NodeJs http SSI Server

Installation

npm install nodejs_http_ssi_server

(please only use the newest version, i.e. 1.1.0 and up)

Usage example

const server = require("nodejs_http_ssi_server");
server(port, root);

Defaults are (port = 3000, root = "/www").

Note that the default root is relative to the "node_modules" folder, so it assumes a "www" directory in the same parent directory as the "node_modules" directory. If the relative path of "/www" does not exist, it defaults to the relative path of "/".

.shtml SSI include directives (SGML comments) have the following format:

<!--#include file="path" --> 

i.e. only a single file is parsed per directive. A path can be specified with double, single or without quotes. That being the exception, the parsing is quite strict otherwise. The path must be relative and contained in the directory of the current .shtml document that is being parsed. Additionally, the pattern "../" is not allowed in the path for simplicity reasons.

Important Note

For simplicity reasons, we assume .shtml and directive attribute files to be utf-8 encoded.

Description

This minimal (server side rendering for static html) webserver is based on express.js and provides a simplified variant of the Server Side Include (SSI) "include" directive. That is, .html files are served as usual, while .shtml files are parsed for directives as described above. If, for any reason, a .shtml file cannot be properly parsed, it is served as a .html file.

Technical Details

The implementation is based on node.js and express.js. This allows us to only care about the middleware layer, i.e. fetching /*.shtml pages to parse for the SSI "include" directive and replace it with the specified file attribute given as path. The decision was made to use synchronous functions for file I/O, as we need to wait for reading .shtml files and the attribute files to replace the "include" directive (if any) anyway, while using asynchronous file I/O would have resulted in unnecessary nesting. However, this point can be improved upon in the future. ~~While the parsing loop does its job sufficiently quickly, it could absolutely be refactored into something more simple, making use of (potentially functional) common JavaScript language constructs.~~ (Edit The initial parsing loop bothered me enough to do a quick refactor using regex parsing) Another low-hanging fruit would be to (at least partially) generalise the directive parser for future extensions. Lastly, the logging is very rudimentary and can be generalised as well.

Statement

Apart from very common lookup of minimal code fragments, no significant amount of code has been copied from anywhere. Except for the imported libraries, the source code seen in this repository is written by me.