Astrowind-based site design #37

caesar commented 2022年12月06日 13:21:59 +01:00

Preview at https://forgejo.codeberg.page/@wip-astrowind/

---

I've based this off the initial Astrowind branch @dachary created, but I've fixed a lot of things and merged changes from main, and began work on customizing the theme a bit.

There's a lot still to be done, but as promised in Matrix I'm pushing my progress so people can see how it's going and contribute if they feel like it.

Assuming there's no other news I'll continue with this tomorrow.

Preview at https://forgejo.codeberg.page/@wip-astrowind/ --- I've based this off the initial Astrowind branch @dachary created, but I've fixed a lot of things and merged changes from main, and began work on customizing the theme a bit. There's a lot still to be done, but as promised in Matrix I'm pushing my progress so people can see how it's going and contribute if they feel like it. Assuming there's no other news I'll continue with this tomorrow.

caesar commented 2022年12月06日 13:23:23 +01:00

CI is failing, not sure why, seems to be a problem installing a binary dependency (the sharp node module). I'll take a look at that tomorrow too but if anyone with more knowledge of the CI can fix it before then that would be great.

CI is failing, not sure why, seems to be a problem installing a binary dependency (the `sharp` node module). I'll take a look at that tomorrow too but if anyone with more knowledge of the CI can fix it before then that would be great.

Ghost commented 2022年12月06日 13:55:51 +01:00

Looks like this is more an installation problem than a CI problem. Does npm install work for you locally?

npm ERR! code 1
npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp
npm ERR! command failed
npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy)
npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br
npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies
npm ERR! sharp: Installation error: aborted
npm ERR! A complete log of this run can be found in:
npm ERR! /root/.npm/_logs/2022-12-06T12_18_39_470Z-debug-0.log

Looks like this is more an installation problem than a CI problem. Does `npm install` work for you locally? ``` npm ERR! code 1 npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp npm ERR! command failed npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy) npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies npm ERR! sharp: Installation error: aborted npm ERR! A complete log of this run can be found in: npm ERR! /root/.npm/_logs/2022-12-06T12_18_39_470Z-debug-0.log ```

Ghost commented 2022年12月06日 14:01:49 +01:00

That's weird, it works when I do this. What's the difference I wonder.

$ docker run --rm -ti node:19 bash
Unable to find image 'node:19' locally
19: Pulling from library/node
f2f58072e9ed: Pull complete 
5c8cfbf51e6e: Pull complete 
aa3a609d1579: Pull complete 
094e7d9bb04e: Pull complete 
2cbfd734f382: Pull complete 
65f9380179fd: Pull complete 
55fb38aec835: Pull complete 
5a5cfed58206: Pull complete 
8a4707ab05af: Pull complete 
Digest: sha256:3341143bf90c80e5b08694c706d49c04eea535f3d3cc0851f1698662f5713518
Status: Downloaded newer image for node:19
root@1a718f95e9f9:/# git clone -b wip-astrowind https://codeberg.org/forgejo/website
Cloning into 'website'...
remote: Enumerating objects: 619, done.
remote: Counting objects: 100% (619/619), done.
remote: Compressing objects: 100% (575/575), done.
remote: Total 739 (delta 350), reused 111 (delta 30), pack-reused 120
Receiving objects: 100% (739/739), 3.43 MiB | 3.53 MiB/s, done.
Resolving deltas: 100% (386/386), done.
root@1a718f95e9f9:/# cd website/
root@1a718f95e9f9:/website# npm install 
npm WARN deprecated stable@0.1.8: Modern JS already guarantees Array#sort() is a stable sort, so this library is deprecated. See the compatibility table on MDN: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/sort#browser_compatibility
npm WARN deprecated sourcemap-codec@1.4.8: Please use @jridgewell/sourcemap-codec instead
added 732 packages, and audited 733 packages in 60s
284 packages are looking for funding
 run `npm fund` for details
found 0 vulnerabilities
npm notice 
npm notice New major version of npm available! 8.19.3 -> 9.1.3
npm notice Changelog: https://github.com/npm/cli/releases/tag/v9.1.3
npm notice Run npm install -g npm@9.1.3 to update!
npm notice 
root@1a718f95e9f9:/website# npm run build
> @forgejo/website@0.0.1 build
> astro build
[MDX] Now inheriting remark and rehype plugins from "markdown" config.
If you applied a plugin to both your Markdown and MDX configs, we suggest removing the duplicate MDX entry.
See "extendPlugins" option to configure this behavior.
12:59:52 PM [build] output target: static
12:59:52 PM [build] Collecting build info...
12:59:52 PM [build] Completed in 484ms.
12:59:52 PM [build] Building static entrypoints...
12:59:54 PM [build] Completed in 2.48s.
 generating static routes 
▶ src/pages/index.astro
 └─ /index.html (+532ms)
λ src/pages/rss.xml.js
 └─ /rss.xml (+4ms)
▶ src/pages/404.astro
 └─ /404.html (+3ms)
▶ src/pages/faq.astro
 └─ /faq/index.html (+24ms)
▶ src/pages/[...tags]/[tag]/[...page].astro
 └─ /tag/blog/index.html (+24ms)
▶ src/pages/[...categories]/[category]/[...page].astro
▶ src/pages/[...blog]/[slug].astro
 ├─ /2022-12-15-hello-forgejo/index.html (+1.03s)
 ├─ /about/index.html (+1.06s)
 ├─ /code-of-conduct/index.html (+1.08s)
 ├─ /imprint/index.html (+1.16s)
 ├─ /privacy-policy/index.html (+1.18s)
 └─ /well-being/index.html (+1.20s)
▶ src/pages/[...blog]/[...page].astro
 ├─ /news/index.html (+443ms)
 └─ /news/2/index.html (+462ms)
 optimizing 2 images in batches of 8 
Completed in 2.32s.
Completed in 4.68s.
@astrojs/sitemap: `sitemap-index.xml` is created.
12:59:59 PM [build] 12 page(s) built in 7.65s
12:59:59 PM [build] Complete!

That's weird, it works when I do this. What's the difference I wonder. ``` $ docker run --rm -ti node:19 bash Unable to find image 'node:19' locally 19: Pulling from library/node f2f58072e9ed: Pull complete 5c8cfbf51e6e: Pull complete aa3a609d1579: Pull complete 094e7d9bb04e: Pull complete 2cbfd734f382: Pull complete 65f9380179fd: Pull complete 55fb38aec835: Pull complete 5a5cfed58206: Pull complete 8a4707ab05af: Pull complete Digest: sha256:3341143bf90c80e5b08694c706d49c04eea535f3d3cc0851f1698662f5713518 Status: Downloaded newer image for node:19 root@1a718f95e9f9:/# git clone -b wip-astrowind https://codeberg.org/forgejo/website Cloning into 'website'... remote: Enumerating objects: 619, done. remote: Counting objects: 100% (619/619), done. remote: Compressing objects: 100% (575/575), done. remote: Total 739 (delta 350), reused 111 (delta 30), pack-reused 120 Receiving objects: 100% (739/739), 3.43 MiB | 3.53 MiB/s, done. Resolving deltas: 100% (386/386), done. root@1a718f95e9f9:/# cd website/ root@1a718f95e9f9:/website# npm install npm WARN deprecated stable@0.1.8: Modern JS already guarantees Array#sort() is a stable sort, so this library is deprecated. See the compatibility table on MDN: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/sort#browser_compatibility npm WARN deprecated sourcemap-codec@1.4.8: Please use @jridgewell/sourcemap-codec instead added 732 packages, and audited 733 packages in 60s 284 packages are looking for funding run `npm fund` for details found 0 vulnerabilities npm notice npm notice New major version of npm available! 8.19.3 -> 9.1.3 npm notice Changelog: https://github.com/npm/cli/releases/tag/v9.1.3 npm notice Run npm install -g npm@9.1.3 to update! npm notice root@1a718f95e9f9:/website# npm run build > @forgejo/website@0.0.1 build > astro build [MDX] Now inheriting remark and rehype plugins from "markdown" config. If you applied a plugin to both your Markdown and MDX configs, we suggest removing the duplicate MDX entry. See "extendPlugins" option to configure this behavior. 12:59:52 PM [build] output target: static 12:59:52 PM [build] Collecting build info... 12:59:52 PM [build] Completed in 484ms. 12:59:52 PM [build] Building static entrypoints... 12:59:54 PM [build] Completed in 2.48s. generating static routes ▶ src/pages/index.astro └─ /index.html (+532ms) λ src/pages/rss.xml.js └─ /rss.xml (+4ms) ▶ src/pages/404.astro └─ /404.html (+3ms) ▶ src/pages/faq.astro └─ /faq/index.html (+24ms) ▶ src/pages/[...tags]/[tag]/[...page].astro └─ /tag/blog/index.html (+24ms) ▶ src/pages/[...categories]/[category]/[...page].astro ▶ src/pages/[...blog]/[slug].astro ├─ /2022-12-15-hello-forgejo/index.html (+1.03s) ├─ /about/index.html (+1.06s) ├─ /code-of-conduct/index.html (+1.08s) ├─ /imprint/index.html (+1.16s) ├─ /privacy-policy/index.html (+1.18s) └─ /well-being/index.html (+1.20s) ▶ src/pages/[...blog]/[...page].astro ├─ /news/index.html (+443ms) └─ /news/2/index.html (+462ms) optimizing 2 images in batches of 8 Completed in 2.32s. Completed in 4.68s. @astrojs/sitemap: `sitemap-index.xml` is created. 12:59:59 PM [build] 12 page(s) built in 7.65s 12:59:59 PM [build] Complete! ```

