Closed zen0bit closed 5 months ago
Before I do this, I need to improve the README... or at least, I need to write an HTML for that. If you see, the URLs are not working and also the menus to expand.
There is a lot of work to do on the README/INDEX
@zen0bit https://markdowntohtml.com/
there is still some work to do, but in usability its ways better, don't you?
Maybe if I use Google Sites as an editor, I can reach the same result as https://bananahackers.net
or https://sites.google.com/view/kaiositalia (the italian version of bananahackers.net)
I only need to have a structure
AM looks promising But application manager is useless there. No one will read it Or for where is this logo?
I am thinking about favicon now...
it's a unique signature, there's no need for others to read it, I've seen lots of AM logos on the web... I wrote this one using gimp/gnome-paint, to distinguish it
I only need to have a structure
of?
I leave this up to you, don't know how you want AM site should looks like... Your choice...
I was thinking of giving it the same structure as bananahackers.net, but with fewer pages... more "Kaios italia"
I like how github kill openning images of others because they cant fix the exploit which was there few years
I only need to have a structure
of?
I leave this up to you, don't know how you want AM site should looks like... Your choice...
I'd like to add 6 pages or urls, to complete the space under the header
And 6 icons
And maybe some music, strictly rock (Nirvana, RATM, AC/DC... I can use one of my playlists )
I only need to have a structure
of? I leave this up to you, don't know how you want AM site should looks like... Your choice...
I'd like to add 6 pages or urls, to complete the space under the header
And 6 icons
And maybe some music, strictly rock (Nirvana, RATM, AC/DC... I can use one of my playlists )
Don't see problem with that
In this moment I've no idea of what pages I should include.
In this moment I've no idea of what pages I should include.
🤪 I thought that you have six, not that you have theme that needs six 🤣
maybe I can try another approach
PS: I suggest not using spoilers in site Implementation lack in real world so it wouldn't work for many people...
I recently study this and is better to leave it, until will be really usable across browsers...
PS: Since I personally hate white themes, i suggest use theme with switchable nigh/day black/white
sure, I hate it too... now its just a draft. But I always prefer black themes.
sure, I hate it too... now its just a draft. But I always prefer black themes.
Why you start with white then...?
because the page of Google Sites is white by default
colors and banners will came later
but after this, when I'll upload the pages, I need also to remove google's references and trackers from it, obviously
I don't like that style
Maybe 10 years back OK
first you dont want any picture behind text. It will make this less readable
You have to count with both color themes... So with white and black text must be seen anywhere...
That's no go It wouldn't work like you expect...
I'll not use a Matrix style, bananahackers.net was inspired to the remake of Nokia 8110 used by "Neo" in the "Matrix" movie. Ence the choice of the theme.
first you dont want any picture behind text. It will make this less readable
You have to count with both color themes... So with white and black text must be seen anywhere...
That's no go It wouldn't work like you expect...
I have to think about the colors later, now I don't even have a structure to think about
Again: Just to not really miss this:
I suggest not using spoilers in site Implementation lack in real world so it wouldn't work for many people...
I recently study this and is better to leave it, until will be really usable across browsers...
Again: Just to not really miss this:
I suggest not using spoilers in site Implementation lack in real world so it wouldn't work for many people...
I recently study this and is better to leave it, until will be really usable across browsers...
OK
the main problem arises from the fact that portable-linux-apps.github.io is already the very heart of the AM project. AM would be meaningless if there were no installation scripts to install applications. And here another problem arises: the database to be rebuilt from scratch.
It should be included in the new site, or at least embedded. People want AM for apps, not for reading help. A main page with the same content as the README would be a useless duplicate.
portable-linux-apps.github.io also includes a JSON file with all applications listed, so for a a good HTML developer would be easy to improve this
Again: Just to not really miss this: I suggest not using spoilers in site Implementation lack in real world so it wouldn't work for many people... I recently study this and is better to leave it, until will be really usable across browsers...
OK
the main problem arises from the fact that portable-linux-apps.github.io is already the very heart of the AM project. AM would be meaningless if there were no installation scripts to install applications. And here another problem arises: the database to be rebuilt from scratch.
