From IRC conversation with Leonard today:
- the apache bug has not been fixed. our version of mod_compress still modifies etags on the way out but not on the way in
- [...] we should not yet turn mod_compress back on in the api vhost
- https://issues.apache.org/bugzilla/show_bug.cgi?id=39727#c31 has roy fielding explaining why the "just strip off -gzip on the way in" fix doesn't work, and that's probably why the fix was never applied
- roy says "The best solution is to implement transfer-encoding as an http protocol filter module", so that intermediaries can understand it and pass it along
We'll discuss next steps soon. They are not clear yet.
From IRC conversation with Leonard today:
- the apache bug has not been fixed. our version of mod_compress still modifies etags on the way out but not on the way in
- [...] we should not yet turn mod_compress back on in the api vhost
- https:/ /issues. apache. org/bugzilla/ show_bug. cgi?id= 39727#c31 has roy fielding explaining why the "just strip off -gzip on the way in" fix doesn't work, and that's probably why the fix was never applied
- roy says "The best solution is to implement transfer-encoding as an http protocol filter module", so that intermediaries can understand it and pass it along
We'll discuss next steps soon. They are not clear yet.