caesar commented 2022年12月06日 14:02:01 +01:00

Looks like this is more an installation problem than a CI problem. Does npm install work for you locally?

npm ERR! code 1
npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp
npm ERR! command failed
npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy)
npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br
npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies
npm ERR! sharp: Installation error: aborted
npm ERR! A complete log of this run can be found in:
npm ERR! /root/.npm/_logs/2022-12-06T12_18_39_470Z-debug-0.log

Yes it does. (On macOS M1.) What about for you?

Is the CI machine really x86? According to https://sharp.pixelplumbing.com/install both Sharp and libvips have to be built from source for x86 😐
Even so, not sure why it wouldn't work...

> Looks like this is more an installation problem than a CI problem. Does `npm install` work for you locally? > > ``` > npm ERR! code 1 > npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp > npm ERR! command failed > npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy) > npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br > npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies > npm ERR! sharp: Installation error: aborted > > npm ERR! A complete log of this run can be found in: > npm ERR! /root/.npm/_logs/2022-12-06T12_18_39_470Z-debug-0.log > ``` Yes it does. (On macOS M1.) What about for you? Is the CI machine really x86? According to https://sharp.pixelplumbing.com/install both Sharp and libvips have to be built from source for x86 😐 Even so, not sure why it wouldn't work...

Ghost commented 2022年12月06日 14:37:35 +01:00

The difference must be in how woodpecker feeds the commands to the container, although I'm not sure what it is. Splitting npm install && npm run... in two different lines gives better results.

https://ci.codeberg.org/forgejo/website/pipeline/161/3

Then there was the issue of having https::// instead of https:// in the URL (notice the double :) which led to the super cryptic error message:

# npm run build
> @forgejo/website@0.0.1 build
> astro build
01:35:04 PM [astro] Unable to load /website/astro.config.mjs
 error Cannot set property message of [object Object] which has only a getter
 File:
 /website/node_modules/astro/dist/core/errors/dev/utils.js:36:22
 Code:
 35 | const { location, pluginName, text } = buildError;
 > 36 | err[i].message = text;
 | ^
 37 | if (location) {
 38 | err[i].loc = { file: location.file, line: location.line, column: location.column };
 39 | err[i].id = err[0].id || (location == null ? void 0 : location.file);
 
 Stacktrace:
TypeError: Cannot set property message of [object Object] which has only a getter
 at file:///website/node_modules/astro/dist/core/errors/dev/utils.js:36:22
 at Array.forEach (<anonymous>)
 at collectErrorMetadata (file:///website/node_modules/astro/dist/core/errors/dev/utils.js:33:14)
 at handleConfigError (file:///website/node_modules/astro/dist/cli/index.js:76:38)
 at async file:///website/node_modules/astro/dist/cli/index.js:123:5
 at async runCommand (file:///website/node_modules/astro/dist/cli/index.js:117:76)
 at async cli (file:///website/node_modules/astro/dist/cli/index.js:168:5)

Which was fixed and now it works and runs a lot faster for some reason...

