OpenCores
URL https://opencores.org/ocsvn/openrisc_2011-10-31/openrisc_2011-10-31/trunk

Subversion Repositories openrisc_2011-10-31

[/] [openrisc/] [trunk/] [gnu-src/] [gcc-4.2.2/] [BUGS] - Diff between revs 38 and 154

Go to most recent revision | Show entire file | Details | Blame | View Log

Rev 38 Rev 154
?rev1line?
?rev2line?
 
 
 
                                   GCC Bugs
 
 
 
   The   latest   version   of  this  document  is  always  available  at
 
   [1]http://gcc.gnu.org/bugs.html.
 
     _________________________________________________________________
 
 
 
Table of Contents
 
 
 
     * [2]Reporting Bugs
 
          + [3]What we need
 
          + [4]What we DON'T want
 
          + [5]Where to post it
 
          + [6]Detailed bug reporting instructions
 
          + [7]Detailed bug reporting instructions for GNAT
 
          + [8]Detailed bug reporting instructions when using a precompiled
 
            header
 
     * [9]Frequently Reported Bugs in GCC
 
          + [10]C++
 
               o [11]Missing features
 
               o [12]Bugs fixed in the 3.4 series
 
          + [13]Fortran
 
     * [14]Non-bugs
 
          + [15]General
 
          + [16]C
 
          + [17]C++
 
               o [18]Common problems when upgrading the compiler
 
     _________________________________________________________________
 
 
 
                                Reporting Bugs
 
 
 
   The main purpose of a bug report is to enable us to fix the bug. The most
 
   important prerequisite for this is that the report must be complete and
 
   self-contained.
 
 
 
   Before you report a bug, please check the [19]list of well-known bugs and,
 
   if possible, try a current development snapshot. If you want to report a bug
 
   with versions of GCC before 3.4 we strongly recommend upgrading to the
 
   current release first.
 
 
 
   Before reporting that GCC compiles your code incorrectly, please compile it
 
   with gcc -Wall and see whether this shows anything wrong with your code that
 
   could be the cause instead of a bug in GCC.
 
 
 
Summarized bug reporting instructions
 
 
 
   After this summary, you'll find detailed bug reporting instructions, that
 
   explain how to obtain some of the information requested in this summary.
 
 
 
  What we need
 
 
 
   Please include in your bug report all of the following items, the first
 
   three of which can be obtained from the output of gcc -v:
 
     * the exact version of GCC;
 
     * the system type;
 
     * the options given when GCC was configured/built;
 
     * the complete command line that triggers the bug;
 
     * the compiler output (error messages, warnings, etc.); and
 
     * the preprocessed file (*.i*) that triggers the bug, generated by adding
 
       -save-temps to the complete compilation command, or, in the case of a
 
       bug report for the GNAT front end, a complete set of source files (see
 
       below).
 
 
 
  What we do not want
 
 
 
     * A source file that #includes header files that are left out of the bug
 
       report (see above)
 
     * That source file and a collection of header files.
 
     * An attached archive (tar, zip, shar, whatever) containing all (or some
 
       :-) of the above.
 
     * A code snippet that won't cause the compiler to produce the exact output
 
       mentioned in the bug report (e.g., a snippet with just a few lines
 
       around  the one that apparently triggers the bug, with some pieces
 
       replaced with ellipses or comments for extra obfuscation :-)
 
     * The  location  (URL) of the package that failed to build (we won't
 
       download it, anyway, since you've already given us what we need to
 
       duplicate the bug, haven't you? :-)
 
     * An error that occurs only some of the times a certain file is compiled,
 
       such that retrying a sufficient number of times results in a successful
 
       compilation; this is a symptom of a hardware problem, not of a compiler
 
       bug (sorry)
 
     * Assembly files (*.s) produced by the compiler, or any binary files, such
 
       as object files, executables, core files, or precompiled header files
 
     * Duplicate  bug  reports,  or  reports of bugs already fixed in the
 
       development tree, especially those that have already been reported as
 
       fixed last week :-)
 
     * Bugs in the assembler, the linker or the C library. These are separate
 
       projects,  with separate mailing lists and different bug reporting
 
       procedures
 
     * Bugs in releases or snapshots of GCC not issued by the GNU Project.
 
