Mercurial > hg > nginx-site

comparison text/en/CHANGES @ 2862:cf7551842617

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
nginx-1.23.0
author Maxim Dounin <mdounin@mdounin.ru>
date Tue, 21 Jun 2022 19:07:59 +0300
parents 1ecf0e0526da
children 9383e934e546
comparison
equal deleted inserted replaced
2861:53734009b0c3 2862:cf7551842617
1
2 Changes with nginx 1.23.0 21 Jun 2022
3
4 *) Change in internal API: now header lines are represented as linked
5 lists.
6
7 *) Change: now nginx combines arbitrary header lines with identical
8 names when sending to FastCGI, SCGI, and uwsgi backends, in the
9 $r->header_in() method of the ngx_http_perl_module, and during lookup
10 of the "$http_...", "$sent_http_...", "$sent_trailer_...",
11 "$upstream_http_...", and "$upstream_trailer_..." variables.
12
13 *) Bugfix: if there were multiple "Vary" header lines in the backend
14 response, nginx only used the last of them when caching.
15
16 *) Bugfix: if there were multiple "WWW-Authenticate" header lines in the
17 backend response and errors with code 401 were intercepted or the
18 "auth_request" directive was used, nginx only sent the first of the
19 header lines to the client.
20
21 *) Change: the logging level of the "application data after close
22 notify" SSL errors has been lowered from "crit" to "info".
23
24 *) Bugfix: connections might hang if nginx was built on Linux 2.6.17 or
25 newer, but was used on systems without EPOLLRDHUP support, notably
26 with epoll emulation layers; the bug had appeared in 1.17.5.
27 Thanks to Marcus Ball.
28
29 *) Bugfix: nginx did not cache the response if the "Expires" response
30 header line disabled caching, but following "Cache-Control" header
31 line enabled caching.
32
1 33
2 Changes with nginx 1.21.6 25 Jan 2022 34 Changes with nginx 1.21.6 25 Jan 2022
3 35
4 *) Bugfix: when using EPOLLEXCLUSIVE on Linux client connections were 36 *) Bugfix: when using EPOLLEXCLUSIVE on Linux client connections were
5 unevenly distributed among worker processes. 37 unevenly distributed among worker processes.