619661 - js source fails to build with --enable-dtrace

Status: RESOLVED DUPLICATE of bug 574403

(Core :: General, defect)

Description

[Matěj Cepl]

Attachment: suggested patch

Description of problem:
When trying to build xulrunner with dtrace enabled (I have the latest systemtap
and systemtap-sdt-devel packages installed on my recently updated i686 rawhide
VM), I get the following errors in js sources: 

In file included from jsdtracef.c:47:
jsdtracef.h:36:30: error: javascript-trace.h: No such file or directory
jsdtracef.c: In function 'jsdtrace_function_entry':
jsdtracef.c:209: warning: implicit declaration of function
'JAVASCRIPT_FUNCTION_ENTRY'
jsdtracef.c: In function 'jsdtrace_function_info':
jsdtracef.c:220: warning: implicit declaration of function
'JAVASCRIPT_FUNCTION_INFO'
jsdtracef.c: In function 'jsdtrace_function_args':
jsdtracef.c:233: warning: implicit declaration of function
'JAVASCRIPT_FUNCTION_ARGS'
jsdtracef.c: In function 'jsdtrace_function_rval':
jsdtracef.c:249: warning: implicit declaration of function
'JAVASCRIPT_FUNCTION_RVAL'
jsdtracef.c: In function 'jsdtrace_function_return':
jsdtracef.c:261: warning: implicit declaration of function
'JAVASCRIPT_FUNCTION_RETURN'
jsdtracef.c: In function 'jsdtrace_object_create_start':
jsdtracef.c:271: warning: implicit declaration of function
'JAVASCRIPT_OBJECT_CREATE_START'
jsdtracef.c: In function 'jsdtrace_object_create_done':
jsdtracef.c:277: warning: implicit declaration of function
'JAVASCRIPT_OBJECT_CREATE_DONE'
jsdtracef.c: In function 'jsdtrace_object_create':
jsdtracef.c:283: warning: implicit declaration of function
'JAVASCRIPT_OBJECT_CREATE'
jsdtracef.c:286: error: 'uintptr_t' undeclared (first use in this function)
jsdtracef.c:286: error: (Each undeclared identifier is reported only once
jsdtracef.c:286: error: for each function it appears in.)
jsdtracef.c:286: error: expected ')' before 'obj'
jsdtracef.c: In function 'jsdtrace_object_finalize':
jsdtracef.c:299: warning: implicit declaration of function
'JAVASCRIPT_OBJECT_FINALIZE'
jsdtracef.c:299: error: 'uintptr_t' undeclared (first use in this function)
jsdtracef.c:299: error: expected ')' before 'obj'


I hand built a javascript-trace.h in the same fashion as the makefile in
js/src/ does and tried to build again. The missing file error for
javascript-trace.h and related warnings were resolved but I still encounter the
uintptr_t related errors at jsdtracef.c:286 and jsdtracef.c 299

Version-Release number of selected component (if applicable):
xulrunner-1.9.1-0.9.beta2.fc11

How reproducible:
Every time

Steps to Reproduce:
1. Add a 'ac_add_options --enable-drace' line to xulrunner-mozconfig 
2. Use rpmbuild to build from sources

Actual results:
The error messages reported above

Expected results:
xulrunner builds

Additional comments:

The js Makefile links the dtrace object into the .so but not the .a.  This
patch adjusts for that:

sed -i '/include.*rules.mk/a\
ifdef HAVE_DTRACE\
PROGOBJS += ../mozjs-dtrace.$(OBJ_SUFFIX)\
endif\
' js/src/shell/Makefile
sed -i '/include.*rules.mk/a\
ifdef HAVE_DTRACE\
PROGOBJS += ../mozjs-dtrace.$(OBJ_SUFFIX)\
endif\
' js/src/jsapi-tests/Makefile
I was building upstream 1.9.1.4 (it is easier to automate since I can just wget
the source).  That links with an explicit object list:
g++ -o js  -fno-rtti -fno-exceptions -Wall -Wpointer-arith -Woverloaded-virtual
-Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wcast-align
-Wno-invalid-offsetof -Wno-long-long -pedantic -g
-I/work/scox/systemtap/bld/../install/include -fno-strict-aliasing
-fshort-wchar -pthread -pipe  -DNDEBUG -DTRIMMED -Os -fstrict-aliasing   js.o
jsdtracef.o jsapi.o jsarena.o jsarray.o jsatom.o jsbool.o jscntxt.o jsdate.o
jsdbgapi.o jsdhash.o jsdtoa.o jsemit.o jsexn.o jsfun.o jsgc.o jshash.o
jsinterp.o jsinvoke.o jsiter.o jslock.o jslog2.o jsmath.o jsnum.o jsobj.o
json.o jsopcode.o jsparse.o jsprf.o jsregexp.o jsscan.o jsscope.o jsscript.o
jsstr.o jsutil.o jsxdrapi.o jsxml.o prmjtime.o mozjs-dtrace.o    -lpthread  
-Wl,-rpath-link,/bin -Wl,-rpath-link,/lib  -L./../../dist/bin
-L./../../dist/lib -L/work/scox/systemtap/bld/testsuite/xul/bld/dist/lib
-lplds4 -lplc4 -lnspr4 -lpthread -ldl -ldl -lm  -lm -ldl

versus 1.9.2.1 which links with:
c++ -o js  -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 -Wno-long-long -pedantic -O2 -g
-pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector
--param=ssp-buffer-size=4 -m64 -mtune=generic -fno-strict-aliasing -pthread
-pipe  -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno-reorder-functions   js.o   
-lpthread -Wl,-rpath,/usr/lib64/xulrunner-1.9.2  -Wl,-rpath-link,/bin
-Wl,-rpath-link,/usr/lib  -L../../../dist/bin -L../../../dist/lib -L/usr/lib64
-lplds4 -lplc4 -lnspr4 -lpthread -ldl ../editline/libeditline.a
../libjs_static.a -ldl -lm

I'm switching to 1.9.2.1 and I'll upgrade the patch.
Created attachment 394365 [details] [diff] [review]
log of the build attempt
Created attachment 394364 [details] [diff] [review]
patch on spec file

-----------------------------
I have a patch for building xulrunner-1.9.2.3-1 with --enable-dtrace.  I had to
workaround a bug in the released sdt.h DTRACE_PROBE10 macro; missing cast.
 SOURCES/mozjs-dtrace.patch New
 xulrunner-mozconfig.patch  Patch to SOURCES/xulrunner-mozconfig
 xulrunner.spec.patch       Patch to SPECS/xulrunner.spec
Just switching on NEEDINFO, so we know we are waiting here.
Please file the patch at upstream bugzilla (bugzilla.mozilla.org).
Reassigning to component owner, as Stan is not in a position to commit the
makefile patch that appears to correct this problem.
systemtap uses the patch in #7 to successfully build and test xul 1.9.2rc1 as
part of its testsuite

Comment 1

[Matěj Cepl]

Build with XULRunner 2.0 from Fedora Rawhide fails
http://koji.fedoraproject.org/koji/taskinfo?taskID=2766561

Comment 2

[Steve Fink [:sfink] [:s:]]

I'm going to dupe this to bug 574403 because I seem to have this building successfully with both dtrace on OSX and systemtap on Linux. I remember needing to fix up DTRACE_PROBE10 as well, but I newer systemtaps seem to have corrected that problem.

I'm not doing everything you did in your patches, but I'd rather new bugs be filed for anything remaining.