Willy, Am 16.05.19 um 11:12 schrieb Willy Tarreau: >> Both frontends will do the correct replacement, but IMO the reqrep one >> is more readable (not that any of these are really readable): > > I'm having a different view on this. The one using the headers hack is > horrible while the reqrep one is simply ugly :-)
I'm fine with that view :-) > Really, this is a good reason for having a "replace-uri" action. It should > mostly be a copy-paste of the replace-header code but applied to the URI, > and will be much simpler and safer to use than any of the tricks above. Yes, definitely. >> The obvious `http-request set-path %[path,regsub(...)]` as suggested in >> the docs for `http-request set-query` does *NOT* work, because the >> `regsub` parameters cannot contain the closing parenthesis required for >> capture groups. > > Yep I see. I've added feature request #100 to keep track of this. I'm Perfect. > willing to merge this even late into 2.0 and to make sure we can warn > on rsp/req*. Just let me know if you're interested. Also feel free to Generally yes, but I'll have to see whether I have time to dive into that before the release. So if someone beats me to it: Even better for me (or rather everyone). > add replace-path to replace only the path component as I suspect it could > be useful as well. I guess for completeness it makes sense, but it can easily be emulated with replace-uri: http-request replace-uri ([^?]*)(?(.*))? /\1/whatever/\2 So the important one is replace-uri, but I guess that it's not too hard to add that one as well. > Oh BTW I've merged your 6 patches into the "next" branch and merged two > of them into master. > Perfect. Best regards Tim Düsterhus