https://ci.codeberg.org/forgejo/website/pipeline/162/3

image

The difference must be in how woodpecker feeds the commands to the container, although I'm not sure what it is. Splitting `npm install && npm run...` in two different lines gives better results. https://ci.codeberg.org/forgejo/website/pipeline/161/3 Then there was the issue of having `https:://` instead of `https://` in the URL (notice the double **:**) which led to the super cryptic error message: ``` # npm run build > @forgejo/website@0.0.1 build > astro build 01:35:04 PM [astro] Unable to load /website/astro.config.mjs error Cannot set property message of [object Object] which has only a getter File: /website/node_modules/astro/dist/core/errors/dev/utils.js:36:22 Code: 35 | const { location, pluginName, text } = buildError; > 36 | err[i].message = text; | ^ 37 | if (location) { 38 | err[i].loc = { file: location.file, line: location.line, column: location.column }; 39 | err[i].id = err[0].id || (location == null ? void 0 : location.file); Stacktrace: TypeError: Cannot set property message of [object Object] which has only a getter at file:///website/node_modules/astro/dist/core/errors/dev/utils.js:36:22 at Array.forEach (<anonymous>) at collectErrorMetadata (file:///website/node_modules/astro/dist/core/errors/dev/utils.js:33:14) at handleConfigError (file:///website/node_modules/astro/dist/cli/index.js:76:38) at async file:///website/node_modules/astro/dist/cli/index.js:123:5 at async runCommand (file:///website/node_modules/astro/dist/cli/index.js:117:76) at async cli (file:///website/node_modules/astro/dist/cli/index.js:168:5) ``` Which was fixed and now it works and runs a lot faster for some reason... https://ci.codeberg.org/forgejo/website/pipeline/162/3 

caesar commented 2022年12月06日 22:09:51 +01:00

Great, thanks @dachary!

Looks like some things are broken in the preview build due to my use of root-relative paths, so I'll fix that.

Great, thanks @dachary! Looks like some things are broken in the preview build due to my use of root-relative paths, so I'll fix that.

caesar commented 2022年12月07日 12:25:09 +01:00

Tonight I have implemented something like @foxlet 's hero design, as well as adjusted a few other things.

I would appreciate help with the content, I think @circlebuilder had some ideas. At the least I'd love to come up with a few bullet points for "features" to add to the home page.

---

@dachary it looks like the CI isn't producing the output images correctly... not sure why. It works locally on my machine.
Although the CI was "succesful" you can see the following in the log:

11:16:46 AM [@astrojs/image] "d/assets/default.c6e14b88.png" image could not be fetched
11:16:46 AM [@astrojs/image] "d/assets/mascot-dark.59062dfc.png" image could not be fetched
11:16:46 AM [@astrojs/image] "d/assets/astronaut.53563f21.jpg" image could not be fetched

and sure enough the images don't exist in the output.

---

Since the CI didn't procuce the right output, I've attatched a screenshot of what it was supposed to look like.

Tonight I have implemented something like @foxlet 's hero design, as well as adjusted a few other things. I would appreciate help with the content, I think @circlebuilder had some ideas. At the least I'd love to come up with a few bullet points for "features" to add to the home page. --- @dachary it looks like the CI isn't producing the output images correctly... not sure why. It works locally on my machine. Although the CI was "succesful" you can see the following in the log: ```log 11:16:46 AM [@astrojs/image] "d/assets/default.c6e14b88.png" image could not be fetched 11:16:46 AM [@astrojs/image] "d/assets/mascot-dark.59062dfc.png" image could not be fetched 11:16:46 AM [@astrojs/image] "d/assets/astronaut.53563f21.jpg" image could not be fetched ``` and sure enough the images [don't exist](https://forgejo.codeberg.page/@wip-astrowind/assets/mascot-dark.59062dfc_Zz1OY3.png) in the output. --- Since the CI didn't procuce the right output, I've attatched a screenshot of what it was supposed to look like. 

Ghost commented 2022年12月07日 13:47:22 +01:00

You can reproduce the problem locally by changing the cat src/config.mjs with something like:

export const SITE = {
	name: 'Forgejo',
	origin: 'https://forgejo.codeberg.page',
	basePathname: '/@wip-astro/',

There is "something" in astro that prepends the last two letters of the basePathname to the asset.

