On Sat, Mar 14, 2020 at 12:07:39PM +0100, Tim Düsterhus wrote: > Willy, > > Am 14.03.20 um 10:13 schrieb Willy Tarreau: > >> This is about the "Table Definition Message", more specifically the > >> "Encoded Table Type". > >> > >> docs/peers-v2.0.txt says: > > > > There's also doc/peers.txt which is more recent and more complete. I don't > > know if there are still some parts of peers-v2.0 which were not covered by > > peers.txt. > > Neither of them are sufficient to understand the protocol. Not even both > of them taken together. Some paragraphs are also pretty unclear. All in > all I had to read the code to understand what's happening. > > The peers.txt is more correct on what it says, but it's missing > essential information (e.g. what incremental updates are). > peers-v2.0.txt contains more information, but also more mistakes > (especially regarding the "magic numbers" such as the table types).
I think peers-v2.0 was older but committed later to complement the missing info from the newer one. > > From now, the best solution likely is to check where the table type is > > used and instead go back to a table-specific type with hard-coded values > > matching what we have now. > > > > Should I file an issue for tracking that? Yes, feel free to do so, this will definitely help get it eventually done. Thanks, Willy