dsw / oink-stack Goto Github PK

Receive a segfault when running elkhound/elkhound elkhound/examples/crash1.gr on Mac OS X. Have built the tools using Apple LLVM version 8.0.0 (clang-800.0.38). Debegging segfault reveals the cause to be dereferencing a null pointer when calling LocString::<< operator (loc.str is NULL). From the comments in ast/locstr.h I see that the loc.str is permitted to be NULL:

  // experimenting with allowing 'str' to be null, which is convenient
  // when the string table isn't available

In that case, the << operators should deal gracefully with str==NULL.

Are there samples for using Cqual++ with openSSL ? Apache ? Others ?

Is there a chance of getting a compiled binary posted on github as a release? I got a segfault in Expression::hasUnparenthesizedGT; and after the things I did force compilation that may very well be my own fault.

g++ -c -o agramlex.yy.o agramlex.yy.cc -g -Wall -Wno-deprecated -D__UNIX__ -DNDEBUG -D__LINUX__ -I../smbase
In file included from ../smbase/sm_flexlexer.h:28,
from agramlex.yy.cc:247:
/usr/local/include/FlexLexer.h:130: error: expected unqualified-id before numeric constant
agramlex.yy.cc: In member function ‘virtual int GrammarLexer::yylex()’:
agramlex.yy.cc:640: error: ‘yy_current_buffer’ was not declared in this scope
agramlex.yy.cc:1087: error: ‘yy_current_buffer’ was not declared in this scope
agramlex.yy.cc:1110: error: ‘yy_current_buffer’ was not declared in this scope

my g++ is: 4.4.0
flex 2.5.4
bison 2.3

When running the ./configure script at the top level, the script died when it went into the smbase directory with the following error:

Can't locate sm_config.pm in @INC (you may need to install the sm_config module) (@INC contains: /usr/lib/perl5/site_perl/5.26.1/x86_64-linux-thread-multi /usr/lib/perl5/site_perl/5.26.1 /usr/lib/perl5/vendor_perl/5.26.1/x86_64-linux-thread-multi /usr/lib/perl5/vendor_perl/5.26.1 /usr/lib/perl5/5.26.1/x86_64-linux-thread-multi /usr/lib/perl5/5.26.1 /usr/lib/perl5/site_perl) at ./configure.pl line 6.
configure failed at ./configure.pl line 128.

For some reason, the smbase/configure.pl:6 require sm_config; line was not able to find the sm_config.pm file in the current directory. My guess is I need to update some config setting to make the @INC path see the local directory? (I freely admit that I don't know what I am doing).

I was able to get the entire configure script to work by changing the line in smbase/configure.pl:6 to do "./sm_config.pm";, but that may be undesired. I've never worked with perl before, so I am not sure if the do statement is bad practice or not.

Hello,

I get this error in a STL header:

In state 296, I expected one of these tokens:
  (, ), [, ], ->, ., +, -, ++, --, &, *, .*, ->*, /, %, <<, >>, <, <=, >, >=, ==, !=, ^, |, &&, ||, ?, :, =, *=, /=, %=, +=, -=, &=, ^=, |=, <<=, >>=, ,, ..., ;, }, __attribute__, <?, >?, 
/usr/include/c++/4.6/bits/stl_algobase.h:731:32: Parse error (state 296) at <name>: __niter

I've reduced the input with delta to this:

namespace std __attribute__ ((__visibility__ ("default"))) {
  template<class _Sp, class _Tp> struct __traitor {};
  template<typename _Tp> struct __is_integer {};
  template<typename _Tp> struct __is_floating {};
  template<typename _Tp> struct __is_pointer {};
  template<typename _Tp> struct __is_arithmetic : public __traitor<__is_integer<_Tp>, __is_floating<_Tp> > {};
  template<typename _Tp> struct __is_scalar : public __traitor<__is_arithmetic<_Tp>, __is_pointer<_Tp> > {};
}
namespace __gnu_cxx __attribute__ ((__visibility__ ("default"))) {
  template<bool, typename> struct __enable_if {};
}
namespace std __attribute__ ((__visibility__ ("default"))) {
  template<typename _OutputIterator, typename _Size, typename _Tp> inline typename __gnu_cxx::__enable_if<!__is_scalar<_Tp>::__value, _OutputIterator>::__type __fill_n_a(_OutputIterator __first, _Size __n, const _Tp& __value) {
    for (__decltype(__n + 0) __niter = __n; __niter > 0; --__niter, ++__first)
      *__first = __value;
  }

I'm using gcc 4.6.3 and oink master with the lattice config on an clean Ubuntu 12.04. The header comes from the libstdc++6-4.6-dev package.

Edit: Ubuntu 12.04 instead of 14.04

When compiling oink on Mac OS X using Apple LLVM version 8.0.0 (clang-800.0.38) the following warning is seen:

gcc -c -o regions.o -I.. -iquote  -DNMEMDEBUG -DSIZEOF_VOIDP=4 -Wall -g -O2 regions.c
In file included from regions.c:288:
../libregion/pages.c:270:3: warning: implicit declaration of function 'set_region_range' is invalid in C99 [-Wimplicit-function-declaration]
  set_region_range(newp, (char *)newp + request_bytes, FREEPAGE);
  ^

In working with the latest code, both packaging for Gentoo and testing the analysis tools on an actual work task, I had a couple of things to report and a question. See here for the package build setup, basically some bash functions to run the package build commands and install things (the latter part is all manual). http://www.gentoogeek.org/portage-overlay/dev-util/oink-stack/