12:42:31 PM [@astrojs/image] "o/assets/default.c6e14b88.png" image could not be fetched
12:42:31 PM [@astrojs/image] "o/assets/mascot-dark.59062dfc.png" image could not be fetched
12:42:31 PM [@astrojs/image] "o/assets/astronaut.53563f21.jpg" image could not be fetched

and in your case it's a d/ because those are the last two letters of '/@wip-astro/'

You can reproduce the problem locally by changing the `cat src/config.mjs` with something like: ``` export const SITE = { name: 'Forgejo', origin: 'https://forgejo.codeberg.page', basePathname: '/@wip-astro/', ``` There is "something" in astro that prepends the last two letters of the basePathname to the asset. ``` 12:42:31 PM [@astrojs/image] "o/assets/default.c6e14b88.png" image could not be fetched 12:42:31 PM [@astrojs/image] "o/assets/mascot-dark.59062dfc.png" image could not be fetched 12:42:31 PM [@astrojs/image] "o/assets/astronaut.53563f21.jpg" image could not be fetched ``` and in your case it's a `d/` because those are the last two letters of '/@wip-astro/'

caesar commented 2022年12月08日 11:01:29 +01:00

The image pipeline is now working.
It turns out that Astro is stripping the trailing slash from the base parameter (if the trailingSlash option is set to never) – even though it wants it to have a trailing slash! This causes problems with @astrojs/image.
Im not sure exactly whther the bug is in Astro or @astrojs/image, but either way, adding a second trailing slash to the base parameter solves it, so I've done that for the CI branch builds.

---

However, there is still a problem with the npm install step, which fails around 50% of the time when installing sharp.
The log output is as follows:

npm ERR! code 1
npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp
npm ERR! command failed
npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy)
npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br
npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies
npm ERR! sharp: Installation error: aborted

I don't have any ideas here really so if someone more familiar with Woodpecker and the machines it's running on has some ideas I'd be grateful.

The image pipeline is now working. It turns out that Astro is stripping the trailing slash from the `base` parameter (if the `trailingSlash` option is set to `never`) – even though it wants it to have a trailing slash! This causes problems with [@astrojs/image](https://docs.astro.build/en/guides/integrations-guide/image/). Im not sure exactly whther the bug is in Astro or @astrojs/image, but either way, adding a *second* trailing slash to the `base` parameter solves it, so I've done that for the CI branch builds. --- However, there is still a problem with the `npm install` step, which fails around 50% of the time when installing `sharp`. The log output is as follows: ```log npm ERR! code 1 npm ERR! path /woodpecker/src/codeberg.org/forgejo/website/node_modules/sharp npm ERR! command failed npm ERR! command sh -c -- (node install/libvips && node install/dll-copy && prebuild-install) || (node install/can-compile && node-gyp rebuild && node install/dll-copy) npm ERR! sharp: Downloading https://github.com/lovell/sharp-libvips/releases/download/v8.13.3/libvips-8.13.3-linux-x64.tar.br npm ERR! sharp: Please see https://sharp.pixelplumbing.com/install for required dependencies npm ERR! sharp: Installation error: aborted ``` I don't have any ideas here really so if someone more familiar with Woodpecker and the machines it's running on has some ideas I'd be grateful.

Ghost commented 2022年12月08日 14:54:07 +01:00

The "sharp" transient error is tracked here #38

The "sharp" transient error is tracked here https://codeberg.org/forgejo/website/issues/38

Ryuno-Ki commented 2022年12月08日 16:08:24 +01:00

Regarding the share button: https://adactio.com/journal/17405 would be my suggestion.
We can address them in a follow-up PR.

Regarding the share button: https://adactio.com/journal/17405 would be my suggestion. We can address them in a follow-up PR.

caesar commented 2022年12月09日 12:14:17 +01:00

Today I started work on styling the blog pages.
I decided to style the blog article list as a "commit graph".
See https://forgejo.codeberg.page/@wip-astrowind/news/ (WIP).

Today I started work on styling the blog pages. I decided to style the blog article list as a "commit graph". See https://forgejo.codeberg.page/@wip-astrowind/news/ (WIP).

caesar commented 2022年12月11日 11:42:49 +01:00

Today I added a (WIP) releases page and releases RSS feed to the site.

These are generated at build time using the Forgejo API to fetch the list of releases from Codeberg, so we will need to trigger rebuilds of the site via a webhook when releases are published.

