GCC 9: special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul #29

srakitnican · 2019-01-22T13:51:49Z

This happens with GCC9, on cbang commit fdd8867

g++ -o build/cbang/enum/CPUExtendedFeature.o -c -std=c++11 -g -Wall -Werror -I/usr/include/v8-3.14/ -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_S
OURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rp
m/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -Wno-error=parentheses -Wno-deprecated-de
clarations -fPIC -DDEBUG -D_REENTRANT -DHAVE_EXPAT -DHAVE_PTHREADS -DHAVE_LIBSQLITE -DHAVE_V8 -DDEBUG_LEVEL=1 -DUSING_CBANG -Iinclude -Isrc -Isrc/boost src/cba
ng/enum/CPUExtendedFeature.cpp
g++ -o build/cbang/enum/CPUFeature.o -c -std=c++11 -g -Wall -Werror -I/usr/include/v8-3.14/ -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2
-Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat
/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -Wno-error=parentheses -Wno-deprecated-declaratio
ns -fPIC -DDEBUG -D_REENTRANT -DHAVE_EXPAT -DHAVE_PTHREADS -DHAVE_LIBSQLITE -DHAVE_V8 -DDEBUG_LEVEL=1 -DUSING_CBANG -Iinclude -Isrc -Isrc/boost src/cbang/enum/
CPUFeature.cpp
In file included from src/boost/boost/date_time/date_facet.hpp:21,
                 from src/boost/boost/date_time/gregorian/gregorian_io.hpp:16,
                 from src/boost/boost/date_time/gregorian/gregorian.hpp:31,
                 from src/boost/boost/date_time/posix_time/time_formatters.hpp:12,
                 from src/boost/boost/date_time/posix_time/posix_time.hpp:24,
                 from src/cbang/time/Time.cpp:42:
src/boost/boost/date_time/special_values_formatter.hpp: In constructor 'boost::date_time::time_facet<time_type, CharT, OutItrT>::time_facet(size_t) [with time_
type = boost::posix_time::ptime; CharT = char; OutItrT = std::ostreambuf_iterator<char, std::char_traits<char> >]':
src/boost/boost/date_time/special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul [-Werror=stringop-overflow=]
   43 |       std::copy(&default_special_value_names[0],
      |       ~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   44 |                 &default_special_value_names[3],
      |                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   45 |                 std::back_inserter(m_special_value_names));
      |                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
src/boost/boost/date_time/special_values_formatter.hpp:89:70: note: referenced argument declared here
   89 |   const typename special_values_formatter<CharT, OutItrT>::char_type special_values_formatter<CharT, OutItrT>::default_special_value_names[3][17] = {
      |                                                                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The text was updated successfully, but these errors were encountered:

jcoffland · 2019-01-24T20:34:52Z

That's an issue with the built-in boost library. Boost is not yet testing with gcc 9.

You could add -Wno-stringop-overflow to your build scripts.

srakitnican · 2019-01-25T13:42:43Z

Hmm Fedora compiles boost with GCC9[1], I don't see Fedora using this option [2] or any patches to fix this. Using the suggested compiler option makes it compile without any further errors, though.

Build log: https://kojipkgs.fedoraproject.org//packages/boost/1.69.0/2.fc30/data/logs/x86_64/build.log

[1] https://koji.fedoraproject.org/koji/buildinfo?buildID=1181814
[2] https://src.fedoraproject.org/cgit/rpms/boost.git/tree/boost.spec#n769

jcoffland · 2019-01-25T19:00:57Z

I've updated the built-in boost to 1.69.0. Perhaps that works without -Wno-stringop-overflow.

srakitnican · 2019-01-28T13:35:39Z

Seems like new version requires additional header I had to install.

g++ -o build/boost/libs/iostreams/src/lzma.o -c -std=c++11 -g -I/usr/include/v8-3.14/ -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -Wno-error=parentheses -Wno-deprecated-declarations -fPIC -DDEBUG -D_REENTRANT -DHAVE_EXPAT -DHAVE_PTHREADS -DHAVE_LIBSQLITE -DHAVE_V8 -DDEBUG_LEVEL=1 -DUSING_CBANG -Iinclude -Isrc -Isrc/boost src/boost/libs/iostreams/src/lzma.cpp
src/boost/libs/iostreams/src/lzma.cpp:15:10: fatal error: lzma.h: No such file or directory
   15 | #include <lzma.h>
      |          ^~~~~~~~
compilation terminated.
scons: *** [build/boost/libs/iostreams/src/lzma.o] Error 1
scons: building terminated because of errors.

