Build your code, only if you need to...
Configuration
Install this package in your project using npm install -D build-if-needed
of yarn add --dev build-if-needed
.
Add the section build-if-needed
in your package.json
. For each of your build commands, specify the list of files from your src
folder and the files from your dist
folder.
Add this to the package.json
...
{
"name": "your-package",
"build-if-needed": {
"build": {
"input": [
"src/**/*.*"
],
"output": [
"dist/**/*.*"
]
},
"build:prod": {
"input": [
"src/**/*.*"
],
"output": [
"dist/**/*.*"
]
}
},
"scripts" {
"build": "...",
"build:if-needed": "build-if-needed --script build",
"build:prod": "...",
"build:prod:if-needed": "build-if-needed --script build:prod",
...
}
}
Add this to the .gitignore
...
node_modules
.build-if-needed
etc.
Then, instead of typing npm run build
to build your code, you can type npm run build:if-needed
.
The first time that you execute npm run build:if-needed
, the command will take just as long as npm run build
. But the second time, it should complete instantly; unless you have edited one of the source files.
How does it work?
This utility will calculate and stores a MD5 hash of all the files within your src
folder and your dist
folder. The hash values are stored inside a .build-if-needed
folder. So the second time you execute npm run build:if-needed
, this utility can workout if the source code has changed and if the application needs to be rebuilt.
- Before the build, we calculate the hash values for the
src
and dist
folders. Then compare them with the last values stored. - If the hash values have changed; it must mean that the source code was edited, or the
dist
folder was tempered with. In that case, the build is executed. - After the build, we recaclulate the hash values for the
src
and dist
folders and store them for the next execution.
Why?
I created this utility for large applications that take a while to compile and where the --watch
command of tsc
, babel
or webpack
cannot be used.
This is expecially useful on a large lerna
monorepo with multiple packages. Hitting npm run build
after going a git pull
was too time consuming. So npm run build:if-needed
becomes really useful to only rebuild the packages that need to be rebuilt.
Monorepo
If you work on a large lerna
monorepo, you may consider adding to the input
a set of your package dependencies.
Given the monorepo below...
packages/
package1/
src/
package.json
package2/
src/
package.json
package.json
lerna.json
Let's consider that package1
has a dependency on package2
. So you may want to add the following to package1/package.json
. This way, this utility knows that package1
needs to be rebuilt when package2
is modified.
"input": [
"src/**/*.**",
"./node_modules/package2/.build-if-needed/*.*" <-- Do this to indicate that package1 has a dependency on package2.
],
Do not scan the whole node_modules
folder. This would scan too many files, and the MD5 hash generation would take too long.
"input": [
"src/**/*.**",
"./node_modules/**/*.*" <-- DO NOT DO THIS!!
],
Command line options
$ build-if-needed --help
Usage: build-if-needed [options]
Options:
-n, --script <script> name of the script configuration
-d, --debug outputs debugging information
-h, --help display help for command
The debug
option will output in the .build-if-needed
folder the list of files that were scanned for updates. This can be useful to debug your configuration.
Tips
To speed up the calculation of the MD5 hash, try to reduce the number of files that build-if-needed
have to scan.
"input": [
"src/**/*.**",
"!**/*.map" <-- Exclude from the MD5 calculation files that are not essential to the compilation of the project.
],
Configuration options
Using the command
parameter, you can configure whether you prefer your build commands to be executed using yarn (e.g. yarn run build
) or npm (e.g. npm run build
).
{
"name": "your-package",
"build-if-needed": {
...
"command": "yarn"|"npm" (Default: "npm")
},
"scripts" { ... }
}