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
By switching to a new model, we could ask people to register new entries in csv/yaml files so we can have tags; then everything markdown is managed by the engine, not by humans.
If we do this, I expect that links could be linked to "tags/categories" with things like "staff picks", but also to "questions", so we can create an automated FAQ.
My approach is to generate a awesome-lint compatible README from a markdown template.
This allows a HTML template to include the rich descriptions stats and media, and be hosted on a GitHub Page or externally.
The awesome-gno content is sourced from a YAML file converted from the previous markdown README.
@poroburu I'd ideally prefer working directly on the markdown rather than editing a yaml. There can be syntaxes understood by the linter and marshalers, but IMO the source of truth should be the markdown.
Counter-arguments welcome. My preference is that 1. i don't particularly like modifying very large yaml files over time and 2. we should be able to add more information and adjust the README dynamically, rather than having to modify a generator in order to change the markdown. -> README is source of truth, rest is derivation
Some basic repository metadata can be extracted in the traditional awesome list markdown entries.
Imagining GitHub repository API stats like: stars, commit graph, contributors.
The yaml entries can contain richer metadata, to be rendered on the HTML.
Possibly like this buildwithdotnet example. As this example could not be generated from an awesome-lint markdown readme: https://builtwithdot.net/
Although, if gno-awesome will follow the awesome manifesto and list guidelines. The markdown cannot be generated.
Prepared by an AI agent (Claude) on behalf of @moul, 2026-11-05.
This needs a decision before anyone writes code, so I've laid out the options here instead of opening a PR. The question this thread hasn't settled: what is the source of truth?@thehowl wants the README, while #58 (draft, conflicting) uses YAML.
A. The README stays the source of truth, and we generate a static page from it. A GitHub Action parses README.md (it's already lint-clean, one - [name](url) - desc. per line) and builds a single HTML page on GitHub Pages, with search, a section filter, and stars/last-push pulled from the GitHub API at build time. Contributors' workflow stays the same. Cost: about one afternoon. Tags would have to be inline markers (e.g. <!-- tags: wallet -->).
B. YAML becomes the source of truth, and the README is generated (#58). This gives the richest metadata (tags, staff picks, FAQ links, media). The downsides: contributors edit YAML instead of markdown, CI has to reject hand edits to the README, and it reverses @thehowl's stated preference. #58 would need a rebase plus the HTML half finished.
C. Render it on gno.land. A realm such as r/gnoland/awesome gets synced from the README by CI. It's on-brand and dogfoods gnoweb, but it adds a deploy key and a chain dependency to a markdown list.
D. Leave it as is. With roughly 70 entries and a ToC, the README still browses fine on GitHub.
Recommendation: A. It's the cheapest option, it keeps the README as the source of truth, and it can be extended to C later by reusing the same parser. If the maintainers agree, it's a small, contained PR. If we go with B, #58 is the starting point.
Activity
moul commentedon Feb 9, 2023
By switching to a new model, we could ask people to register new entries in csv/yaml files so we can have tags; then everything markdown is managed by the engine, not by humans.
If we do this, I expect that links could be linked to "tags/categories" with things like "staff picks", but also to "questions", so we can create an automated FAQ.
solofomo commentedon Feb 22, 2023
this tool can be used to build the awesome page
https://github.com/anuraghazra/github-readme-stats
for the tags/categories part, need to do more research, working on it
poroburu commentedon Sep 23, 2024
My approach is to generate a
awesome-lintcompatible README from a markdown template.This allows a HTML template to include the rich descriptions stats and media, and be hosted on a GitHub Page or externally.
The
awesome-gnocontent is sourced from a YAML file converted from the previous markdown README.thehowl commentedon Sep 23, 2024
@poroburu I'd ideally prefer working directly on the markdown rather than editing a yaml. There can be syntaxes understood by the linter and marshalers, but IMO the source of truth should be the markdown.
Counter-arguments welcome. My preference is that 1. i don't particularly like modifying very large yaml files over time and 2. we should be able to add more information and adjust the README dynamically, rather than having to modify a generator in order to change the markdown. -> README is source of truth, rest is derivation
poroburu commentedon Sep 25, 2024
@thehowl
Some basic repository metadata can be extracted in the traditional awesome list markdown entries.
Imagining GitHub repository API stats like: stars, commit graph, contributors.
The yaml entries can contain richer metadata, to be rendered on the HTML.
Possibly like this
buildwithdotnetexample. As this example could not be generated from anawesome-lintmarkdown readme: https://builtwithdot.net/Although, if
gno-awesomewill follow the awesome manifesto and list guidelines. The markdown cannot be generated.moul commentedon Nov 5, 2026
Prepared by an AI agent (Claude) on behalf of @moul, 2026-11-05.
This needs a decision before anyone writes code, so I've laid out the options here instead of opening a PR. The question this thread hasn't settled: what is the source of truth? @thehowl wants the README, while #58 (draft, conflicting) uses YAML.
A. The README stays the source of truth, and we generate a static page from it. A GitHub Action parses README.md (it's already lint-clean, one
- [name](url) - desc.per line) and builds a single HTML page on GitHub Pages, with search, a section filter, and stars/last-push pulled from the GitHub API at build time. Contributors' workflow stays the same. Cost: about one afternoon. Tags would have to be inline markers (e.g.<!-- tags: wallet -->).B. YAML becomes the source of truth, and the README is generated (#58). This gives the richest metadata (tags, staff picks, FAQ links, media). The downsides: contributors edit YAML instead of markdown, CI has to reject hand edits to the README, and it reverses @thehowl's stated preference. #58 would need a rebase plus the HTML half finished.
C. Render it on gno.land. A realm such as
r/gnoland/awesomegets synced from the README by CI. It's on-brand and dogfoods gnoweb, but it adds a deploy key and a chain dependency to a markdown list.D. Leave it as is. With roughly 70 entries and a ToC, the README still browses fine on GitHub.
Recommendation: A. It's the cheapest option, it keeps the README as the source of truth, and it can be extended to C later by reusing the same parser. If the maintainers agree, it's a small, contained PR. If we go with B, #58 is the starting point.