After I've installed that still the same error showed up.

g++ -o build/cbang/enum/CPUFeature80000001.o -c -std=c++11 -g -Wall -Werror -I/usr/include/v8-3.14/ -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_S
OURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rp
m/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -Wno-error=parentheses -Wno-deprecated-de
clarations -fPIC -DDEBUG -D_REENTRANT -DHAVE_EXPAT -DHAVE_PTHREADS -DHAVE_LIBSQLITE -DHAVE_V8 -DDEBUG_LEVEL=1 -DUSING_CBANG -Iinclude -Isrc -Isrc/boost src/cba
ng/enum/CPUFeature80000001.cpp
In file included from src/boost/boost/date_time/date_facet.hpp:21,
                 from src/boost/boost/date_time/gregorian/gregorian_io.hpp:16,
                 from src/boost/boost/date_time/gregorian/gregorian.hpp:31,
                 from src/boost/boost/date_time/posix_time/time_formatters.hpp:12,
                 from src/boost/boost/date_time/posix_time/posix_time.hpp:24,
                 from src/cbang/time/Time.cpp:42:
src/boost/boost/date_time/special_values_formatter.hpp: In constructor 'boost::date_time::time_facet<time_type, CharT, OutItrT>::time_facet(size_t) [with time_
type = boost::posix_time::ptime; CharT = char; OutItrT = std::ostreambuf_iterator<char, std::char_traits<char> >]':
src/boost/boost/date_time/special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul [-Werror=stringop-overflow=]
   43 |       std::copy(&default_special_value_names[0],
      |       ~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   44 |                 &default_special_value_names[3],
      |                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   45 |                 std::back_inserter(m_special_value_names));
      |                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
src/boost/boost/date_time/special_values_formatter.hpp:89:70: note: referenced argument declared here
   89 |   const typename special_values_formatter<CharT, OutItrT>::char_type special_values_formatter<CharT, OutItrT>::default_special_value_names[3][17] = {
      |                                                                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

jcoffland · 2019-02-01T10:15:35Z

I fixed the lzma.h error.

You noted that Fedora does not add -Wno-stringop-overflow. I can see this is true from the log you linked above, but the log also shows this warning:
libs/log/src/trivial.cpp:77:53: warning: 'strlen' argument missing terminating nul [-Wstringop-overflow=]

A different place but, it's the same error. Building the boost package does not actually build much of boost. Boost is mostly C++ header files that get compiled only when used from other programs.

So Fedora has the same problem and it's not fixed there either.

srakitnican · 2019-02-03T17:49:26Z

Hmm turns out -Werror is not a standard Fedora compiler flag, turns out this becomes an issue due to me setting the SCons debug flag long time ago. I remember setting it because the resulted binary was not in the wanted format, maybe debuginfo stripped or lacking. Anyway without -Werror cbang compiles fine. Sorry for the noise. CAMotics still fails but that is another issue.

jcoffland · 2019-02-04T07:08:29Z

Yes, I use -Werror in debug mode to make sure I fix any compiler warnings.

…iling error on gcc9 Add "-Wno-stringop-overflow" option to avoid the following error: orch.cpp: In member function 'ref_resolve_status Orch::resolveFieldRefArray(type_map&, const string&, swss::KeyOpFieldsValuesTuple&, std::vector<long unsigned int>&, std::string&)': orch.cpp:609:41: error: 'strlen' argument missing terminating nul [-Werror=stringop-overflow=] 609 | object_name_list += string(&list_item_delimiter); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from orch.cpp:6: orch.h:25:12: note: referenced argument declared here 25 | const char list_item_delimiter = ','; | ^~~~~~~~~~~~~~~~~~~ cc1plus: all warnings being treated as errors According to CauldronDevelopmentLLC/cbang#29, this is an error from boost and can be suppressed by using this option Signed-off-by: Stephen Sun <[email protected]>

jcoffland closed this as completed Sep 3, 2019

stephenxs mentioned this issue Jan 27, 2021

Fix issue #5157 by identifying the dependency among objects and avoiding releasing an object still being referenced sonic-net/sonic-swss#1440

Merged

This was referenced Jan 27, 2021

Add diagnose option "-Wno-stringop-overflow" option to avoid the compiling errors stephenxs/sonic-swss#12

Closed

Fix the compiling errors in gcc9 sonic-net/sonic-swss#1621

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

GCC 9: special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul #29

GCC 9: special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul #29

srakitnican commented Jan 22, 2019 •

edited

Loading

jcoffland commented Jan 24, 2019

srakitnican commented Jan 25, 2019

jcoffland commented Jan 25, 2019

srakitnican commented Jan 28, 2019

jcoffland commented Feb 1, 2019

srakitnican commented Feb 3, 2019 •

edited

Loading

jcoffland commented Feb 4, 2019

GCC 9: special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul #29

GCC 9: special_values_formatter.hpp:43:16: error: 'strlen' argument missing terminating nul #29

Comments

srakitnican commented Jan 22, 2019 • edited Loading

jcoffland commented Jan 24, 2019

srakitnican commented Jan 25, 2019

jcoffland commented Jan 25, 2019

srakitnican commented Jan 28, 2019

jcoffland commented Feb 1, 2019

srakitnican commented Feb 3, 2019 • edited Loading

jcoffland commented Feb 4, 2019

srakitnican commented Jan 22, 2019 •

edited

Loading

srakitnican commented Feb 3, 2019 •

edited

Loading