Re: [PATCH 2/5] Save trace into CTF format

[Joel Brobecker <brobecker at adacore dot com>]

> 2013-03-03  Hui Zhu  <hui_zhu@mentor.com>
> 	    Yao Qi  <yao@codesourcery.com>
> 
> 	* Makefile.in (REMOTE_OBS): Add ctf.o.
> 	(SFILES): Add ctf.c.
> 	(HFILES_NO_SRCDIR): Add ctf.h.
> 	* ctf.c, ctf.h: New files.
> 	* tracepoint.c : Include 'ctf.h'.
> 	(collect_pseudocommand): Remove static.
> 	(trace_save_command): Parse option "-ctf".
> 	Produce different trace file writers per option.
> 	Adjust output message.
> 	(trace_save_tfile, trace_save_ctf): New.
> 	* tracepoint.h (trace_save_tfile, trace_save_ctf): Declare.
> 	* mi/mi-main.c: Include 'ctf.h'.
> 	(mi_cmd_trace_save): Handle option '-ctf'.  Call either
> 	trace_save_tfile or trace_save_ctf.
[...]
> +  ctf_save_fwrite_format (handler->metadata_fd, metadata_fmt,
> +			  CTF_SAVE_MAJOR, CTF_SAVE_MINOR,
> +			  BYTE_ORDER == LITTLE_ENDIAN ? "le" : "be");

This is causing a build failure on ppc-aix, for instance, due to:

    /[...]ctf.c:223:28: error: 'BYTE_ORDER' undeclared (first use in this function)
    /[...]ctf.c:223:28: note: each undeclared identifier is reported only once for each function it appears in
    /[...]ctf.c:223:42: error: 'LITTLE_ENDIAN' undeclared (first use in this function)

I don't know how to fix this, because I am not sure whether host
or target endianness were meant. I asssume from the code that it
was host, but not sure how the host endianess would make sense...

What also worries me is that I don't seem to see approval for commit
for this patch. I know both of you are very careful, so perhaps in
a separate message not linked to this sub-thread?

-- 
Joel