This is the thing that npm uses to read package.json files. It validates some stuff, and loads some default things.
It keeps a cache of the files you've read, so that you don't end up reading the same package.json file multiple times.
Note that if you just want to see what's literally in the package.json
file, you can usually do var data = require('some-module/package.json')
.
This module is basically only needed by npm, but it's handy to see what npm will see when it looks at your package.
var readJson = require('read-package-json')
// readJson(filename, [logFunction=noop], [strict=false], cb)
readJson('/path/to/package.json', console.error, false, function (er, data) {
if (er) {
console.error("There was an error reading the file")
return
}
console.error('the package data is', data)
});
-
file
{String} The path to the package.json file -
logFn
{Function} Function to handle logging. Defaults to a noop. -
strict
{Boolean} True to enforce SemVer 2.0 version strings, and other strict requirements. -
cb
{Function} Gets called with(er, data)
, as is The Node Way.
Reads the JSON file and does the things.
See man 5 package.json
or npm help json
.
By default this is a reference to the npmlog
module. But if that
module can't be found, then it'll be set to just a dummy thing that does
nothing.
Replace with your own {log,warn,error}
object for fun loggy time.
Run all the extra stuff relative to the file, with the parsed data.
Modifies the data as it does stuff. Calls the cb when it's done.
Array of functions that are called by extras
. Each one receives the
arguments fn(file, data, cb)
and is expected to call cb(er, data)
when done or when an error occurs.
Order is indeterminate, so each function should be completely independent.
Mix and match!
Some other files have an effect on the resulting data object, in the following ways:
If there is a README
or README.*
file present, then npm will attach
a readme
field to the data with the contents of this file.
Owing to the fact that roughly 100% of existing node modules have Markdown README files, it will generally be assumed to be Markdown, regardless of the extension. Please plan accordingly.
If there is a server.js
file, and there is not already a
scripts.start
field, then scripts.start
will be set to node server.js
.
If there is not already a contributors
field, then the contributors
field will be set to the contents of the AUTHORS
file, split by lines,
and parsed.
If a bindings.gyp file exists, and there is not already a
scripts.install
field, then the scripts.install
field will be set to
node-gyp rebuild
.
If the json file does not exist, but there is a index.js
file
present instead, and that file has a package comment, then it will try
to parse the package comment, and use that as the data instead.
A package comment looks like this:
/**package
* { "name": "my-bare-module"
* , "version": "1.2.3"
* , "description": "etc...." }
**/
// or...
/**package
{ "name": "my-bare-module"
, "version": "1.2.3"
, "description": "etc...." }
**/
The important thing is that it starts with /**package
, and ends with
**/
. If the package.json file exists, then the index.js is not
parsed.
If there is not already a man
field defined as an array of files or a
single file, and
there is a directories.man
field defined, then that directory will
be searched for manpages.
Any valid manpages found in that directory will be assigned to the man
array, and installed in the appropriate man directory at package install
time, when installed globally on a Unix system.
If there is not already a bin
field defined as a string filename or a
hash of <name> : <filename>
pairs, then the directories.bin
directory will be searched and all the files within it will be linked as
executables at install time.
When installing locally, npm links bins into node_modules/.bin
, which
is in the PATH
environ when npm runs scripts. When
installing globally, they are linked into {prefix}/bin
, which is
presumably in the PATH
environment variable.
If you do not have a types
field, then it will check if a
corresponding *.d.ts
file exists for your package entry file and add
it to the package.json
.