9]>
! /lib/keep/hoon:<[274 5].[292 7]>
! /lib/keep/hoon:<[270 5].[292 7]>
! /lib/keep/hoon:<[269 5].[292 7]>
! /sys/vane/gall/hoon:<[1.392 9].[1.392 47]>

Got this when playing around with recurring backups both to local disk and to own ship. maybe non-unique behn wires?

See this comment.

The saved state has slightly richer data than the UI is using so far.

• remove 'simple' occurrences from desk.bill

currently i have a moon in 'pending' that is also already taking recurring backups.

this feature also might make "remove from pending" more desirable

This is issue that the github project board creates upon linking. ignore.