Summary: | mod_ratelimit breaks HEAD requests | ||
---|---|---|---|
Product: | Apache httpd-2 | Reporter: | Jean Weisbuch <jean> |
Component: | mod_ratelimit | Assignee: | Apache HTTPD Bugs Mailing List <bugs> |
Status: | RESOLVED FIXED | ||
Severity: | major | CC: | marcin |
Priority: | P2 | Keywords: | FixedInTrunk |
Version: | 2.4.38 | ||
Target Milestone: | --- | ||
Hardware: | PC | ||
OS: | Linux | ||
Attachments: |
Don't eat EOS in ap_http_header_filter() for single brigade HEAD responses (trunk)
Don't eat EOS in ap_http_header_filter() for single brigade HEAD responses (2.4.x) |
Description
Jean Weisbuch
2019-02-20 17:39:40 UTC
> Since upgrading to 2.4.34
In order to help diagnose the potential regression, which version were you using before the upgrade?
(In reply to Christophe JAILLET from comment #1) > > Since upgrading to 2.4.34 > > In order to help diagnose the potential regression, which version were you > using before the upgrade? It first appeared with 2.4.34 if i recall correctly, still happens on both 2.4.37 and 2.4.38. The only modification that occurred in mod_ratelimit in 2.4.34 is r1835168. I've reverted this commit and the behavior looks correct again (i.e. curl does not exits with "(52) Empty reply from server") Created attachment 36448 [details]
Don't eat EOS in ap_http_header_filter() for single brigade HEAD responses (trunk)
Does this patch work?
Created attachment 36449 [details]
Don't eat EOS in ap_http_header_filter() for single brigade HEAD responses (2.4.x)
This version applies to 2.4.x branch.
(In reply to Yann Ylavic from comment #4) > Created attachment 36448 [details] > Don't eat EOS in ap_http_header_filter() for single brigade HEAD responses > (trunk) > > Does this patch work? The 2.4.x version seems to fix the issue on 2.4.38, thanks. Backported to 2.4.x, will be in upcoming 2.4.39. |