Rewrite of the website

[kazam0180]

I have looked through it fixed the issues I found. You might want to check FAQ and Wiki page @ivan-hc

[kazam0180]

Some things that need to be done:

• Add APPIMAGE URLs from old markdown files to the sources field in the new json files
• Add json parsing support in AM, AppGallery (and vappman too?)
• Add pla-install URI handling in AM-GUI and AppGallery

The CONTRIBUTING.md is on it's way. Did I miss something?

[ivan-hc]

@andy5995 can you review this?

We also talked about this on Discord.

[kazam0180]

added source urls from both UNOFFICIAL APPIMAGE: and APPIMAGE: lines in some apps

[andy5995]

@ivan-hc I will try.

In my opinion, if a lot of changes are needed for a website rewrite, then it normally should be done in smaller chunks so changes can be efficiently reviewed by the maintainer. If this were my project, I'd suggest a contributor submit the first required stage in a pull request. After that is reviewed and merged, then a "stage 2" commit and PR.

Why are over 3000 files getting changed in this single pull request? The site is normally generated using tools and scripts from the AM repo. If the existing site generator can't write the proposed changes, then my first thought would be a single converter script is needed in the AM repo that will implement the proposed changes to all the appimage entries.

I can't think of any more feedback right now, but if I do, I'll post here. And maybe I'll be able to review more once my questions are answered.

[ivan-hc]

@andy5995 I completely agree. Changes like this need to be analyzed in gradual phases. @coyoteclan here are too many changes to analyze.

[kazam0180]

Why are over 3000 files getting changed in this single pull request? The site is normally generated using tools and scripts from the AM repo. If the existing site generator can't write the proposed changes, then my first thought would be a single converter script is needed in the AM repo that will implement the proposed changes to all the appimage entries.

This is a catalogue, it's expected. The generator script and workflow is updated. It does generate the new website. But I had to test it so I enabled the workflow and ran it. You can see individual commits.
Of course I had to convert the existing apps manually first to retain their data. But new apps and stats are generated like they were previously.

[andy5995]

Why are over 3000 files getting changed in this single pull request? The site is normally generated using tools and scripts from the AM repo. If the existing site generator can't write the proposed changes, then my first thought would be a single converter script is needed in the AM repo that will implement the proposed changes to all the appimage entries.

This is a catalogue, it's expected. The generator script and workflow is updated. It does generate the new website. But I had to test it so I enabled the workflow and ran it. You can see individual commits. Of course I had to convert the existing apps manually first to retain their data. But new apps and stats are generated like they were previously.

But there are thousands of changed files and thousands of changed lines in this commit. In cases like this, you need to eliminate all the changes that aren't required, submit changes for the script additions and changes so the maintainer (@ivan-hc ) can examine them without the extra noise. I understand you wanted to test, and that's great, but I believe @ivan-hc would need you to remove your test output/artifacts from the PR.

[kazam0180]

Why are over 3000 files getting changed in this single pull request? The site is normally generated using tools and scripts from the AM repo. If the existing site generator can't write the proposed changes, then my first thought would be a single converter script is needed in the AM repo that will implement the proposed changes to all the appimage entries.

This is a catalogue, it's expected. The generator script and workflow is updated. It does generate the new website. But I had to test it so I enabled the workflow and ran it. You can see individual commits. Of course I had to convert the existing apps manually first to retain their data. But new apps and stats are generated like they were previously.

But there are thousands of changed files and thousands of changed lines in this commit. In cases like this, you need to eliminate all the changes that aren't required, submit changes for the script additions and changes so the maintainer (@ivan-hc ) can examine them without the extra noise. I understand you wanted to test, and that's great, but I believe @ivan-hc would need you to remove your test output/artifacts from the PR.

okay cool. I need further testing. I will copy the generated output in apps/ and re-add the old files temporarily when i am done. after the PR is merged, we can re-add the new app files back.