       Report them to whoever provided you with the release
 
     * Questions about the correctness or the expected behavior of certain
 
       constructs that are not GCC extensions. Ask them in forums dedicated to
 
       the discussion of the programming language
 
 
 
  Where to post it
 
 
 
   Please  submit  your  bug report directly to the [20]GCC bug database.
 
   Alternatively, you can use the gccbug script that mails your bug report to
 
   the bug database.
 
   Only  if  all  this  is absolutely impossible, mail all information to
 
   [21]gcc-bugs@gcc.gnu.org.
 
 
 
Detailed bug reporting instructions
 
 
 
   Please refer to the [22]next section when reporting bugs in GNAT, the Ada
 
   compiler, or to the [23]one after that when reporting bugs that appear when
 
   using a precompiled header.
 
 
 
   In general, all the information we need can be obtained by collecting the
 
   command line below, as well as its output and the preprocessed file it
 
   generates.
 
 
 
     gcc -v -save-temps all-your-options source-file
 
 
 
   The only excuses to not send us the preprocessed sources are (i) if you've
 
   found a bug in the preprocessor, (ii) if you've reduced the testcase to a
 
   small file that doesn't include any other file or (iii) if the bug appears
 
   only when using precompiled headers. If you can't post the preprocessed
 
   sources because they're proprietary code, then try to create a small file
 
   that triggers the same problem.
 
 
 
   Since we're supposed to be able to re-create the assembly output (extension
 
   .s), you usually should not include it in the bug report, although you may
 
   want to post parts of it to point out assembly code you consider to be
 
   wrong.
 
 
 
   Please avoid posting an archive (.tar, .shar or .zip); we generally need
 
   just a single file to reproduce the bug (the .i/.ii/.f preprocessed file),
 
   and, by storing it in an archive, you're just making our volunteers' jobs
 
   harder. Only when your bug report requires multiple source files to be
 
   reproduced should you use an archive. This is, for example, the case if you
 
   are using INCLUDE directives in Fortran code, which are not processed by the
 
   preprocessor, but the compiler. In that case, we need the main file and all
 
   INCLUDEd files. In any case, make sure the compiler version, error message,
 
   etc, are included in the body of your bug report as plain text, even if
 
   needlessly duplicated as part of an archive.
 
 
 
Detailed bug reporting instructions for GNAT
 
 
 
   See the [24]previous section for bug reporting instructions for GCC language
 
   implementations other than Ada.
 
 
 
   Bug reports have to contain at least the following information in order to
 
   be useful:
 
     * the exact version of GCC, as shown by "gcc -v";
 
     * the system type;
 
     * the options when GCC was configured/built;
 
     * the exact command line passed to the gcc program triggering the bug (not
 
       just the flags passed to gnatmake, but gnatmake prints the parameters it
 
       passed to gcc)
 
     * a collection of source files for reproducing the bug, preferably a
 
       minimal set (see below);
 
     * a description of the expected behavior;
 
     * a description of actual behavior.
 
 
 
   If  your  code  depends  on  additional  source files (usually package
 
   specifications), submit the source code for these compilation units in a
 
   single file that is acceptable input to gnatchop, i.e. contains no non-Ada
 
   text. If the compilation terminated normally, you can usually obtain a list
 
   of dependencies using the "gnatls -d main_unit" command, where main_unit is
 
   the file name of the main compilation unit (which is also passed to gcc).
 
 
 
   If you report a bug which causes the compiler to print a bug box, include
 
   that bug box in your report, and do not forget to send all the source files
 
   listed after the bug box along with your report.
 
 
 
   If you use gnatprep, be sure to send in preprocessed sources (unless you
 
   have to report a bug in gnatprep).
 
