Otto van der Schaaf

I meant to say we wouldn't want an intermediary cdn/proxy/browser to cache the html, because we can't purge those. Having said that, if you see room for doc improvements, PR's...

Yes, changes have landed that are release-worthy IMHO; I've been trying to find some time to prepare a new RC because I wanted to do this earlier, but I have...

I've worked on producing a release candidate, I'll shoot for getting that review by the incubator PMC this week. I'm not sure how long we'll be going back and forth,...

@Fierralin could you paste the commands you ran? (and their outputs)?

There currently is a VOTE thread in progress for release candidate number 4 on releasing `Apache PageSpeed (incubating) 1.14.26.1`. Hopefully the VOTE will pass, and then we can ship :rocket:!

We don't have any tests with `proxy_protocol` turned on in the configuration, but I would expect the feature to be transparent to our module. So before diving in deeper, it...

Thanks. These lines look relevant: https://github.com/nginx/nginx/blob/master/src/http/v2/ngx_http_v2.c#L686 https://github.com/nginx/nginx/blob/master/src/core/ngx_proxy_protocol.c#L120 Could you try if configuring the module to load files directly from disk avoids these log lines?: https://modpagespeed.com/doc/domains#ModPagespeedLoadFromFile Alternatively, I think that pointing...

That is some really good information. Intuitively I'd think that there is an issue specific to interaction with compression and the combiner(s). One more thing I'd also love to see...

In this scenario, nginx is not expected to play a role in the gzipping. As a workaround, adding `pagespeed HttpCacheCompressionLevel 0;` might work around the issue (as pagespeed will stop...

I suspect that impact on required cache space of the workaround may not be that big: images probably make up the majority of that, and are not subject to compression....