[kazam0180]

@andy5995 i proposed a simple syntax for defining app info on pkgforge discord:

# some app
description of your app.
only use bold, italic, underline, lists markdown here
do not add links or images here
use SOURCES field if you package someone else's app

# SCREENSHOTS: https://someapp.io/1.png https://someapp.io/2.png
# SITES: https://someapp.io
# SOURCES: https://github.com/name/someapp-appimage
# BUTTONS: Label_of_Button::https://discord.gg Donation_Link::https://someapp.io/donate

the app.html file uses json

this info would then be converted to json by a script during pages deployment and served on the website
so this essentially gives us a nice API on website and AM can still get the raw info files from github repo because @ivan-hc prefers not to use jq and complicate the am -a option

this new syntax will simplify the AM about logic too because the format was pretty inconsistent previously

i am currently working on this

[kazam0180]

@andy5995 @ivan-hc done.

this app info (AM can use this)
gets converted to this json
and shows as this page

check it out, check other parts of the website if you have the time, if there are changes to be made, i will do it. Then i will remove the 3369 apps generated and replace them with old md files for PR review

[andy5995]

I don't see this as being desirable. It would be an extra maintenance burden for this project to have links to the Discord, donation buttons for other projects. How would that information get updated when the links change or become dead over time? @ivan-hc agree?

buttons	
0	"Label_of_Button::https://discord.gg"
1	"Donation_Link::https://someapp.io/donate"

That's from https://coyoteclan.github.io/Portable-Linux-Apps.github.io/apps/testapp.json

[kazam0180]

I don't see this as being desirable. It would be an extra maintenance burden for this project to have links to the Discord, donation buttons for other projects. How would that information get updated when the links change or become dead over time? @ivan-hc agree?

buttons	
0	"Label_of_Button::https://discord.gg"
1	"Donation_Link::https://someapp.io/donate"

That's from https://coyoteclan.github.io/Portable-Linux-Apps.github.io/apps/testapp.json

That's for an example. The original contributing.md said you can add extra buttons. So I included a way to define extra buttons. The matter of links becoming dead is same as it was before.
We could use a workflow to check for dead links, an action for that already exists.

[ivan-hc]

To write a markdown is simpler than compiling a json, and more customizable for common users.

[kazam0180]

To write a markdown is simpler than compiling a json, and more customizable for common users.

i think you didn't read my message properly. Users write this:

# some app
description of your app.
only use **bold**, *italic*, ++underline++, lists markdown here
- do not add links or images here
- use SOURCES field if you package someone else's app

# SCREENSHOTS: https://coyoteclan.github.io/Portable-Linux-Apps.github.io/contribute_ss.png https://coyoteclan.github.io/Portable-Linux-Apps.github.io/contribute_ss.png
# SITES: https://someapp.io
# SOURCES: https://github.com/name/someapp-appimage
# BUTTONS: Label_of_Button::https://discord.gg Donation_Link::https://someapp.io/donate

it gets converted to json for site deployment automatically by the workflow but this info file remains intact in the repo. no json files are committed. the json files are only used on website. users write and edit the above mentioned markdown-ish format.

so like i said before, we are getting both benefits now:

• contributors only edit a simple info file
• we get a nice json api on the website

i updated the faq entry on how to improve app pages

[andy5995]

That's for an example. The original contributing.md said you can add extra buttons. So I included a way to define extra buttons.

Oh, I see now. :)

[andy5995]

@andy5995 I completely agree. Changes like this need to be analyzed in gradual phases. @coyoteclan here are too many changes to analyze.

This general guide has more information about "best practices" https://www.deployhq.com/blog/the-perfect-pull-request-best-practices-for-collaborative-development

[andy5995]

@andy5995 @ivan-hc done.

this app info (AM can use this) gets converted to this json and shows as this page

check it out, check other parts of the website if you have the time, if there are changes to be made, i will do it. Then i will remove the 3369 apps generated and replace them with old md files for PR review

@coyoteclan Please post the screenshots here instead of the links, then remove all the generated artifacts, update the PR. Then we will be able to review the code that changes the website infrastructure.

[kazam0180]

@andy5995 @ivan-hc done.
this app info (AM can use this) gets converted to this json and shows as this page
check it out, check other parts of the website if you have the time, if there are changes to be made, i will do it. Then i will remove the 3369 apps generated and replace them with old md files for PR review

@coyoteclan Please post the screenshots here instead of the links, then remove all the generated artifacts, update the PR. Then we will be able to review the code that changes the website infrastructure.