Today I added a (WIP) [releases page](https://forgejo.codeberg.page/@wip-astrowind/releases/) and [releases RSS feed](https://forgejo.codeberg.page/@wip-astrowind/releases/rss.xml) to the site. These are generated at build time using the Forgejo API to fetch the list of releases from Codeberg, so we will need to trigger rebuilds of the site via a webhook when releases are published.

caesar commented 2022年12月12日 12:13:00 +01:00

Today I improved the releases page and the static text pages somewhat, and added a download page which also automatically shows instructions for the latest release (generated at site build time).

The download page needs a bit more content (explanation of what the steps do).

Today I improved the [releases page](https://forgejo.codeberg.page/@wip-astrowind/releases/) and the static text pages somewhat, and added a [download page](https://forgejo.codeberg.page/@wip-astrowind/download/) which also automatically shows instructions for the latest release (generated at site build time). The download page needs a bit more content (explanation of what the steps do).

Ghost commented 2022年12月12日 14:00:21 +01:00

As a Gitea admin or contributor, the download page will be enough for me. I agree it would need more content for a wider audience.

It looks excellent 👍

As a Gitea admin or contributor, the download page will be enough for me. I agree it would need more content for a wider audience. It looks excellent 👍

caesar commented 2022年12月13日 12:44:57 +01:00

Tonight I worked on the theme toggle. You won't see much difference, but it should be robust now and respect the user's preference when Javascript is disabled...

Tomorrow I'll go back to more visible things. 😆

Only a coujple of days left to launch now, so the content is a priority. Not much is needed; mostly just a few "features" for the home page. I can probably pull some things from the FAQ or About page (there also seems to be a lot of overlap between those two.)

Tonight I worked on the theme toggle. You won't see much difference, but it should be robust now and respect the user's preference when Javascript is disabled... Tomorrow I'll go back to more visible things. 😆 Only a coujple of days left to launch now, so the content is a priority. Not much is needed; mostly just a few "features" for the home page. I can probably pull some things from the FAQ or About page (there also seems to be a lot of overlap between those two.)

Ghost commented 2022年12月13日 13:48:35 +01:00

Tomorrow I'll go back to more visible things.

I got used to a daily eye candy, I must say :-)

Not much is needed; mostly just a few "features" for the home page.

I'll do that today.

> Tomorrow I'll go back to more visible things. I got used to a daily eye candy, I must say :-) > Not much is needed; mostly just a few "features" for the home page. I'll do that today.

Ghost commented 2022年12月14日 00:27:01 +01:00

I pushed df16653018 for the features. It's a lazy copy/paste but maybe it's good enough?

I pushed https://codeberg.org/forgejo/website/commit/df166530186bcab033f63ba56e84bced8432bca5 for the features. It's a lazy copy/paste but maybe it's good enough?

caesar commented 2022年12月14日 14:28:20 +01:00

I think I've done about as much as I can before the launch. I'll give it a final check over tomorrow, but that's already the morning of launch day in Europe.

Please could folks give it a look over and let me know if anything importatnt needs to be changed urgently before the launch?

---

To publish this when we're ready, I suggest that instead of trying to merge it into main, it's better to reset main to this branch and force-push. We should add a new name to the current main branch first to keep it around for the moment.

I think I've done about as much as I can before the launch. I'll give it a final check over tomorrow, but that's already the morning of launch day in Europe. Please could folks give it a look over and let me know if anything importatnt needs to be changed urgently before the launch? --- To publish this when we're ready, I suggest that instead of trying to merge it into `main`, it's better to reset `main` to this branch and force-push. We should add a new name to the current `main` branch first to keep it around for the moment.

Ghost commented 2022年12月14日 14:39:23 +01:00

Here are my comments after carefully reading everything. It looks great 👍 I went ahead and pushed commits to fix trivial typos / wording that do not require discussions (see b1ba448183 and following).

I would remove whatever needs further discussion because it is not obviously agreed upon by everyone, from my point of view. This includes:

Philosophy

• Philosophy: Simplicity, Best Practices and High Performance

because it will likely be something like (or related to) the "Core Values" of Forgejo once they are agreed on, once the decision process is in place.

Feature Intuitive software project management

Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen.

Feature Lightweight and performant

I would keep this on "Lightweight" and remove "and performant". The reason there is a focus on scaling in Forgejo is because it does not perform well at the Codeberg scale and needs work. Reason why I think it is accurate to write that there is a focus on scaling. But claiming that it performs well at scale is going one step too far.

I propose this as an alternative:

With a rich feature set, Forgejo still has a low server profile and can run on an inexpensive Raspberry Pi. Save your machine energy!

Here are my comments after carefully reading everything. It looks great 👍 I went ahead and pushed commits to fix trivial typos / wording that do not require discussions (see b1ba44818367c8b8889d5145d055b8c2ac110397 and following). I would remove whatever needs further discussion because it is not obviously agreed upon by everyone, from my point of view. This includes: ### Philosophy * Philosophy: Simplicity, Best Practices and High Performance because it will likely be something like (or related to) the "Core Values" of Forgejo once they are agreed on, once the decision process is in place. ### Feature Intuitive software project management Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen. ### Feature Lightweight and performant I would keep this on "Lightweight" and remove "and performant". The reason there is a focus on scaling in Forgejo is because it does not perform well at the Codeberg scale and needs work. Reason why I think it is accurate to write that there is a focus on scaling. But claiming that it performs well at scale is going one step too far. I propose this as an alternative: > With a rich feature set, Forgejo still has a low server profile and can run on an inexpensive Raspberry Pi. Save your machine energy!

Ghost commented 2022年12月14日 15:19:04 +01:00

We should add a new name to the current main branch first to keep it around for the moment.

I pushed the current main under the name 2022年12月14日-main for archive. I agree to the force push.

> We should add a new name to the current main branch first to keep it around for the moment. I pushed the current main under the name [2022年12月14日-main](https://codeberg.org/forgejo/website/src/branch/2022-12-14-main) for archive. I agree to the force push.

circlebuilder commented 2022年12月14日 17:06:57 +01:00

Re: @dachary

Feature Intuitive software project management

Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen.

This feature describes something that already there. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github. Feature has nothing to do with any further user research / UX track for new roadmap items and whether people are available for that.

Re: @dachary >Feature Intuitive software project management > >Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen. This feature describes something that _already there_. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github. Feature has nothing to do with any further user research / UX track for new roadmap items and whether people are available for that.

Ghost commented 2022年12月14日 17:58:19 +01:00

This feature describes something that already there. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github.

To be more specific, Forgejo:

• is prone to UI regressions (primarily because there is no automated testing)
• has many inconsistent user interfaces (there is no UX work prior to adding a UI)
• contains a large number of features that are unmaintained (because it has only partial integration testing coverage)
• has a fair number of surprising user experience pathways

I could be even more specific by providing examples for each of those, in case you think I'm mistaken. But I think some of them come to mind already 😄

That being said, the user experience is nevertheless considered to be good enough by all people I had the opportunity to talk to. Including myself: I'm careful to not use what I know to be fragile and I learned my way around the most unintuitive parts.

For this reason I think Forgejo is a long way from being "Intuitive software project management". This could be replaced by "Good enough software project management" but maybe that's not very appealing in a feature list.

> This feature describes something that already there. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github. To be more specific, Forgejo: * is prone to UI regressions (primarily because there is no automated testing) * has many inconsistent user interfaces (there is no UX work prior to adding a UI) * contains a large number of features that are unmaintained (because it has only partial integration testing coverage) * has a fair number of surprising user experience pathways I could be even more specific by providing examples for each of those, in case you think I'm mistaken. But I think some of them come to mind already 😄 That being said, **the user experience is nevertheless considered to be good enough by all people I had the opportunity to talk to**. Including myself: I'm careful to not use what I know to be fragile and I learned my way around the most unintuitive parts. For this reason I think Forgejo is a long way from being "Intuitive software project management". This could be replaced by "Good enough software project management" but maybe that's not very appealing in a feature list.

caesar commented 2022年12月14日 21:03:24 +01:00

Philosophy

Philosophy: Simplicity, Best Practices and High Performance

because it will likely be something like (or related to) the "Core Values" of Forgejo once they are agreed on, once the decision process is in place.

As I mentioned above to @circlebuilder, this is placeholder text that I have already removed from the site precisely for the reason you mention. (The source file is still present because we will probably reuse it later, but it's not included on the site anywhere.)

Re: @dachary

Feature Intuitive software project management

Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen.

This feature describes something that already there. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github. Feature has nothing to do with any further user research / UX track for new roadmap items and whether people are available for that.