 
 
   When you have checked that your report meets these criteria, please submit
 
   it according to our [25]generic instructions. (If you use a mailing list for
 
   reporting, please include an "[Ada]" tag in the subject.)
 
 
 
Detailed bug reporting instructions when using a precompiled header
 
 
 
   If you're encountering a bug when using a precompiled header, the first
 
   thing to do is to delete the precompiled header, and try running the same
 
   GCC command again. If the bug happens again, the bug doesn't really involve
 
   precompiled headers, please report it without using them by following the
 
   instructions [26]above.
 
 
 
   If you've found a bug while building a precompiled header (for instance, the
 
   compiler crashes), follow the usual instructions [27]above.
 
 
 
   If you've found a real precompiled header bug, what we'll need to reproduce
 
   it is the sources to build the precompiled header (as a single .i file), the
 
   source file that uses the precompiled header, any other headers that source
 
   file includes, and the command lines that you used to build the precompiled
 
   header and to use it.
 
 
 
   Please don't send us the actual precompiled header. It is likely to be very
 
   large and we can't use it to reproduce the problem.
 
     _________________________________________________________________
 
 
 
                        Frequently Reported Bugs in GCC
 
 
 
   This is a list of bugs in GCC that are reported very often, but not yet
 
   fixed. While it is certainly better to fix bugs instead of documenting them,
 
   this document might save people the effort of writing a bug report when the
 
   bug is already well-known.
 
 
 
   There are many reasons why a reported bug doesn't get fixed. It might be
 
   difficult to fix, or fixing it might break compatibility. Often, reports get
 
   a low priority when there is a simple work-around. In particular, bugs
 
   caused by invalid code have a simple work-around: fix the code.
 
     _________________________________________________________________
 
 
 
C++
 
 
 
  Missing features
 
 
 
   The export keyword is not implemented.
 
          Most C++ compilers (G++ included) do not yet implement export, which
 
          is necessary for separate compilation of template declarations and
 
          definitions. Without export, a template definition must be in scope
 
          to be used. The obvious workaround is simply to place all definitions
 
          in the header itself. Alternatively, the compilation unit containing
 
          template definitions may be included from the header.
 
 
 
  Bugs fixed in the 3.4 series
 
 
 
   The following bugs are present up to (and including) GCC 3.3.x. They have
 
   been fixed in 3.4.0.
 
 
 
   Two-stage name-lookup.
 
          GCC did not implement two-stage name-lookup (also see [28]below).
 
 
 
   Covariant return types.
 
          GCC did not implement non-trivial covariant returns.
 
 
 
   Parse errors for "simple" code.
 
          GCC gave parse errors for seemingly simple code, such as
 
 
 
struct A
 
{
 
  A();
 
  A(int);
 
};
 
 
 
struct B
 
{
 
  B(A);
 
  B(A,A);
 
  void foo();
 
};
 
 
 
A bar()
 
{
 
  B b(A(),A(1));  // Variable b, initialized with two temporaries
 
  B(A(2)).foo();  // B temporary, initialized with A temporary
 
  return (A());   // return A temporary
 
}
 
 
 
          Although being valid code, each of the three lines with a comment was
 
          rejected  by  GCC. The work-arounds for older compiler versions
 
          proposed below do not change the semantics of the programs at all.
 
 
 
          The problem in the first case was that GCC started to parse the
 
          declaration  of  b as a function called b returning B, taking a
 
          function returning A as an argument. When it encountered the 1, it
 
          was too late. To show the compiler that this should be really an
 
          expression, a comma operator with a dummy argument could be used:
 
 
 
B b((0,A()),A(1));
 
 
 
          The work-around for simpler cases like the second one was to add
 
          additional parentheses around the expressions that were mistaken as
 
          declarations:
 
 
 
(B(A(2))).foo();
 
 
 
          In the third case, however, additional parentheses were causing the
 
          problems: The compiler interpreted A() as a function (taking no
 
          arguments, returning A), and (A()) as a cast lacking an expression to
 
          be casted, hence the parse error. The work-around was to omit the
 
          parentheses:
 
 
 
return A();
 
 
 
          This problem occurred in a number of variants; in throw statements,
 
          people also frequently put the object in parentheses.
 