How apps are defined by contributors or "dumb" apps generated by sync with AM workflow:

This info file is converted to json and served on the deployed site:

And the app.html page renders the json data like this:

real app example, virtualbox:

[kazam0180]

thanks for the article link. i will do it now

[kazam0180]

@ivan-hc i noticed now i have missing categories. you pointed it out before but i didn't pay much attention. i used the script that was present on this repo instead of the AM one. fixing it now and improving the overall logic of category and category page generation

[kazam0180]

done, we are down to 71 files changed, most of them are deleted files.

run by sync with AM worklfow:

am2pla-site (in this repo) now does only these things:

• creates apps.json (all apps, used on the Applications list)
• updates the stats in index.html

run by pages deployment workflow:

simple2json.py replaces the simple app info files in apps/ with json files (used by app.html)

gen_cats.py generates categories/*.json files

gen_cat_pages.py generates <category>.html files in root that render the category data from categories/

@andy5995 @ivan-hc

[kazam0180]

actually, apps.json too shouldn't be tracked in git, i will move that to the website deployment

[kazam0180]

done, now we're golden, apps.json is no longer tracked in git. it only lives on website, stats are also not updated in the repo

sync with AM now only removes unnecessary app info files and creates dumb info and icon files
when sync with AM commits new apps/removes apps, the website deployment workflow runs, it downloads x86_64-apps, stats-appimages and stats-portable.
it generates category json files, category html files, apps.json in root and updates the stats in index.html

[kazam0180]

there is a problem in website deployment, i am debugging it at https://github.com/Wolf-Pack-Clan/Portable-Linux-Apps.github.io

[kazam0180]

okay now everything seems good

[kazam0180]

@andy5995 @ivan-hc

[andy5995]

@kazam0180 50,000 changed lines is still way too many changes for a PR like this. And there are merge conflicts that would have to be fixed. Have you AI agent review the PR and the comments left by @ivan-hc and me, and post what it says here.

[kazam0180]

@kazam0180 50,000 changed lines is still way too many changes for a PR like this. And there are merge conflicts that would have to be fixed. Have you AI agent review the PR and the comments left by @ivan-hc and me, and post what it says here.

@andy5995 first of all i am not that stupid, i dont fully vibe code. i used AI to win time because it would take me like 3 or 4 days to get the html and css like this and AI is pretty good at UI and UX stuff.

all of the comments are already addressed since #167 (comment). i cant reduce any more changes. this is the minimum possible.
those 48k deletions are just the long category pages deleted (which are now no longer committed to the repository, generated and deployed on website only like i said before)

the conflicts are because of the deleted category pages. maybe i should re-add them 🤔

[andy5995]

@kazam0180 50,000 changed lines is still way too many changes for a PR like this. And there are merge conflicts that would have to be fixed. Have you AI agent review the PR and the comments left by @ivan-hc and me, and post what it says here.

@andy5995 first of all i am not that stupid, i dont fully vibe code. i used AI to win time because it would take me like 3 or 4 days to get the html and css like this and AI is pretty good at UI and UX stuff.

I didn't say you were stupid. I also use LLM to save time.

[kazam0180]

what i meant by "not that stupid" was that i can and have solved the issues in the PR myself

so you gonna take a look?
just see the new files and the modified workflow

[kazam0180]

@andy5995 should i create a new fork with the new files only? or discard commits then add new files

[ivan-hc]

Guys, I'd love to have a modicum of your web development skills. Unfortunately, my knowledge doesn't extend beyond BASH/SHELL and Markdown, so I can't be of much help. Sorry for the silence.

[kazam0180]

Guys, I'd love to have a modicum of your web development skills. Unfortunately, my knowledge doesn't extend beyond BASH/SHELL and Markdown, so I can't be of much help. Sorry for the silence.

No problem at all!

[andy5995]

@kazam0180 If this were my project, I'd say... before you fork and make any more changes, start a Discussion on the AM repo that sketches an outline of the architecture of the changes you'd like to introduce. From there, we'll see about taking things step-by-step.

@ivan-hc do you agree?

[ivan-hc]

Yes @andy5995 , the ideas are good and it looks even better... but I feel inundated with information and strings to analyze that I wouldn't be able to decipher without the proper knowledge.

[kazam0180]

alright, will do