[Mono-dev] 2.4RC3 build problems (and code problems with the build)

[Jonathan Pobst]

You can compare the warnings to a Linux build to see which are something 
unique to ppc:

http://mono.ximian.com/monobuild/builds/HEAD/sles-10-x86_64/mono/129882/logs/build.log

Jay reports the same warnings on Linux as well.

Jonathan


Paul wrote:
> Hi,
> 
> I'm a tad worried about the build problems on mono. Looking at the
> throwback and comments coming from the compiler, I'm seeing lots of
> these sorts of problems which indicates bison problems and quite a lot
> of warnings while compiling the likes of boehm-gc.c (gc_priv.h line 1931
> - warning function declaration isn't a prototype - lots of these) and
> also a bit later on
> 
> ../doltcompile gcc -DHAVE_CONFIG_H -I. -I.. -I/usr/include/glib-2.0
> -I/usr/lib/glib-2.0/include    -DGC_LINUX_THREADS -D_GNU_SOURCE
> -D_REENTRANT -DUSE_MMAP -DUSE_MUNMAP -D_FILE_OFFSET_BITS=64
> -D__mono_ppc__ -DUSE_COMPILER_TLS  -O2 -g -pipe -Wall
> -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector
> --param=ssp-buffer-size=4 -m32 -fno-strict-aliasing -fno-strict-aliasing
> -Wdeclaration-after-statement -g -Wall -Wunused -Wmissing-prototypes
> -Wmissing-declarations -Wstrict-prototypes  -Wmissing-prototypes
> -Wnested-externs -Wpointer-arith -Wno-cast-qual -Wcast-align
> -Wwrite-strings -MT jni.lo -MD -MP -MF .deps/jni.Tpo -c -o jni.lo jni.c
> jni.c:490: warning: function declaration isn't a prototype
> jni.c:494: warning: no previous prototype for 'ikvm_MarshalDelegate'
> jni.c:501: warning: no previous prototype for 'ikvm_CallOnLoad'
> mv -f .deps/jni.Tpo .deps/jni.Plo
> 
> a load more of the same sort of warnings in macros.c, serial.c,
> zlib_macros.c and minizip/zip.c
> 
> and finally before System.Xml.dll throws a big style wobbler,
> 
> make[8]: Entering directory
> `/builddir/build/BUILD/mono-2.4/mcs/class/System.XML'
> ./../../jay/jay -ct < ./../../jay/skeleton.cs System.Xml.XPath/Parser.jay >System.Xml.XPath/Parser.cs
> ./../../jay/jay: 21 rules never reduced
> ./../../jay/jay: 1 shift/reduce conflict, 42 reduce/reduce conflicts.
> sed "s/\%start Expr/\%start Pattern/" System.Xml.XPath/Parser.jay >Mono.Xml.Xsl/PatternParser.jay
> echo "#define XSLT_PATTERN" > Mono.Xml.Xsl/PatternParser.cs
> ./../../jay/jay -ct Mono.Xml.Xsl/PatternParser.jay < ./../../jay/skeleton.cs >>Mono.Xml.Xsl/PatternParser.cs
> ./../../jay/jay: 3 rules never reduced
> ./../../jay/jay: 1 shift/reduce conflict, 46 reduce/reduce conflicts.
> 
> These conflicts are the most worrying and is more than likely the cause
> for the ppc build to fail.
> 
> I'm not sure if this is a bison style problem but it is worrying that
> the likes of not checking a return value for fwrite when building mcs to
> start with is still being overlooked.
> 
> I would suggest that before 2.4 hits the roads properly, these issues
> are carefully looked at (that is unless I'm needlessly worrying, but
> given that the likes of the gcc people stop when these problems occur
> suggests otherwise).
> 
> TTFN
> 
> Paul
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Mono-devel-list mailing list
> Mono-devel-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-devel-list