Intro

cblog is web software, the CMS engine in particular. It is aimed at providing content creators, system administrators, and businesses with an easy way to quickly make a website with their content. Despite its name, you can make any CMS-oriented site, like a news site.

The engine is aimed at:

1. Use as few resources as possible.
2. Provide a lightning-fast response time.
3. have the possibility to run on low-power machines (cheap VPS or even MCUs)
4. Making it possible to deploy in a few minutes without additional technical knowledge (currently WIP)
5. Hosting multiple websites on one engine instance

Current state of the project: under active development, a preview version (0.1) with basic functionality is available.

Build the app

First of all you need to download dependencies.

git clone https://github.com/xdevelnet/ssb.git
git clone https://github.com/xdevelnet/md4c

If you want to make project run via fastcgi, you would need nginx and fastcgi:

sudo apt-get install libfcgi-dev nginx

If you want to make project run via embedded web server mongoose:

sudo apt-get install libev-dev
git clone https://github.com/cesanta/mongoose.git

Now download project itself and build it. Replace obj with fcgi or mon

git clone https://github.com/xdevelnet/cblog
cd cblog
make obj

Done! Now you can run your app via freshly created binary.

Use external tools/software to control process.

Feature availability tables

Application features:

Feature	Status	Comment
application abstraction	Done	application code is independent from web server or any kind of layer
supported backend engines	Done	Mongoose: allows to compile app and web server to single binary. Fastcgi (nginx): Requires web server configuration
Multiple template support	Done, but only one template is enabled	Abstracted in essb library that were written particularly for this project. Two templates were already done
Multiple L10n support	Done, but not enabled at all	Abstracted in tssb library that were written particularly for this project
Markdown pages	Done	When user makes new blog record, app may expect markdown format. If no html provided it will automatically transform markdown into html
Tags support	Mostly done	Tags are displaying on top of each record, engine can show records that are filtered by tag, during adding new record new tag could be added if it's not exist
CI Pre-built packages, Docker images	Planned	CI done only for subprojects (ssb)
Tests	Partially done	Done only for one module, multiple hand tests are available
Pages	Done	Title page with displaying multiple records, filter page with displaying multiple records by specific filter (currently tag filter is supported) with length limiter, single record page with displaying tags, 404 page. 500 page, user login page, logout, user panel page, adding new record page
User management	Done	User can login and logout. Each user stores it's id, name, email, hashed password (called "credentials"), user creating time, approve code for email sending, status, time values depending on status. Making new users currently can be done manually
User management pages	Partially done	Only login/logout is available. User panel is for preview only. Register page and administrator management are planned features.
RBACL (Role-Based Access Control Lists)	Planned	Some internal parts are available
Sessions	Done, currently used mostly for storing user sessions

Supported data storage:

Engine	Status
Files	Done
MySQL	WIP
SQLite	Planned
Single file/ embedded storage	Planned

Platform support

Platform	Status
Linux	Done
FreeBSD	Done
POSIX	Not tested but will probably work
MCUs (like ESP32)	Planned
Windows	Small preview is planned

External tools/software/references:

Tool/software	Description
Nginx	Web server that were mainly used with fastcgi library. https://nginx.org/
Fastcgi	Interface that allows to move out from CGI and achieve great performance. Unfortunately, original site is down, only archive left: https://web.archive.org/web/20060221235908/http://fastcgi.com/ Active development is here: https://github.com/FastCGI-Archives/fcgi2
Mongoose	Web server that could be embedded in application. The easiest way to demonstrate app capabilities
ssb (essb, tssb)	library that allows to parse special formatted ssb data in order to have a handy use in this project: https://github.com/xdevelnet/ssb
template2essb	application that transforms html templates to (e)ssb format: https://github.com/xdevelnet/ssb
tcsv2tssb	converter from csv data to (t)ssb format. Mainly used to make text translations. https://github.com/xdevelnet/tcsv2tssb

More

Why I am doing this?

1. Modern websites are fat and ugly. They are using too many resources on both sides - on server and on client. In most cases such amount of resources usage is not justified. If you have small business, need single-page website (business card website), or even if you just want a place where you write your thoughts sometimes - there is no reason to host whole relational database just to serve it.
2. I don't like complexity that people have deal with during work with such software. By complexity, I mean not only personal rejection/hostility interpreted programming languages, but effects that they cause, specifically improving for sake of improving, which leads to unnecessary clutter and complexity.
3. As consequence of points 1 and 2, I would like to give people ability to run their web apps on very weak devices (by weak I mean not "bad" but "not powerful"): starting from small VPS (on 128mb RAM WordPress isn't working well), ending by specialized MCU (plugged in PoE ethernet and god ready to use service) and embedded into other apps (like Android).

Pros and cons

Pros are marked with green color, cons - red. Points that can't be categorized by pros/cons are marked with blue.

1. Resources. 🟩 Using small amount of resources, especially on some configurations
2. Compatibility. 🟩 Much better compatibility with all possible devices - because of small resource usage it can be embedded into anything
3. Development speed. 🟥️ Very slow. Developing such apps gives no right to make a mistake, therefore developer should deal carefully with that. Wherein, additional techniques to prove reliability and faultlessness/correctness are mandatory.
4. Reliability. 🟦 In theory, even during complicated OS states, such software will continue to function, because in most cases running process already prepared all data and did all preparation operations. So, after serving each request, such data will be not freed, and continue to exist in order to serve further requests. Hardly speaking, if OS is capable to evaluate standard Berkely sockets operations (socket()/send()/recv()/poll()/close()), the application will probably continue to work.
5. Security. 🟦 Despite the fact that compiled software is well known for dangerous vulnerabilities because of developers mistakes, vulnerabilities on websites that are made by using modern tools/ways aren't much better. The whole tragedy is that because of complexity and bloatwareness that's between "received request" and "send a response", such flaws are still exists and found every year, just look at php and WordPress.

I also have something to say!

1. I just want to take this opportunity to make a web a little bit better. I am accounting that someone will use this software and as an author I will have DIRECT influence to websites quality.
2. In order to read few text paragraphs and look at small image, I don't want to wait for 3 seconds until page loads on my PC (8 seconds on my flagship Android phone of 2020: Samsung S20), then pressing "Close" button in the corner of frame that just JUMPS ON YOU, hiding all the content and asks for subscribing to VERY IMPORTANT BREAKING NEWS, and then again I need to close another frame that asks me for a cookies. SITE DON'T NEED TO SET ANY COOKIES ON MY BROWSER JUST TO SHOW ME SOME INFORMATION. Thank you. Looking at all of this I just want to open https://motherfuckingwebsite.com/ and enjoy my life being.
3. There would be bundled templates with my app, which servers enough compatibility, but not using much client's resources. In other words, it would be lightweight html without javascript, or just very small amount of javascript which is not required to site to run website.
4. I just want to cull/take away web-monkeys that are making a mess/hell/sodomy in modern web by high entry threshold into C programming language and related technologies. No offence dear colleagues, but business pushes on market, and demand on crap creates supply.
5. I have a mental trauma after php (while I was freelancing at the dawn of a career), which I attempted to cure with perl (during work on PortaOne Inc.), then I finished myself by using Python (during making some help to friends with their projects). These programming languages themselves aren't bad, I am not their hater... Issues that they made and the ways that are solved are just so ugly and causes my rejection. After billions of useless frameworks I just spat and left into Embedded.