You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
docs(conventions): settle the member-order page wording #2765 left open #2772
What #2765 (closing #2490) applied rule c, subgroup d and registry-map order, using my recommendations on five wording questions that protocol-member-order.rst leaves open. #2490 closed before they were ruled on, so they move here with needs: decision.
Several registries in one class: list all readers before all makers (the MH/PCAPNG shape)? Rec: yes; the guard checks only within each registry.
PCAPNG.__option__ order: namespace then value, or by opt_value? Rec: namespace then value, written on the page.
Int keys with no enum: change "alphabetically" to "by key value", which is what TCP/UDP/FRAD now do? Rec: yes. This is the one place where the page and the code currently disagree.
PCAP-NG Option.__enum__: in scope? Rec: out of scope, since Option.register's warning relies on its order.
.rst member lists (e.g. IPv4's) no longer follow source order. Rec: realign them in this issue's docs-only PR.
Reconciled. On 1, 2, 3 and 5 you took the recommendations. On 4, yes: it can be brought in scope.
Several registries in one class: every reader comes before every maker, the MH/PCAPNG shape. The page says so explicitly.
PCAPNG.__option__: ordered by namespace, then by value. That order is written on the page.
Int keys without an enum: "alphabetically" becomes "by key value", matching what TCP, UDP and FRAD already do.
The PCAP-NG Option.__enum__: today its order matters only because Option.register's clash warning names the clashing namespaces in map order. Make that warning list them in a fixed order of its own, and the map can then follow the page like the rest. The guard gains a check for it.
The .rst member lists (IPv4's and the others): realigned to source order.
A worker is starting now. The Option.register change touches code, so this goes as a PR rather than a docs-only push.
Four small choices on #2779, the PR for this issue. The PR already follows my recommendation on each, so a ruling only adjusts it:
Namespace order in PCAPNG.__option__. (a) Keep OptionType's declared order, which is the map as it stands, with opt first. (b) Go alphabetical. Rec: (a).
The Protocol docs page. It documents members that ProtocolBase defines, so I left its order alone. (a) Leave it. (b) Realign it to ProtocolBase's source order. Rec: (a).
A guard for .rst member order. Nothing yet checks that the pages stay in source order. (a) Add one as a follow-up. (b) Don't. Rec: (a).
A reply such as "take all recommendations" here or on #2779 settles it.
Activity
JarryShaw commentedon Nov 4, 2026
JarryShaw commentedon Nov 4, 2026
Reconciled. On 1, 2, 3 and 5 you took the recommendations. On 4, yes: it can be brought in scope.
PCAPNG.__option__: ordered by namespace, then by value. That order is written on the page.Option.__enum__: today its order matters only becauseOption.register's clash warning names the clashing namespaces in map order. Make that warning list them in a fixed order of its own, and the map can then follow the page like the rest. The guard gains a check for it..rstmember lists (IPv4's and the others): realigned to source order.A worker is starting now. The
Option.registerchange touches code, so this goes as a PR rather than a docs-only push.JarryShaw commentedon Nov 4, 2026
what needs my decision?
JarryShaw commentedon Nov 4, 2026
Four small choices on #2779, the PR for this issue. The PR already follows my recommendation on each, so a ruling only adjusts it:
PCAPNG.__option__. (a) KeepOptionType's declared order, which is the map as it stands, withoptfirst. (b) Go alphabetical. Rec: (a).Protocoldocs page. It documents members thatProtocolBasedefines, so I left its order alone. (a) Leave it. (b) Realign it toProtocolBase's source order. Rec: (a)..rstmember order. Nothing yet checks that the pages stay in source order. (a) Add one as a follow-up. (b) Don't. Rec: (a).A reply such as "take all recommendations" here or on #2779 settles it.
JarryShaw commentedon Nov 4, 2026
take all recommendations
JarryShaw commentedon Nov 4, 2026
Reconciled: all four recommendations stand. #2779 needs no change.
PCAPNG.__option__keepsOptionType's namespace order.Protocoldocs page keeps its order..rstmember lists in source order goes in a follow-up issue, filed now. It is test-only, so it can land straight on main once docs(conventions): settle the member-order page wording and align the protocol registry with the other maps (#2772, #2773) #2779 merges.Once its cross-review and CI finish, #2779 closes this issue.
docs(conventions): settle the member-order page wording and align the…
test(project): guard that protocol docs pages list members in class s…