Disable mod_http2 and mod_ssl_ct for prefork builds, since the former

shouldn't be used under prefork and the latter isn't tested at all.
Possibly related to infrequent prefork child segfaults under pool-debug.

e.g. https://travis-ci.org/github/apache/httpd/jobs/736044109
is a multi-threaded prefork child dying with both mod_h2 and mod_ssl_ct
threads active.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@1882548 13f79535-47bb-0310-9956-ffa450edef68

diff --git a/.travis.yml b/.travis.yml
index 16febaaf05f55e8f965ca71b53c1764c5ce9c3e9..d682cca4f9f0a750e81164017cdbb3788164163b 100644 (file)
--- a/.travis.yml
+++ b/.travis.yml
@@ -240,7 +240,7 @@ jobs:
     - name: Linux Ubuntu, Shared MPMs (prefork), pool-debug, SSL/TLS variants
       env: APR_VERSION=1.7.0 APR_CONFIG="--enable-pool-debug"
            APU_VERSION=1.6.1 APU_CONFIG="--with-crypto --with-ldap"
-           CONFIG="--enable-mods-shared=reallyall --enable-mpms-shared=all --with-mpm=prefork"
+           CONFIG="--enable-mods-shared=reallyall --enable-mpms-shared=all --with-mpm=prefork --disable-http2 --disable-ssl-ct"
            TEST_MALLOC=1 TEST_SSL=1
     # -------------------------------------------------------------------------
     - name: Linux Ubuntu, APR 1.7.0, APR-util 1.6.1, pool-debug, LDAP

diff --git a/test/README.travis b/test/README.travis
index 01a5bfcee49d7e59af5d629aa52d2c1a289e4c33..5d41e120e70862368ae3accb051c80cd0dd3832d 100644 (file)
--- a/test/README.travis
+++ b/test/README.travis
@@ -60,8 +60,29 @@ TODO list
 * Windows build
 * clang-on-Linux build
 * Use containers for non-Ubuntu-based Linux testing
-* Known test failures
- - "apt-get install" timeout/fails - workaround by moving apt install to before_script phase?
+
+Known Failures
+--------------
+
+Some known failures:
+
+* "apt-get install" timeout/fails - workaround by moving apt install
+  to before_script phase?
+
+* prefork, and more rarely non-prefork testing sometimes catches child
+  segfaults under pool-debug from assert()s in (e.g.)
+  __pthread_tpp_change_priority where one child is destroying threads
+  which another is waiting for, or iterating through via
+  apr_pool_walk_tree().
+
+  See dev@httpd threads:
+  msg <5f4abde1b5789_13fde2ecacb40795a1@travis-tasks-5b566d48fc-drkb9.mail>
+  msg <73060f7b-df7f-ad3c-a9fa-dd666a59b31e@kippdata.de> and
+  https://bz.apache.org/bugzilla/show_bug.cgi?id=63098
+  https://bz.apache.org/bugzilla/show_bug.cgi?id=46185
+
+  Not clear if there is a real bug here which can be reproduced
+  outside of pool-debug.
 
 Testing from a Feature Branch
 -----------------------------