lmi
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lmi] Should we update submodules's submodules?


From: Vadim Zeitlin
Subject: Re: [lmi] Should we update submodules's submodules?
Date: Wed, 9 Mar 2022 23:57:44 +0100

On Wed, 23 Feb 2022 20:38:19 +0000 Greg Chicares <gchicares@sbcglobal.net> 
wrote:

GC> With gcc-11, I see apparently-harmless warnings (example below)
GC> for what I believe is a libpng submodule embedded in a wx
GC> submodule. This doesn't seem to cause any actual problem, but
GC> perhaps it's worth mentioning.
GC> 
GC> /opt/lmi/src/lmi/third_party/wx/src/png/png.c: In function 
'wx_png_convert_to_rfc1123_buffer':
GC> /opt/lmi/src/lmi/third_party/wx/src/png/pngprefix.h:136:27: warning: 
'number_buf' may be used uninitialized [-Wmaybe-uninitialized]
GC>   136 | #define png_format_number wx_png_format_number
GC> /opt/lmi/src/lmi/third_party/wx/src/png/png.c:757:69: note: in definition 
of macro 'APPEND_STRING'
GC>   757 | #     define APPEND_STRING(string) pos = png_safecat(out, 29, pos, 
(string))
GC>       |                                                                     
^~~~~~
GC> /opt/lmi/src/lmi/third_party/wx/src/png/pngpriv.h:1758:4: note: in 
expansion of macro 'png_format_number'
GC>  1758 |    png_format_number(buffer, buffer + (sizeof buffer), format, 
number)
GC>       |    ^~~~~~~~~~~~~~~~~
GC> /opt/lmi/src/lmi/third_party/wx/src/png/png.c:759:24: note: in expansion of 
macro 'PNG_FORMAT_NUMBER'
GC>   759 |          APPEND_STRING(PNG_FORMAT_NUMBER(number_buf, format, 
(value)))
GC>       |                        ^~~~~~~~~~~~~~~~~

 Just for the record, I've fixed this warning in wxWidgets in

https://github.com/wxWidgets/wxWidgets/commit/3533decf93a43e208135840e2016906e641301a4

with the real change being in the submodule commit

https://github.com/wxWidgets/libpng/commit/dc6985f9d2d8625ac54717d698cae8af1a53d21b

 I'm pretty sure that the original warning was spurious, i.e. it's just a
compiler bug, but I couldn't produce a simple self-contained test case
reproducing it reasonably quickly, so I didn't report this as a gcc bug and
just added the workaround above, so that the warning won't appear any more
once lmi wx submodule is updated soon.

 Regards,
VZ

Attachment: pgpUVAKcukun6.pgp
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]