[Webkit-unassigned] [Bug 174252] Speedometer 2.0: Document es2015-babel-webpack build process

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Sun Jul 9 02:28:59 PDT 2017


https://bugs.webkit.org/show_bug.cgi?id=174252

--- Comment #8 from Mathias Bynens <mathias at qiwi.be> ---
The (In reply to Ryosuke Niwa from comment #7)
> (In reply to Mathias Bynens from comment #6)
> > Everything in `dist/` is the result of running the now-documented build
> > steps. You can reproduce these versions locally by following the build steps.
> 
> What I'm asking is rather how your new build step is different from what we
> had earlier.
> 
> If there is any configuration change, I'd have to manually run the test and
> verify each test step is functioning correctly.

The configuration for this test is defined in `package.json` and `webpack.config.babel.js`. Of those two files, this patch only changes `package.json` by removing some unnecessary entries.

In `package.json`, the `dependencies`/`devDependencies` used for building `dist/*` are pinned to specific versions (e.g. `1.2.3` instead of `~1.2.3` or `^1.2.3`). This patch doesn’t change those versions.

So, the configuration has not changed, and assuming these build steps match how the previous `dist/*` output was generated (Addy?), the build process has not changed either.

The only difference is the generated chunk hashes and some variable names in the minified output. Why did these change? The only reason I can think of is that the dependencies of those dependencies might have gotten updates since the last time `dist/*` was created. For example, if webpack v1.0.0 has a dependency `"foo-bar": "^1.0.0"` in its `package.json`, running `npm install webpack at 1.0.0` might install `foo-bar at 1.0.0` at one point in time, but `foo-bar at 1.2.3` a couple of weeks later. I suspect this is what’s causing the slight difference in output. If we want deterministic `npm install`s, we should check in `package-lock.json` files. Would you like me to do that?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170709/62cc7c50/attachment.html>


More information about the webkit-unassigned mailing list