Flakes don't appear to be well-supported

[cole-h]

14:43:05 postgres.1           | 2021-11-09 14:43:05.147 PST [169780] ERROR:  new row for relation "jobsets" violates check constraint "jobsets_flake_paths_check"
14:43:05 postgres.1           | 2021-11-09 14:43:05.147 PST [169780] DETAIL:  Failing row contains (pr-2, 3, test, Test pr by cole-h: https://github.com/cole-h/hydra-playground/pu..., null, null, null, null, null, null, 1, 0, 0, , 3, 300, 1, null, null, null, 0, git+ssh://[email protected]:cole-h/hydra-playground.git?ref=test-pr...).
14:43:05 postgres.1           | 2021-11-09 14:43:05.147 PST [169780] STATEMENT:  INSERT INTO jobsets ( checkinterval, description, emailoverride, enabled, enableemail, flake, hidden, keepnr, name, project, schedulingshares) VALUES ( $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11 ) RETURNING id
14:43:05 hydra-notify.1       | ERROR: failed to process declarative jobset test:pr-2, DBIx::Class::Storage::DBI::_dbh_execute(): DBI Exception: DBD::Pg::st execute failed: ERROR:  new row for relation "jobsets" violates check constraint "jobsets_flake_paths_check"
14:43:05 hydra-notify.1       | DETAIL:  Failing row contains (pr-2, 3, test, Test pr by cole-h: https://github.com/cole-h/hydra-playground/pu..., null, null, null, null, null, null, 1, 0, 0, , 3, 300, 1, null, null, null, 0, git+ssh://[email protected]:cole-h/hydra-playground.git?ref=test-pr...). [for Statement "INSERT INTO jobsets ( checkinterval, description, emailoverride, enabled, enableemail, flake, hidden, keepnr, name, project, schedulingshares) VALUES ( ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ? ) RETURNING id" with ParamValues: 1='300', 2='Test pr by cole-h: https://github.com/cole-h/hydra-playground/pull/2', 3='', 4='1', 5='0', 6='git+ssh://[email protected]:cole-h/hydra-playground.git?ref=test-pr&rev=c3f4581eeab7525f91d204723d9261ba2e7edfe8', 7='0', 8='3', 9='pr-2', 10='test', 11='1'] at /home/vin/workspace/vcs/hydra/src/lib/Hydra/Helper/AddBuilds.pm line 65
14:43:05 hydra-notify.1       | 
14:43:05 postgres.1           | 2021-11-09 14:43:05.150 PST [169780] ERROR:  current transaction is aborted, commands ignored until end of transaction block
14:43:05 postgres.1           | 2021-11-09 14:43:05.150 PST [169780] STATEMENT:  SELECT me.name, me.id, me.project, me.description, me.nixexprinput, me.nixexprpath, me.errormsg, me.errortime, me.lastcheckedtime, me.triggertime, me.enabled, me.enableemail, me.hidden, me.emailoverride, me.keepnr, me.checkinterval, me.schedulingshares, me.fetcherrormsg, me.forceeval, me.starttime, me.type, me.flake FROM jobsets me WHERE ( ( me.name = $1 AND me.project = $2 ) )
14:43:05 hydra-notify.1       | ERROR: failed to process declarative jobset test:pr-4, DBIx::Class::Storage::DBI::_dbh_execute(): DBI Exception: DBD::Pg::st execute failed: ERROR:  current transaction is aborted, commands ignored until end of transaction block [for Statement "SELECT me.name, me.id, me.project, me.description, me.nixexprinput, me.nixexprpath, me.errormsg, me.errortime, me.lastcheckedtime, me.triggertime, me.enabled, me.enableemail, me.hidden, me.emailoverride, me.keepnr, me.checkinterval, me.schedulingshares, me.fetcherrormsg, me.forceeval, me.starttime, me.type, me.flake FROM jobsets me WHERE ( ( me.name = ? AND me.project = ? ) )" with ParamValues: 1='pr-4', 2='test'] at /home/vin/workspace/vcs/hydra/src/lib/Hydra/Helper/AddBuilds.pm line 65
14:43:05 hydra-notify.1       | 
14:43:05 postgres.1           | 2021-11-09 14:43:05.153 PST [169780] ERROR:  current transaction is aborted, commands ignored until end of transaction block
14:43:05 postgres.1           | 2021-11-09 14:43:05.153 PST [169780] STATEMENT:  SELECT me.name, me.id, me.project, me.description, me.nixexprinput, me.nixexprpath, me.errormsg, me.errortime, me.lastcheckedtime, me.triggertime, me.enabled, me.enableemail, me.hidden, me.emailoverride, me.keepnr, me.checkinterval, me.schedulingshares, me.fetcherrormsg, me.forceeval, me.starttime, me.type, me.flake FROM jobsets me WHERE ( ( me.name = $1 AND me.project = $2 ) )
14:43:05 hydra-notify.1       | ERROR: failed to process declarative jobset test:pr-3, DBIx::Class::Storage::DBI::_dbh_execute(): DBI Exception: DBD::Pg::st execute failed: ERROR:  current transaction is aborted, commands ignored until end of transaction block [for Statement "SELECT me.name, me.id, me.project, me.description, me.nixexprinput, me.nixexprpath, me.errormsg, me.errortime, me.lastcheckedtime, me.triggertime, me.enabled, me.enableemail, me.hidden, me.emailoverride, me.keepnr, me.checkinterval, me.schedulingshares, me.fetcherrormsg, me.forceeval, me.starttime, me.type, me.flake FROM jobsets me WHERE ( ( me.name = ? AND me.project = ? ) )" with ParamValues: 1='pr-3', 2='test'] at /home/vin/workspace/vcs/hydra/src/lib/Hydra/Helper/AddBuilds.pm line 65
14:43:05 hydra-notify.1       | 

See my naive attempt here: https://github.com/cole-h/hydra-playground/tree/flake

Maybe I was just holding it wrong, though. (Entirely possible, I kinda just cobbled together the configuration and flake without any references.)

At the very least, flake support isn't well-documented.

[cole-h]

After #17 merges, the only "issue" will be that the generator assumes SSH access to the flake.

[grahamc]

The issue at this point is we're generating invalid flakerefs:

We're generating: git+ssh://[email protected]:NixOS/nix?ref=v1.2.3

Nix wants: git+ssh://[email protected]/NixOS/nix?ref=v1.2.3.