On 01 Nov (20:46:05), nusenu wrote:

David Goulet:

...

On 29 Oct (22:48:53), nusenu wrote:

...

Hi,

I'm wondering if the current version of the text is the latest available version of it or if there is somewhere a newer version that hasn't been pushed yet?

https://gitlab.torproject.org/tpo/core/torspec/-/blob/main/proposals/328-rel...

"Status: Draft" but it is already in released tor versions.

It should actually be set to "Closed" now and we need to merge it in dir-spec.txt.

"Implemented-In" would also be nice.

my understanding of the changelog https://gitlab.torproject.org/tpo/core/tor/-/merge_requests/361/diffs#821ec6...

was that only the "overload-general" line got moved (not all lines from this spec) from the extra-info descriptor to the server descriptor, but this change implies that all lines are now located in the server descriptors?

https://gitlab.torproject.org/tpo/core/torspec/-/commit/3424a245774e2ee56115... https://gitlab.torproject.org/tpo/core/torspec/-/blob/main/proposals/328-rel...

Ah ! my mistake. I'll fix that right away.

You are absolutely right, the other overload lines are in the extra-info only. The general one is in the server descriptor. I confirmed with the implementation and will fix the spec asap!

Good catch!

Has the version field in the "overload-general" line been increase when the semantics for DNS timeouts changed? (the 1 to 1%/10min change)

Yes, in theory but we didn't go for this considering that the version 1 here is absolutely broken and at this early time, we wanted to be agile with this feature and so we backported this as a "fix" to a feature.

Any new features to that very line will see a version bump and a proposal for sure.

The overload-general line implementation had a mis-communication between the proposal and the coding work and so we thought we had the X% over Y% but we didn't in the end.

Cheers! David