     _________________________________________________________________
 
 
 
Fortran
 
 
 
   G77 bugs are documented in the G77 manual rather than explicitly listed
 
   here. Please see [29]Known Causes of Trouble with GNU Fortran in the G77
 
   manual.
 
     _________________________________________________________________
 
 
 
                                   Non-bugs
 
 
 
   The  following are not actually bugs, but are reported often enough to
 
   warrant a mention here.
 
 
 
   It is not always a bug in the compiler, if code which "worked" in a previous
 
   version, is now rejected. Earlier versions of GCC sometimes were less picky
 
   about standard conformance and accepted invalid source code. In addition,
 
   programming languages themselves change, rendering code invalid that used to
 
   be conforming (this holds especially for C++). In either case, you should
 
   update your code to match recent language standards.
 
     _________________________________________________________________
 
 
 
General
 
 
 
   Problems with floating point numbers - the [30]most often reported non-bug.
 
          In  a  number  of  cases, GCC appears to perform floating point
 
          computations incorrectly. For example, the C++ program
 
 
 
#include 
 
 
 
int main()
 
{
 
  double a = 0.5;
 
  double b = 0.01;
 
  std::cout << (int)(a / b) << std::endl;
 
  return 0;
 
}
 
 
 
          might print 50 on some systems and optimization levels, and 49 on
 
          others.
 
 
 
          This is the result of rounding: The computer cannot represent all
 
          real numbers exactly, so it has to use approximations. When computing
 
          with  approximation, the computer needs to round to the nearest
 
          representable number.
 
 
 
          This is not a bug in the compiler, but an inherent limitation of the
 
          floating  point  types.  Please  study  [31]this paper for more
 
          information.
 
     _________________________________________________________________
 
 
 
C
 
 
 
   Increment/decrement operator (++/--) not working as expected - a [32]problem
 
          with many variations.
 
          The following expressions have unpredictable results:
 
 
 
x[i]=++i
 
foo(i,++i)
 
i*(++i)                 /* special case with foo=="operator*" */
 
std::cout << i << ++i   /* foo(foo(std::cout,i),++i)          */
 
 
 
          since the i without increment can be evaluated before or after ++i.
 
 
 
          The  C  and C++ standards have the notion of "sequence points".
 
          Everything that happens between two sequence points happens in an
 
          unspecified order, but it has to happen after the first and before
 
          the second sequence point. The end of a statement and a function call
 
          are examples for sequence points, whereas assignments and the comma
 
          between function arguments are not.
 
 
 
          Modifying a value twice between two sequence points as shown in the
 
          following examples is even worse:
 
 
 
i=++i
 
foo(++i,++i)
 
(++i)*(++i)               /* special case with foo=="operator*" */
 
std::cout << ++i << ++i   /* foo(foo(std::cout,++i),++i)        */
 
 
 
          This leads to undefined behavior (i.e. the compiler can do anything).
 
 
 
   Casting does not work as expected when optimization is turned on.
 
          This is often caused by a violation of aliasing rules, which are part
 
          of the ISO C standard. These rules say that a program is invalid if
 
          you try to access a variable through a pointer of an incompatible
 
          type. This is happening in the following example where a short is
 
          accessed through a pointer to integer (the code assumes 16-bit shorts
 
          and 32-bit ints):
 
 
 
#include 
 
 
 
int main()
 
{
 
  short a[2];
 
 
 
  a[0]=0x1111;
 
  a[1]=0x1111;
 
 
 
  *(int *)a = 0x22222222; /* violation of aliasing rules */
 
 
 
  printf("%x %x\n", a[0], a[1]);
 
  return 0;
 
}
 
 
 
          The aliasing rules were designed to allow compilers more aggressive
 
          optimization. Basically, a compiler can assume that all changes to
 
          variables happen through pointers or references to variables of a
 
          type compatible to the accessed variable. Dereferencing a pointer
 
          that violates the aliasing rules results in undefined behavior.
 
