site: eight families, real RAM figures, and a contract - #1302
Merged
Conversation
Il sito diceva "Currently shipping v1.7.0" con sei famiglie: quattro release e due modelli indietro. Mancavano GLM-5.3-Flash e Qwen3.8-Flash-Next, cioe' i due piu' nuovi e gli unici con vision. Le card ora rispondono alle due domande di chi arriva: cosa sa fare (badge reasoning/tools/vision/audio, presi dal registry e dal server, non inventati) e con quanta RAM gira (i numeri della tabella del README, non nuovi). "Size - Status" diceva "Live" su tutto, che non distingue niente; "Size - RAM" e' la tesi del progetto in una riga: 744B, da 16 GB. DeepSeek V4 si chiamava "DeepSeek V4 Flash" ovunque tranne che sul sito; ora anche li'. Il contratto: site/index.html era la sesta copia del numero di versione e la quinta lista di famiglie, e nessuno dei contratti esistenti (#1287, #1288) la guardava. Il nuovo test confronta la riga "Currently shipping" con version.py e pretende ogni display_name del registry. Controlli negativi: versione vecchia fallisce; famiglia rimossa da TUTTO il file fallisce (rimuoverla solo dall'h3 non basta, il footer la nomina ancora, ed e' giusto cosi': il contratto e' che il sito la nomini, non dove).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The site said "Currently shipping v1.7.0" and listed six families: four releases and two models behind. GLM-5.3-Flash and Qwen3.8-Flash-Next — the two newest, the only ones with vision — were absent.
What the cards say now
Each card answers the two questions a visitor actually has:
glm53_image.pyandqwen38_image.pyare wired into the server; Inkling gets audio; OLMoE gets none, honestly.Order: GLM-5.2, then the two new ones, then the rest as before. DeepSeek V4 is now "DeepSeek V4 Flash", its registry name, on the site too.
The contract
site/index.htmlwas the sixth copy of the version number and the fifth list of families, and neither existing contract (#1287 READMEs, #1288 banners) looked at it. The site is the first thing a visitor sees and the last thing anyone remembers to update: exactly the place for a contract.The new test pins the "Currently shipping" line to
version.pyand requires every registrydisplay_name. Negative controls: stale version fails; a family stripped from the whole file fails. (Stripping it from just the heading does not fail, because the footer still names it — which is correct: the contract is that the site names the family, not where.)Note on going live
Pages serves
main, so this reaches the live site at the next dev→main merge. Since the site is four versions stale, I would follow this with a small docs-only dev→main PR (no tag, so no release build) rather than waiting for v1.11.0.