Iain Buclaw <ibuc...@gdcproject.org> writes:
> On Tue, 16 Oct 2018 at 17:28, Richard Sandiford
> <richard.sandif...@arm.com> wrote:
>>
>> Iain Buclaw <ibuc...@gdcproject.org> writes:
>> > diff --git a/gcc/Makefile.in b/gcc/Makefile.in
>> > index 4b7cec82382..0b2daa320c3 100644
>> > --- a/gcc/Makefile.in
>> > +++ b/gcc/Makefile.in
>> > @@ -2496,6 +2525,7 @@ s-tm-texi: build/genhooks$(build_exeext)
>> > $(srcdir)/doc/tm.texi.in
>> > && ( test $(srcdir)/doc/tm.texi -nt $(srcdir)/target.def \
>> > || test $(srcdir)/doc/tm.texi -nt
>> > $(srcdir)/c-family/c-target.def \
>> > || test $(srcdir)/doc/tm.texi -nt
>> > $(srcdir)/common/common-target.def \
>> > + || test $(srcdir)/doc/tm.texi -nt $(srcdir)/d/d-target.def \
>> > ); then \
>> > echo >&2 ; \
>> > echo You should edit $(srcdir)/doc/tm.texi.in rather than
>> > $(srcdir)/doc/tm.texi . >&2 ; \
>> > [...]
>> > @@ -2784,7 +2815,7 @@ build/genrecog.o : genrecog.c $(RTL_BASE_H)
>> > $(BCONFIG_H) $(SYSTEM_H) \
>> > $(CORETYPES_H) $(GTM_H) errors.h $(READ_MD_H) $(GENSUPPORT_H)
>> > \
>> > $(HASH_TABLE_H) inchash.h
>> > build/genhooks.o : genhooks.c $(TARGET_DEF) $(C_TARGET_DEF) \
>> > - $(COMMON_TARGET_DEF) $(BCONFIG_H) $(SYSTEM_H) errors.h
>> > + $(COMMON_TARGET_DEF) $(D_TARGET_DEF) $(BCONFIG_H) $(SYSTEM_H) errors.h
>> > build/genmddump.o : genmddump.c $(RTL_BASE_H) $(BCONFIG_H) $(SYSTEM_H)
>> > \
>> > $(CORETYPES_H) $(GTM_H) errors.h $(READ_MD_H) $(GENSUPPORT_H)
>> > build/genmatch.o : genmatch.c $(BCONFIG_H) $(SYSTEM_H) \
>>
>> I was initially a bit worried about this because it makes the build
>> depend on the existence of the d/ directory. But it doesn't look
>> like we try to ship separate tarballs for specific source languages
>> any more, so that's probably not a problem.
>>
>> > @@ -10659,6 +10668,22 @@ unloaded. The default is to return false.
>> > Return target-specific mangling context of @var{decl} or @code{NULL_TREE}.
>> > @end deftypefn
>> >
>> > +@node D Language and ABI
>> > +@section D ABI parameters
>> > +@cindex parameters, d abi
>> > +
>> > +@deftypefn {D Target Hook} void TARGET_D_CPU_VERSIONS (void)
>> > +Declare all environmental version identifiers relating to the target CPU
>> > using the function @code{builtin_version}, which takes a string
>> > representing the name of the version. Version identifiers predefined by
>> > this hook apply to all modules and being compiled and imported.
>> > +@end deftypefn
>>
>> "and being"? Does this mean "that are being"?
>>
>> > +@deftypefn {D Target Hook} void TARGET_D_OS_VERSIONS (void)
>> > +Similarly to @code{TARGET_D_CPU_VERSIONS}, but is used for versions
>> > relating to the target operating system.
>> > +@end deftypefn
>> > +
>> > +@deftypefn {D Target Hook} unsigned TARGET_D_CRITSEC_SIZE (void)
>> > +Returns the size of the data structure used by the targeted operating
>> > system for critical sections and monitors. For example, on Microsoft
>> > Windows this would return the @code{sizeof(CRITICAL_SECTION)}, while other
>> > platforms that implement pthreads would return
>> > @code{sizeof(pthread_mutex_t)}.
>> > +@end deftypefn
>>
>> Please reflow the .def so that these fit within 80 chars.
>>
>
> This file is auto-generated, but looking at other entries, some remain
> within 80 chars, others are one line such as this. I'll see if I can
> spot the difference in the tm.texi.in file.
It doesn't come from tm.texi.in but from the *target.def file.
Normally the docstring lines should be 80 chars long there, with each
line ending in \n\
Thanks,
Richard