 
 
          In the case above, the compiler may assume that no access through an
 
          integer pointer can change the array a, consisting of shorts. Thus,
 
          printf may be called with the original values of a[0] and a[1]. What
 
          really happens is up to the compiler and may change with architecture
 
          and optimization level.
 
 
 
          Recent versions of GCC turn on the option -fstrict-aliasing (which
 
          allows alias-based optimizations) by default with -O2. And some
 
          architectures  then really print "1111 1111" as result. Without
 
          optimization the executable will generate the "expected" output "2222
 
          2222".
 
 
 
          To disable optimizations based on alias-analysis for faulty legacy
 
          code, the option -fno-strict-aliasing can be used as a work-around.
 
 
 
          The option -Wstrict-aliasing (which is included in -Wall) warns about
 
          some  - but not all - cases of violation of aliasing rules when
 
          -fstrict-aliasing is active.
 
 
 
          To fix the code above, you can use a union instead of a cast (note
 
          that  this  is  a GCC extension which might not work with other
 
          compilers):
 
 
 
#include 
 
 
 
int main()
 
{
 
  union
 
  {
 
    short a[2];
 
    int i;
 
  } u;
 
 
 
  u.a[0]=0x1111;
 
  u.a[1]=0x1111;
 
 
 
  u.i = 0x22222222;
 
 
 
  printf("%x %x\n", u.a[0], u.a[1]);
 
  return 0;
 
}
 
 
 
          Now the result will always be "2222 2222".
 
 
 
          For  some  more insight into the subject, please have a look at
 
          [33]this article.
 
 
 
   Cannot use preprocessor directive in macro arguments.
 
          Let me guess... you used an older version of GCC to compile code that
 
          looks something like this:
 
 
 
  memcpy(dest, src,
 
#ifdef PLATFORM1
 
         12
 
#else
 
         24
 
#endif
 
        );
 
 
 
          and you got a whole pile of error messages:
 
 
 
test.c:11: warning: preprocessing directive not recognized within macro arg
 
test.c:11: warning: preprocessing directive not recognized within macro arg
 
test.c:11: warning: preprocessing directive not recognized within macro arg
 
test.c: In function `foo':
 
test.c:6: undefined or invalid # directive
 
test.c:8: undefined or invalid # directive
 
test.c:9: parse error before `24'
 
test.c:10: undefined or invalid # directive
 
 
 
          This is because your C library's  happens to define memcpy
 
          as a macro - which is perfectly legitimate. In recent versions of
 
          glibc,  for  example, printf is among those functions which are
 
          implemented as macros.
 
 
 
          Versions of GCC prior to 3.3 did not allow you to put #ifdef (or any
 
          other preprocessor directive) inside the arguments of a macro. The
 
          code therefore would not compile.
 
 
 
          As of GCC 3.3 this kind of construct is always accepted and the
 
          preprocessor will probably do what you expect, but see the manual for
 
          detailed semantics.
 
 
 
          However,  this  kind  of code is not portable. It is "undefined
 
          behavior" according to the C standard; that means different compilers
 
          may do different things with it. It is always possible to rewrite
 
          code which uses conditionals inside macros so that it doesn't. You
 
          could write the above example
 
 
 
#ifdef PLATFORM1
 
   memcpy(dest, src, 12);
 
#else
 
   memcpy(dest, src, 24);
 
#endif
 
 
 
          This is a bit more typing, but I personally think it's better style
 
          in addition to being more portable.
 
 
 
   Cannot initialize a static variable with stdin.
 
          This has nothing to do with GCC, but people ask us about it a lot.
 
          Code like this:
 
 
 
#include 
 
 
 
FILE *yyin = stdin;
 
 
 
          will not compile with GNU libc, because stdin is not a constant. This
 
          was  done  deliberately,  to  make it easier to maintain binary
 
          compatibility  when  the  type  FILE needs to be changed. It is
 
          surprising for people used to traditional Unix C libraries, but it is
 
          permitted by the C standard.
 
