Rails 6.0 introduces parallel testing and the default degree of parallelism is configured based on the number of CPU.
refer https://github.com/rails/rails/pull/34735https://github.com/rails/rails/pull/31900
When any minitest executed under the OS where 2 or more CPU available,
SQLite 3 database files are not git-ignored.
Also updated other files which ignores SQLite database files.
* Steps to reproduce
```
$ git clone https://github.com/yahonda/rep_ignore_sqlite3_databases.git
$ cd rep_ignore_sqlite3_databases/
$ bin/rails test
$ git status
```
* Expected behavior:
- No `Untracked files:`
* Actual behavior:
- SQLite 3 database files appeared
```
$ git status
On branch master
Untracked files:
(use "git add <file>..." to include in what will be committed)
db/test.sqlite3-0
db/test.sqlite3-1
db/test.sqlite3-2
db/test.sqlite3-3
db/test.sqlite3-4
db/test.sqlite3-5
nothing added to commit but untracked files present (use "git add" to track)
$
```
* Global ignores at toplevel .gitignore
* Component-specific ignores in each toplevel directory
* Remove `actionview/test/tmp/.keep` for JRuby
```
rm actionview/test/tmp/ -fr
cd actionview/
bundle exec jruby -Itest test/template/digestor_test.rb
```
Related to #11743, #30392.
Closes#29978.
This allows activestorage users to ship smaller javascript bundles to
visitors using modern browsers, as demonstrated in this repository:
https://github.com/rmacklin/activestorage-es2015-build-example
In that example, the bundle shrinks by 5K (24%).
In addition to allowing smaller bundles for those who ship untranspiled
code to modern browsers, including the source code in the published
package can be useful in other ways:
1. Users can import individual modules rather than the whole library
2. As a result of (1), users can also monkey patch parts of
activestorage by importing the relevant module, modifying the
exported object, and then importing the rest of activestorage (which
would then use the patched object).
Note:
In order to allow the source code to be depended on rather than the
compiled code, we have to declare the external dependency on spark-md5
as a regular dependency, not a development dependency.
This means that even users who depend on the compiled code will have to
download this package. However, spark-md5 is a small package, so this
tradeoff seems worth it.
