681837 - xpcshell Test Failures on latest s-c merge and m-c

Status: RESOLVED INVALID

(Firefox :: Sync, defect)

Description

[Gregory Szorc [:gps]]

Attachment: services/sync/tests xpcshell test log

I'm seeing lots of test failures on the latest s-c merge and on current m-c tip.

s-c version:

changeset:   75738:947fe0ebad68
tag:         tip
parent:      75737:fc9a011616cc
parent:      75667:51fb40284211
user:        Allison Naaktgeboren <[e-mail]>
date:        Tue Aug 23 14:44:11 2011 -0700
description:
Merge mozilla-central into services-central

m-c version:
changeset:   75845:e58e98a89827
tag:         tip
parent:      75817:b354d9b3e9e1
parent:      75844:427a6b2313db
user:        Ed Morley <[e-mail]>
date:        Thu Aug 25 01:09:05 2011 +0100
summary:     Merge last green changeset of mozilla-inbound to mozilla-central

.mozconfig:
mk_add_options MOZ_OBJDIR=@TOPSRCDIR@/obj-ff-debug
mk_add_options MOZ_MAKE_FLAGS="-j8"
ac_add_options --enable-application=browser
ac_add_options --enable-tests
ac_add_options --with-ccache=/usr/bin/ccache
ac_add_options --enable-optimize=-O2

xpcshell test output from `make -C services/sync/tests xpcshell-tests` is attached.

For a general `make xpcshell-tests`, I'm seeing a few scattered test failures and longer-than-expected execution times. A number of tests seem to pause for a while before resuming. When I compare times to tinderbox logs, they are way out of whack (by an order of magnitude). I find it hard to believe the Tinderbox machines are that much faster than my i7-2600K!

I have a build without '-O2' churning now. I'll report if that is any different.

Comment 1

[Gregory Szorc [:gps]]

I forgot to mention this is on a Linux 2.6.38 64-bit virtual machine. I had no problems with this build configuration until a recent pull. No clue when the previous time I pulled was.

Comment 2

[Gregory Szorc [:gps]]

I removed the '--enable-optimize=-O2' line from my .mozconfig and I still see the same behavior. Confirmed that gcc is being invoked with '-Os'

Here is an example compile command:

/usr/bin/ccache c++ -o SQLFunctions.o -c -I../../../dist/stl_wrappers -I../../../dist/system_wrappers -include /home/gps/src/mozilla-central/config/gcc_hidden.h -DSQLITE_DEFAULT_PAGE_SIZE=32768 -DXPCOM_TRANSLATE_NSGM_ENTRY_POINT=1 -DMOZILLA_INTERNAL_API -D_IMPL_NS_COM -DEXPORT_XPT_API -DEXPORT_XPTC_API -D_IMPL_NS_GFX -D_IMPL_NS_WIDGET -DIMPL_XREAPI -DIMPL_NS_NET -DIMPL_THEBES  -DSTATIC_EXPORTABLE_JS_API -DOSTYPE=\"Linux2.6\" -DOSARCH=Linux -DEXCLUDE_SKIA_DEPENDENCIES  -DOS_LINUX=1 -DOS_POSIX=1  -I/home/gps/src/mozilla-central/toolkit/components/places/../build -I/home/gps/src/mozilla-central/ipc/chromium/src -I/home/gps/src/mozilla-central/ipc/glue -I../../../ipc/ipdl/_ipdlheaders  -I/home/gps/src/mozilla-central/toolkit/components/places -I. -I../../../dist/include -I../../../dist/include/nsprpub  -I/home/gps/src/mozilla-central/obj-ff-debug/dist/include/nspr -I/home/gps/src/mozilla-central/obj-ff-debug/dist/include/nss       -fPIC -fno-rtti -fno-exceptions -Wall -Wpointer-arith -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wcast-align -Wno-invalid-offsetof -Wno-variadic-macros -Werror=return-type -Wno-long-long -fno-strict-aliasing -std=gnu++0x -pthread -ffunction-sections -fdata-sections -pipe -DNDEBUG -DTRIMMED -g -fno-omit-frame-pointer -Os -freorder-blocks -finline-limit=50 -fomit-frame-pointer   -DMOZILLA_CLIENT -include ../../../mozilla-config.h -MD -MF .deps/SQLFunctions.pp /home/gps/src/mozilla-central/toolkit/components/places/SQLFunctions.cpp

Comment 3

[Philipp von Weitershausen [:philikon]]

We resolved this via IRC. Greg was running a server on port 8080 which broke all tests that use nsHttpServer.

Comment 4

[Richard Newman [:rnewman]]

See also Bug 681840, which causes us to print a helpful warning in this situation, to avoid further wasted time :D

Comment 5

[Gregory Szorc [:gps]]

To be clear, stopping hte server on port 8080 did NOT fix the issue. It was merely one of the issues preventing tests from running.

That being said, I just created a fresh Ubuntu 11.04 64-bit VM and compiled a fresh pull of mozilla-central. And, things run fine!

On my original, problematic VM, I did a pull to latest m-c and the tests still have issues. I have no clue what is going on, but it appears it is specific to my VM. So, the INVALID resolution makes sense to me.