 
 
          This construct commonly occurs in code generated by old versions of
 
          lex  or yacc. We suggest you try regenerating the parser with a
 
          current version of flex or bison, respectively. In your own code, the
 
          appropriate fix is to move the initialization to the beginning of
 
          main.
 
 
 
          There  is  a  common  misconception that the GCC developers are
 
          responsible for GNU libc. These are in fact two entirely separate
 
          projects; please check the [34]GNU libc web pages for details.
 
     _________________________________________________________________
 
 
 
C++
 
 
 
   Nested classes can access private members and types of the containing class.
 
          Defect report 45 clarifies that nested classes are members of the
 
          class  they are nested in, and so are granted access to private
 
          members of that class.
 
 
 
   G++ emits two copies of constructors and destructors.
 
          In general there are three types of constructors (and destructors).
 
 
 
         1. The complete object constructor/destructor.
 
         2. The base object constructor/destructor.
 
         3. The allocating constructor/deallocating destructor.
 
 
 
          The first two are different, when virtual base classes are involved.
 
 
 
   Global destructors are not run in the correct order.
 
          Global  destructors should be run in the reverse order of their
 
          constructors  completing. In most cases this is the same as the
 
          reverse  order  of  constructors  starting, but sometimes it is
 
          different, and that is important. You need to compile and link your
 
          programs with --use-cxa-atexit. We have not turned this switch on by
 
          default, as it requires a cxa aware runtime library (libc, glibc, or
 
          equivalent).
 
 
 
   Classes in exception specifiers must be complete types.
 
          [15.4]/1  tells you that you cannot have an incomplete type, or
 
          pointer  to  incomplete  (other than cv void *) in an exception
 
          specification.
 
 
 
   Exceptions don't work in multithreaded applications.
 
          You  need  to  rebuild g++ and libstdc++ with --enable-threads.
 
          Remember, C++ exceptions are not like hardware interrupts. You cannot
 
          throw an exception in one thread and catch it in another. You cannot
 
          throw an exception from a signal handler and catch it in the main
 
          thread.
 
 
 
   Templates, scoping, and digraphs.
 
          If you have a class in the global namespace, say named X, and want to
 
          give it as a template argument to some other class, say std::vector,
 
          then std::vector<::X> fails with a parser error.
 
 
 
          The reason is that the standard mandates that the sequence <: is
 
          treated  as  if  it  were  the token [. (There are several such
 
          combinations of characters - they are called digraphs.) Depending on
 
          the version, the compiler then reports a parse error before the
 
          character : (the colon before X) or a missing closing bracket ].
 
 
 
          The simplest way to avoid this is to write std::vector< ::X>, i.e.
 
          place  a  space between the opening angle bracket and the scope
 
          operator.
 
 
 
   Copy constructor access check while initializing a reference.
 
          Consider this code:
 
 
 
class A
 
{
 
public:
 
  A();
 
 
 
private:
 
  A(const A&);   // private copy ctor
 
};
 
 
 
A makeA(void);
 
void foo(const A&);
 
 
 
void bar(void)
 
{
 
  foo(A());       // error, copy ctor is not accessible
 
  foo(makeA());   // error, copy ctor is not accessible
 
 
 
  A a1;
 
  foo(a1);        // OK, a1 is a lvalue
 
}
 
 
 
          Starting with GCC 3.4.0, binding an rvalue to a const reference
 
          requires an accessible copy constructor. This might be surprising at
 
          first sight, especially since most popular compilers do not correctly
 
          implement this rule.
 
 
 
          The C++ Standard says that a temporary object should be created in
 
          this context and its contents filled with a copy of the object we are
 
          trying to bind to the reference; it also says that the temporary copy
 
          can be elided, but the semantic constraints (eg. accessibility) of
 
          the copy constructor still have to be checked.
 
 
 
          For further information, you can consult the following paragraphs of
 
          the C++ standard: [dcl.init.ref]/5, bullet 2, sub-bullet 1, and
 
          [class.temporary]/2.
 