I think we should keep this bullet point as the overall gist of it is both largely true currently (though there are a lot of improvements to be made) and an important goal even if UI research is not currently underway.

However we could swap the word intuitive for easy which would be perhaps less of a strong claim. Would that be okay?

> Philosophy > > Philosophy: Simplicity, Best Practices and High Performance > > because it will likely be something like (or related to) the "Core Values" of Forgejo once they are agreed on, once the decision process is in place. As I mentioned above to @circlebuilder, this is placeholder text that I have already removed from the site precisely for the reason you mention. (The source file is still present because we will probably reuse it later, but it's not included on the site anywhere.) > Re: @dachary > > >Feature Intuitive software project management > > > >Although I think User Research and User eXperience should be a focus of Forgejo, there is a (very) long way to go. Until it happens I don't think "Intuitive" is something that accurately applies to Forgejo and most importantly there is noone committed to make it happen. > > This feature describes something that _already there_. The software we are using now (Codeberg) is rather slick and intuitive esp. as FOSS goes, and its almost like Github. Feature has nothing to do with any further user research / UX track for new roadmap items and whether people are available for that. I think we should keep this bullet point as the overall gist of it is both *largely true* currently (though there are a *lot* of improvements to be made) and an important goal even if UI research is not currently underway. However we could swap the word `intuitive` for `easy` which would be perhaps less of a strong claim. Would that be okay?

caesar commented 2022年12月14日 21:27:40 +01:00

Feature Lightweight and performant

I would keep this on "Lightweight" and remove "and performant". The reason there is a focus on scaling in Forgejo is because it does not perform well at the Codeberg scale and needs work. Reason why I think it is accurate to write that there is a focus on scaling. But claiming that it performs well at scale is going one step too far.

This was discussed on Matrix with input from @fnetX and @dachary.

I will change to the following wording:

With a rich feature set, Forgejo still has a low server profile and requires an order of magnitude less resources than other forges.

> ### Feature Lightweight and performant > > I would keep this on "Lightweight" and remove "and performant". The reason there is a focus on scaling in Forgejo is because it does not perform well at the Codeberg scale and needs work. Reason why I think it is accurate to write that there is a focus on scaling. But claiming that it performs well at scale is going one step too far. This was [discussed on Matrix](https://matrix.to/#/!qjPHwFPdxhpLkXMkyP:matrix.org/$SM_TICmmQ77Nuvkq0FlNfBaTduhrGhClGKPCprWCYio) with input from @fnetX and @dachary. I will change to the following wording: > With a rich feature set, Forgejo still has a low server profile and requires an order of magnitude less resources than other forges.

Ghost commented 2022年12月14日 21:29:52 +01:00

However we could swap the word intuitive for easy which would be perhaps less of a strong claim. Would that be okay?

Yes. As @fr33domlover would say, this is something I'm happy to live with 😄

> However we could swap the word intuitive for easy which would be perhaps less of a strong claim. Would that be okay? Yes. As @fr33domlover would say, this is something I'm happy to live with :smile:

caesar commented 2022年12月14日 23:01:33 +01:00

This is merged as of 78356c6552 but I will continue to make some small improvements so please continue to leave suggestions here.

This is merged as of https://codeberg.org/forgejo/website/commit/78356c65523b428546a1fa7965911fd1f179232e but I will continue to make some small improvements so please continue to leave suggestions here.

circlebuilder commented 2022年12月15日 08:19:18 +01:00

One more change needed in the footer to attribute David Revoy properly. See: https://www.davidrevoy.com/article605/best-practices-for-attribution

Text would be (as markdown):

["Forgejo"](https://www.peppercarrot.com/en/viewer/comissions-src__2022年11月27日_Forgejo_by-David-Revoy.html) mascot by David Revoy, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)

One more change needed in the footer to attribute David Revoy properly. See: https://www.davidrevoy.com/article605/best-practices-for-attribution Text would be (as markdown): ```markdown ["Forgejo"](https://www.peppercarrot.com/en/viewer/comissions-src__2022年11月27日_Forgejo_by-David-Revoy.html) mascot by David Revoy, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) ```

caesar commented 2022年12月15日 09:27:00 +01:00

Thanks @circlebuilder, fixed in 20f3d614f9.

Thanks @circlebuilder, fixed in https://codeberg.org/forgejo/website/commit/20f3d614f993c24370065b484354efe393213d26.