Re: Strange gcc warning...

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Sat, 18 Aug 2018 12:44:42 +0200
    From:        Kamil Rytarowski <n54%gmx.com@localhost>
    Message-ID:  <9a8d0076-72c3-8055-f386-5a3a389f1b4b%gmx.com@localhost>


  | >        fprintf ((FILE), ","HOST_WIDE_INT_PRINT_UNSIGNED",%u\n",  \

  | This is not a strange warning but an invalid C++ code.

If you had read my message you would have seen that that one was
not the one to which the Subject was referring, the other report (that
you ignored in this reply) was - this one I was just having a rant about,
as (it seems to me anyway) that C is adopting more and more from C++
(as strange as that is, given that C++ always was pretty disgusting)
and that if this C++ issue turns into a C issue, it is going to be a real PITA.

  | Here and in other places with reports from compilers, it's better to fix
  | the code rather than papering issues with a pile of -Wno*.

No-one was papering over anything - there is no need, all those warnings
are being ignored anyway (none of them stopped the build - which went on
way past that section, until gcc eventually suicided by attemtping to use too
much RAM - I was building on my test system, which has no configured
swap, and only 1GiB RAM configured - a XEN DomU ... I was not using -j,
it was a purely serial build, and still gcc apparently needs more than
something approaching 1GB of space ... the system was doing nothing
else at the time, aside from the standard daemons like syslogd, init, and
my login shell on the console which was running the build.sh)

It is worth noting also that the errors gcc was complaining about were
in the gcc sources themselves (that is, in what we have in gcc.old, which
I think is where the tools gcc is built from - that one built fine - and which
was then building itself to make the installed compiler - no cross compiles,
this was an amd64 host (running HEAD) compiling the exact same set
of sources it was built from).

But let's ignore all that, let the gcc people sort it out for themselves, and
instead please explain the 2nd warning that gcc issued (still about its
own code).

kre