 
 
  Common problems when upgrading the compiler
 
 
 
    ABI changes
 
 
 
   The C++ application binary interface (ABI) consists of two components: the
 
   first defines how the elements of classes are laid out, how functions are
 
   called, how function names are mangled, etc; the second part deals with the
 
   internals of the objects in libstdc++. Although we strive for a non-changing
 
   ABI, so far we have had to modify it with each major release. If you change
 
   your compiler to a different major release you must recompile all libraries
 
   that contain C++ code. If you fail to do so you risk getting linker errors
 
   or malfunctioning programs. Some of our Java support libraries also contain
 
   C++ code, so you might want to recompile all libraries to be safe. It should
 
   not be necessary to recompile if you have changed to a bug-fix release of
 
   the same version of the compiler; bug-fix releases are careful to avoid ABI
 
   changes. See also the [35]compatibility section of the GCC manual.
 
 
 
   Remark: A major release is designated by a change to the first or second
 
   component  of the two- or three-part version number. A minor (bug-fix)
 
   release is designated by a change to the third component only. Thus GCC 3.2
 
   and 3.3 are major releases, while 3.3.1 and 3.3.2 are bug-fix releases for
 
   GCC 3.3. With the 3.4 series we are introducing a new naming scheme; the
 
   first release of this series is 3.4.0 instead of just 3.4.
 
 
 
    Standard conformance
 
 
 
   With each release, we try to make G++ conform closer to the ISO C++ standard
 
   (available  at  [36]http://www.ncits.org/cplusplus.htm).  We have also
 
   implemented  some of the core and library defect reports (available at
 
   [37]http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html     &
 
   [38]http://www.open-std.org/jtc1/sc22/wg21/docs/lwg-defects.html
 
   respectively).
 
 
 
   Non-conforming legacy code that worked with older versions of GCC may be
 
   rejected by more recent compilers. There is no command-line switch to ensure
 
   compatibility in general, because trying to parse standard-conforming and
 
   old-style  code  at  the  same  time  would  render  the  C++ frontend
 
   unmaintainable. However, some non-conforming constructs are allowed when the
 
   command-line option -fpermissive is used.
 
 
 
   Two  milestones in standard conformance are GCC 3.0 (including a major
 
   overhaul of the standard library) and the 3.4.0 version (with its new C++
 
   parser).
 
 
 
    New in GCC 3.0
 
 
 
     * The  standard  library is much more conformant, and uses the std::
 
       namespace (which is now a real namespace, not an alias for ::).
 
     * The standard header files for the c library don't end with .h, but begin
 
       with c (i.e.  rather than ). The .h names are still
 
       available, but are deprecated.
 
     *  is deprecated, use  instead.
 
     * streambuf::seekoff  &  streambuf::seekpos are private, instead use
 
       streambuf::pubseekoff & streambuf::pubseekpos respectively.
 
     * If std::operator << (std::ostream &, long long) doesn't exist, you need
 
       to recompile libstdc++ with --enable-long-long.
 
 
 
   If you get lots of errors about things like cout not being found, you've
 
   most likely forgotten to tell the compiler to look in the std:: namespace.
 
   There are several ways to do this:
 
     * Say std::cout at the call. This is the most explicit way of saying what
 
       you mean.
 
     * Say using std::cout; somewhere before the call. You will need to do this
 
       for each function or type you wish to use from the standard library.
 
     * Say  using  namespace  std; somewhere before the call. This is the
 
       quick-but-dirty fix. This brings the whole of the std:: namespace into
 
       scope. Never do this in a header file, as every user of your header file
 
       will be affected by this decision.
 
 
 
    New in GCC 3.4.0
 
 
 
   The  new  parser  brings  a lot of improvements, especially concerning
 
   name-lookup.
 