Issue 1)
Building as a Gentoo package using "./configure perform" everything builds fine with a current toolchain (gcc-4.5.2 plus the tools below) but the regression tests fail with an assert error:

x86_64-pc-linux-gnu-g++ -o triv/aSEb.gr.exe -DGRAMMAR_NAME=\"triv/aSEb.bin\" -g -Wall -Wno-deprecated -D__UNIX__ -DUSE_ENDSOURCELOC=1 -O2 -DNDEBUG -D__LINUX__ -I. -I../smbase -I../ast -Ic -fno-strict-aliasing triv/aSEb.gr.gen.o c/cc_lang.o c/parssppt.o c/lexer1.o c/lexer1yy.o c/lexer2.o trivmain.cc trivlex.o libelkhound.a -g -Wall -Werror ../ast/libast.a ../smbase/libsmbase.a
rm triv/aSEb.gr.gen.o
make[2]: Leaving directory `/var/tmp/portage/dev-util/oink-stack-9999/work/oink-stack-9999/elkhound'
------------ [1] triv/aSEb.gr.exe triv/aSEb.in1 ------------
Exception thrown: Assertion failed: !isStatic(base), file ../smbase/srcloc.h line 369
Assertion failed: !isStatic(base), file ../smbase/srcloc.h line 369

[1] A regression test command failed:
  triv/aSEb.gr.exe triv/aSEb.in1
make[1]: *** [check] Error 2
make[1]: Leaving directory `/var/tmp/portage/dev-util/oink-stack-9999/work/oink-stack-9999/elkhound'
make: *** [check-rec] Error 1

System uname: Linux-2.6.36.3-rt-x86_64-AMD
sys-libs/glibc: 2.11.3
sys-devel/binutils: 2.21
sys-devel/gcc: 3.4.6-r2, 4.4.4-r2, 4.5.2
sys-devel/libtool: 2.4-r1
sys-devel/make: 3.81-r2

The same assert error is induced when scanning a particular source file with qual.

Issue 2)
Your docs say platform requires gcc 3.4, so I tried 3.4.6 since that's the only 3.4.x version in the Gentoo package tree. So, in building the platform stuff with gcc-3.4.6 there is a missing symbol error from an older gcc libc++:

$ make
cd libc_glibc-2.3.5 && make all
make[1]: Entering directory `oink-stack-9999/platform-model/libc_glibc-2.3.5'
cd src && make all
make[2]: Entering directory `oink-stack-9999/platform-model/libc_glibc-2.3.5/src'
-- aio.c
      gcc       aio.c           -> aio.o
      cpp       aio.c           -> aio.oink.i
      oinkx     aio.oink.i      -> aio.odir
oink/oink: /usr/lib/gcc/x86_64-pc-linux-gnu/3.4.6/libstdc++.so.6: version `GLIBCXX_3.4.11' not found (required by oink-stack-9999/oink/oink)
oink/oink: /usr/lib/gcc/x86_64-pc-linux-gnu/3.4.6/libstdc++.so.6: version `GLIBCXX_3.4.9' not found (required by oink-stack-9999/oink/oink)
make[2]: *** [aio.odir] Error 1
rm aio.oink.i
make[2]: Leaving directory `oink-stack-9999/platform-model/libc_glibc-2.3.5/src'
make[1]: *** [all] Error 2
make[1]: Leaving directory `oink-stack-9999/platform-model/libc_glibc-2.3.5'

It turns out those symbols are only found in dbg_util.o as shown below:

$ readelf -s ../oink/oink | grep 'GLIBCXX_3\.4\.9' | c++filt | grep _ZN
    50: 0000000000000000     0 FUNC    GLOBAL DEFAULT  UND _ZNSo9_M_insertIPKvEERSoT@GLIBCXX_3.4.9 (6)

$ nm -A *.o | grep "_ZNSo9_M_insertIPKvEERSoT"
dbg_util.o:                 U _ZNSo9_M_insertIPKvEERSoT_

So my question is...

Did you have any plans to update the above source file so it does not depend on such a narrow version of the C++ library implementation? If so, that should make it possible to package it (relatively) normally, otherwise it seems like I'd need to build it in a chroot with a very specific toolchain just so I can use it (and I would not have an easy way to package it to use the platform-model). Although shared libs/executables are preferred, I would at least be able to package it without the platform stuff.

Thanks in advance again

I tried to compile a (big, c++) project, and got that PARSING error in a standard C++ header:

In state 1138, I expected one of these tokens:
try, {,
/usr/include/c++/4.6/x86_64-linux-gnu/bits/c++config.h:154:1: Parse error (state 1138) at {

I think there is some inconsistencies between the #define in ast and elkhound

Commit 3d1b8bc introduces conflicting changes in definition and usage of yylex macro in ./elkhound/grampar.tab.cc

$ ./configure
Can't use 'defined(@array)' (Maybe you should just omit the defined()?) at ./configure.pl line 109.

It seems to be a problem with recent Perl versions. kentcdodds/cloc#1

I've just did a conversion of https://github.com/dsw/oink-stack/blob/master/elsa/cc.gr to an EBNF that can be used to generate a nice navigable railroad diagram and it can be viewd here https://mingodad.github.io/cpp-grammars/elsa-cc.gr.ebnf.xhtml , the EBNF is here https://github.com/mingodad/cpp-grammars/blob/main/elsa-cc.gr.ebnf .

Thee you'll find EBNF and railroad diagrams for other C++ grammars.

When compiling on Mac OS X using Apple LLVM version 8.0.0 (clang-800.0.38) the following error is seen

gcc -c -o regions.o -I.. -iquote  -DNMEMDEBUG -DSIZEOF_VOIDP=4 -Wall -g -O2 regions.c
In file included from regions.c:288:
../libregion/pages.c:272:56: error: use of undeclared identifier 'zeroregion'; did you mean 'newregion'?
  set_region_range(newp, (char *)newp + request_bytes, FREEPAGE);