It should be included in the new site, or at least embedded. People want AM for apps, not for reading help. A main page with the same content as the README would be a useless duplicate.
Why should be rebuild?
For now I'll left it as the readme of the repository. If I get inspired, I'll come back to work on it.
If you have other ideas, feel free to add/edit whatever you want.
Address is hardcoded everywhere? Can't database be just copied?
Why should be rebuild?
I had another idea about refactoring of the installation scripts database... but for now its far from reality: to resume all commands on a JSON file or an online database, including all descriptions and commands to use to unstall the apps and create their AM-updaters.
The issue is not much the huge quantity of the installation scripts but to have a structure for that.
Address is hardcoded everywhere? Can't database be just copied?
in what sense?
Why should be rebuild?
I had another idea about refactoring of the installation scripts database... but for now its far from reality: to resume all commands on a JSON file or an online database, including all descriptions and commands to use to unstall the apps and create their AM-updaters.
The issue is not much the huge quantity of the installation scripts but to have a structure for that.
Which structure you mean? or better to say structure of what?
You don't want everything in one JSON file...
Which structure you mean?
One file? More files?
How should each file be compiled to resume and mege together a message (for -a), the download urls and the extraction (-i) and update (-u)?
I'd like to have a database with the updated urls for each application, so the installation scripts should no more rely on what is hosted but tho the still working url, see for example https://github.com/ivan-hc/AM/issues/539
again, this is far from reality for now... its just a fantasy of mine
but this is an off topic
One template per app as I said
As simple as posible. Main power will be that anyone can add own AppImage.
So It must be really simple..
Main question is how much from installation scripts differ between apps.
Just theoreticaly: Sort them to (from github, sourceforge, gitlab, gitea, ... , own)
If you still have most apps in own categories we will have problems. If you have just few we are good.
more than an app template... a reference URL for download. See here, from the lmstudio script I've fixed yesterday:
version=$(mkdir -p "$HOME"/.cache && cd "$HOME"/.cache && wget -rq https://lmstudio.ai 2>&1 && cat ./lmstudio.ai/static/js/main*.js | tr '"' '\n' | grep -i appimage && rm -Rf "$HOME"/.cache/lmstudio.ai)
I had to change it because the download page has been changed. This approach will slow down also the update, because the script must download and check an hidden url into a deep js file.
The approach I'd like to have is a file that lists all download urls directly. Github actions can take case of updating these.
No more missing urls this way. We have a snapshot for the apps and people can still have the apps they want, also if not updated... but better than nothing.
Main question is how much from installation scripts differ between apps.
Just theoreticaly: Sort them to (from github, sourceforge, gitlab, gitea, ... , own)
If you still have most apps in own categories we will have problems. If you have just few we are good.
also this is a big issue, use the command
am -d platform-tools brave firefox-nightly nootka node python3.9 coreuniverse inkscape blender kdegames libreoffice wine kdenlive ffwa-whatsapp windows95
they are just samples
No more missing urls this way. We have a snapshot for the apps and people can still have the apps they want, also if not updated... but better than nothing. Not understand this..
Don't undestand this.
CI 👍
Don't undestand this.
The ChaoticAUR, for example, already compiles the packages from AUR and then stores the package in its repos. If a PKGBUILD is broken, ChaoticAUR still have a working URL for the ready binary.
The same way, I'd like to have a snapshot of the existing URLs, all the installation scripts should only rely on the URL listed. So what happened at https://github.com/ivan-hc/AM/issues/539 should no more happen, at least to have a package, and not nothing.
am -d platform-tools brave firefox-nightly nootka node python3.9 coreuniverse inkscape blender kdegames libreoffice wine kdenlive ffwa-whatsapp windows95
they are just samples
Problem?
run the command, download these scripts and compare them
EDIT: are you afraid to use the -d option because on your fork something bay be broken? Better! You would understand more how I feel when I take tests. xD
run the command, download these scripts and compare them
EDIT: are you afraid to use the -d option because on your fork something bay be broken? Better! You would understand more how I feel when I take tests. xD
OK, I was right.
@zen0bit R.I.P.
Here somewhere shoould be mention of am website