     * The "implicit typename" extension got removed (it was already deprecated
 
       since GCC 3.1), so that the following code is now rejected, see [14.6]:
 
 
 
template  struct A
 
{
 
    typedef int X;
 
};
 
 
 
template  struct B
 
{
 
    A::X          x;  // error
 
    typename A::X y;  // OK
 
};
 
 
 
B b;
 
 
 
     * For  similar reasons, the following code now requires the template
 
       keyword, see [14.2]:
 
 
 
template  struct A
 
{
 
    template  struct X {};
 
};
 
 
 
template  struct B
 
{
 
    typename A::X<0>          x;  // error
 
    typename A::template X<0> y;  // OK
 
};
 
 
 
B b;
 
 
 
     * We  now  have two-stage name-lookup, so that the following code is
 
       rejected, see [14.6]/9:
 
 
 
template  int foo()
 
{
 
    return i;  // error
 
}
 
 
 
     * This also affects members of base classes, see [14.6.2]:
 
 
 
template  struct A
 
{
 
    int i, j;
 
};
 
 
 
template  struct B : A
 
{
 
    int foo1() { return i; }       // error
 
    int foo2() { return this->i; } // OK
 
    int foo3() { return B::i; } // OK
 
    int foo4() { return A::i; } // OK
 
 
 
    using A::j;
 
    int foo5() { return j; }       // OK
 
};
 
 
 
   In addition to the problems listed above, the manual contains a section on
 
   [39]Common Misunderstandings with GNU C++.
 
 
 
References
 
 
 
   1. http://gcc.gnu.org/bugs.html
 
   2. http://gcc.gnu.org/bugs.html#report
 
   3. http://gcc.gnu.org/bugs.html#need
 
   4. http://gcc.gnu.org/bugs.html#dontwant
 
   5. http://gcc.gnu.org/bugs.html#where
 
   6. http://gcc.gnu.org/bugs.html#detailed
 
   7. http://gcc.gnu.org/bugs.html#gnat
 
   8. http://gcc.gnu.org/bugs.html#pch
 
   9. http://gcc.gnu.org/bugs.html#known
 
  10. http://gcc.gnu.org/bugs.html#cxx
 
  11. http://gcc.gnu.org/bugs.html#missing
 
  12. http://gcc.gnu.org/bugs.html#fixed34
 
  13. http://gcc.gnu.org/bugs.html#fortran
 
  14. http://gcc.gnu.org/bugs.html#nonbugs
 
  15. http://gcc.gnu.org/bugs.html#nonbugs_general
 
  16. http://gcc.gnu.org/bugs.html#nonbugs_c
 
  17. http://gcc.gnu.org/bugs.html#nonbugs_cxx
 
  18. http://gcc.gnu.org/bugs.html#upgrading
 
  19. http://gcc.gnu.org/bugs.html#known
 
  20. http://gcc.gnu.org/bugzilla/
 
  21. mailto:gcc-bugs@gcc.gnu.org
 
  22. http://gcc.gnu.org/bugs.html#gnat
 
  23. http://gcc.gnu.org/bugs.html#pch
 
  24. http://gcc.gnu.org/bugs.html#detailed
 
  25. http://gcc.gnu.org/bugs.html#where
 
  26. http://gcc.gnu.org/bugs.html#detailed
 
  27. http://gcc.gnu.org/bugs.html#detailed
 
  28. http://gcc.gnu.org/bugs.html#new34
 
  29. http://gcc.gnu.org/onlinedocs/gcc-3.4.6/g77/Trouble.html
 
  30. http://gcc.gnu.org/PR323
 
  31. http://www.validlab.com/goldberg/paper.ps
 
  32. http://gcc.gnu.org/PR11751
 
  33. http://mail-index.NetBSD.org/tech-kern/2003/08/11/0001.html
 
  34. http://www.gnu.org/software/libc/
 
  35. http://gcc.gnu.org/onlinedocs/gcc/Compatibility.html
 
  36. http://www.ncits.org/cplusplus.htm
 
  37. http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html
 
  38. http://www.open-std.org/jtc1/sc22/wg21/docs/lwg-defects.html
 
  39. http://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Misunderstandings.html

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.