hdf5-18 1.8.6_1 build failure

[astrofitz (Michael Fitzgerald)]

I am encountering a build failure. This occurs while trying to build the file H5dbg.o, with the first error message "H5TSprivate.h:42: error: expected specifier-qualifier-list before 'CRITICAL_SECTION'" (see attached log).

[astrofitz (Michael Fitzgerald)]

build log

[mamoll (Mark Moll)]

mmoll-main.log

[mamoll (Mark Moll)]

It works for me. There are small differences in how files are compiled. The H5dbg.c is compiled like so for you:

:info:build libtool: compile:  
/usr/bin/gcc-4.2 -DHAVE_CONFIG_H -I. -DNDEBUG -UH5_DEBUG_API -I/opt/local/include -std=c99 -pedantic -Wall -Wextra -Wundef -Wshadow -Wpointer-arith -Wbad-function-cast -Wcast-qual -Wcast-align -Wwrite-strings -Wconversion -Waggregate-return -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations -Wredundant-decls -Wnested-externs -Winline -Wno-long-long -Wfloat-equal -Wmissing-format-attribute -Wmissing-noreturn -Wpacked -Wdisabled-optimization -Wformat=2 -Wunreachable-code -Wendif-labels -Wdeclaration-after-statement -Wold-style-definition -Winvalid-pch -Wvariadic-macros -Wnonnull -Winit-self -Wmissing-include-dirs -Wswitch-default -Wswitch-enum -Wunused-macros -Wunsafe-loop-optimizations -Wc++-compat -Wvolatile-register-var -Wstrict-overflow -O3 -fomit-frame-pointer -finline-functions -pipe -O2 -arch x86_64 -MT H5dbg.lo -MD -MP -MF .deps/H5dbg.Tpo -c H5dbg.c  -fno-common -DPIC -o .libs/H5dbg.o

On my machine (OS X 10.6.6), it compiles like so:

:info:build /bin/sh ../libtool --tag=CC   --mode=compile 
/usr/bin/gcc-4.2 -DHAVE_CONFIG_H -I.  -DNDEBUG -UH5_DEBUG_API -I/opt/local/include -std=c99 -pedantic -Wall -Wextra -Wundef -Wshadow -Wpointer-arith -Wbad-function-cast -Wcast-qual -Wcast-align -Wwrite-strings -Wconversion -Waggregate-return -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations -Wredundant-decls -Wnested-externs -Winline -Wno-long-long -Wfloat-equal -Wmissing-format-attribute -Wmissing-noreturn -Wpacked -Wdisabled-optimization -Wformat=2 -Wunreachable-code -Wendif-labels -Wdeclaration-after-statement -Wold-style-definition -Winvalid-pch -Wvariadic-macros -Wnonnull -Winit-self -Wmissing-include-dirs -Wswitch-default -Wswitch-enum -Wunused-macros -Wunsafe-loop-optimizations -Wc++-compat -Wvolatile-register-var -Wstrict-overflow -O3 -fomit-frame-pointer -finline-functions -pipe -O2 -arch x86_64 -MT H5Fdbg.lo -MD -MP -MF .deps/H5Fdbg.Tpo -c -o H5Fdbg.lo H5Fdbg.c

What could be different between your setup and mine?

[jmroot (Joshua Root)]

Everything before the build phase is missing from the reporter's log so it's hard to know. Does the build system perhaps re-libtoolize itself?

[mamoll (Mark Moll)]

Maybe. How do I force this (if this is what it should be doing) or disable it (if it causing the compilation to fail)?

[mamoll (Mark Moll)]

It looks like the reporter had the threadsafe variant enabled. With this variant enabled, compilation fails for me, too, but in a different file with a different error. The release notes at ​http://www.hdfgroup.org/ftp/HDF5/current/src/hdf5-1.8.6-RELEASE.txt say that the thread-safe version is not supported on OS X. I'll remove this variant and close the ticket. If the reporter has this problem without the threadsafe variant, I'll reopen the ticket.

[astrofitz (Michael Fitzgerald)]

I was indeed able to successfully build the port with the threadsafe option disabled. It may be splitting hairs, but the release notes suggest the threadsafe option on OSX is merely untested, rather than unsupported. Perhaps at this point it is an upstream problem.