You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Dec 25, 2018. It is now read-only.
The contenthash is unreliable for cache breaking because it's calculated on the source css, so if the source html changes such that the purifier includes different content into the build, then the contenthash remains the same despite the fact that the file contents are now actually different.
As you can see that adding additional markup to the html causes the purifier to not exclude the respective style, but the resulting contenthash for the build remains main.a4b8c75c90e070022e83e3602ea0b64b.css
If possible the contenthash from the ExtractTextPlugin should be re-calculated for the build product. At the very least this should be documented in the README, because I think as it stands the usage of the contenthash is somewhat misleading as it's generally relied upon for cache breaking.
The text was updated successfully, but these errors were encountered:
I can't recreate it, but can confirm that while the size of files changes, [hash] isn't.
And I can't isolate if it's this module or webpack itself (feeling a bit overwhelmed with all the conventions being thrown around to give a more useful report), but think it is.
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
purifycss-webpack: 0.7.0
purify-css: 1.2.5
node v7.5.0
npm 4.1.2
The contenthash is unreliable for cache breaking because it's calculated on the source css, so if the source html changes such that the purifier includes different content into the build, then the contenthash remains the same despite the fact that the file contents are now actually different.
Here is a repository to demonstrate the issue.
As you can see that adding additional markup to the html causes the purifier to not exclude the respective style, but the resulting contenthash for the build remains
main.a4b8c75c90e070022e83e3602ea0b64b.css
If possible the contenthash from the ExtractTextPlugin should be re-calculated for the build product. At the very least this should be documented in the README, because I think as it stands the usage of the contenthash is somewhat misleading as it's generally relied upon for cache breaking.
The text was updated successfully, but these errors were encountered: