[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: warnings in win32 builds from trunk
From: |
Corwin Brust |
Subject: |
Re: warnings in win32 builds from trunk |
Date: |
Sun, 3 Apr 2022 11:30:47 -0500 |
On Sun, Apr 3, 2022 at 1:17 AM Eli Zaretskii <eliz@gnu.org> wrote:
>
> > From: Corwin Brust <corwin@bru.st>
> > Date: Sat, 2 Apr 2022 18:33:27 -0500
> >
> > Attempting to build a snapshot from the master branch today I noticed
> > some warnings. Here they are; please let me know if there's else I
> > might helpfully do to troubleshoot/debug these (e.g. open a bug
> > report).
>
> Please file a bug report for each one of them.
Now done (except for the one you indicated should be ignored).
Replying here also in case that's helpful for posterity; see detail below.
>
> > In file included from
> > C:/Users/corwi/emacs-build/git/master/src/process.c:33:
> > C:/Users/corwi/emacs-build/git/master/src/process.c: In function
> > 'Fmake_process':
> > C:/Users/corwi/emacs-build/git/master/src/lisp.h:1649:31: warning:
> > null pointer dereference [-Wnull-dereference]
> > 1649 | return XSTRING (string)->u.s.data;
> > | ~~~~~~~~~~~~~~~~~~~~~^~~~~
>
> Line 1649 of lisp.h is inside SDATA. But since Fmake_process doesn't
> call SDATA directly, I don't know which string object is this about.
> (Really, the compiler should do a better job when reporting such
> issues, by showing all the relevant source locations.)
bug#54695: 29.0.50; w32notify.o Fmake_process warning building on win32
>
> > C:/Users/corwi/emacs-build/git/master/src/w32menu.c: In function
> > 'set_frame_menubar':
> > C:/Users/corwi/emacs-build/git/master/src/w32menu.c:324:9: warning:
> > 'memcpy' offset [3, 10] from the object at '<unknown>' is out of the
> > bounds of referenced subobject 'contents' with type 'struct Lisp_X
> > *[]' at offset 3 [-Warray-bounds]
> > 324 | memcpy (previous_items, XVECTOR
> > (f->menu_bar_vector)->contents,
> > |
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > 325 | previous_menu_items_used * word_size);
> > | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > In file included from
> > C:/Users/corwi/emacs-build/git/master/src/w32menu.c:26:
> > C:/Users/corwi/emacs-build/git/master/src/lisp.h:1725:17: note:
> > subobject 'contents' declared here
> > 1725 | Lisp_Object contents[FLEXIBLE_ARRAY_MEMBER];
> > | ^~~~~~~~
>
> This is some general problem with FLEXIBLE_ARRAY_MEMBER, not specific
> to w32. Sounds like latest compilers don't like that. Are you
> building with custom compiler switches, per chance?
bug#54694: 29.0.50; w32menu.o set_frame_menubar warning building on win32
I don't think I am building with any custom compiler switches.
Configured using:
'configure --without-dbus --with-native-compilation
--without-compress-install -C 'CFLAGS=-O2 -g3''
Configured features:
ACL GIF GMP GNUTLS HARFBUZZ JPEG JSON LCMS2 LIBXML2 MODULES NATIVE_COMP
NOTIFY W32NOTIFY PDUMPER PNG RSVG SOUND SQLITE3 THREADS TIFF
TOOLKIT_SCROLL_BARS WEBP XPM ZLIB
>
> > C:/Users/corwi/emacs-build/git/master/src/w32heap.c: In function
> > 'getrlimit':
> > C:/Users/corwi/emacs-build/git/master/src/w32heap.c:853:14: warning:
> > 'm' may be used uninitialized [-Wmaybe-uninitialized]
> > 853 | if (!VirtualQuery ((LPCVOID) &m, &m, sizeof m))
> > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> This is a bogus warning, ignore it. The compiler doesn't understand
> the semantics of VirtualQuery.
So I didn't create a bug report for this one. Please LMK if it's
helpful for me to do so.
>
> Thanks.