[Bug dyninst/28790] New: Replace deprecated C++98 junk with lambdas

[Bug dyninst/28790] New: Replace deprecated C++98 junk with lambdas

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

            Bug ID: 28790
           Summary: Replace deprecated C++98 junk with lambdas
           Product: systemtap
           Version: unspecified
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: dyninst
          Assignee: systemtap at sourceware dot org
          Reporter: jwakely.gcc at gmail dot com
  Target Milestone: ---

I have no idea what the correct Component is, sorry.

GCC 12 now (rightly) warns about systemtap's use of std::ptr_fun

util.cxx: In function 'void ltrim(std::string&)':
util.cxx:1764:56: error: 'std::pointer_to_unary_function<_Arg, _Result>
std::ptr_fun(_Result (*)(_Arg)) [with _Arg = int; _Result = int]' is
deprecated: use 'std::mem_fn' instead [-Werror=deprecated-declarations]
 1764 |                        std::not1(std::ptr_fun<int,
int>(std::isspace))));
      |                                  ~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~
In file included from
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/string:48,
                 from util.h:8,
                 from util.cxx:17:
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/bits/stl_function.h:1126:5:
note: declared here
 1126 |     ptr_fun(_Result (*__x)(_Arg))
      |     ^~~~~~~
util.cxx:1764:33: error: 'constexpr std::unary_negate<_Predicate>
std::not1(const _Predicate&) [with _Predicate = pointer_to_unary_function<int,
int>]' is deprecated: use 'std::not_fn' instead
[-Werror=deprecated-declarations]
 1764 |                        std::not1(std::ptr_fun<int,
int>(std::isspace))));
      |                        ~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/string:48,
                 from util.h:8,
                 from util.cxx:17:
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/bits/stl_function.h:1046:5:
note: declared here
 1046 |     not1(const _Predicate& __pred)
      |     ^~~~
util.cxx: In function 'void rtrim(std::string&)':
util.cxx:1772:43: error: 'std::pointer_to_unary_function<_Arg, _Result>
std::ptr_fun(_Result (*)(_Arg)) [with _Arg = int; _Result = int]' is
deprecated: use 'std::mem_fn' instead [-Werror=deprecated-declarations]
 1772 |           std::not1(std::ptr_fun<int, int>(std::isspace))).base(),
s.end());
      |                     ~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~
In file included from
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/string:48,
                 from util.h:8,
                 from util.cxx:17:
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/bits/stl_function.h:1126:5:
note: declared here
 1126 |     ptr_fun(_Result (*__x)(_Arg))
      |     ^~~~~~~
util.cxx:1772:20: error: 'constexpr std::unary_negate<_Predicate>
std::not1(const _Predicate&) [with _Predicate = pointer_to_unary_function<int,
int>]' is deprecated: use 'std::not_fn' instead
[-Werror=deprecated-declarations]
 1772 |           std::not1(std::ptr_fun<int, int>(std::isspace))).base(),
s.end());
      |           ~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/string:48,
                 from util.h:8,
                 from util.cxx:17:
/nix/store/w4d0lkjirdzr19cqdq428j0kqlzc4lgd-gcc-12.0.0/include/c++/12.0.0/bits/stl_function.h:1046:5:
note: declared here
 1046 |     not1(const _Predicate& __pred)
      |     ^~~~
cc1plus: all warnings being treated as errors

std::ptr_fun has not been part of the C++ standard since C++17, so you should
avoid using it (or ensure you compile as C++14).

The code is wrong for platforms with signed char anyway, with potentially
undefined behaviour. The argument to std::isspace(int) must have the value of
an unsigned char, i.e. it must be between (int)0 and (int)255. If you call it
with a signed char that has a negative value, you get a negative int, which is
outside the allowed range.

You can kill two birds with one stone by just getting rid of the std::ptr_fun
uses entirely:

diff --git a/util.cxx b/util.cxx
index c20f76003..76c751f18 100644
--- a/util.cxx
+++ b/util.cxx
@@ -1763,7 +1763,7 @@ ltrim(std::string &s)
 {
   s.erase(s.begin(),
          std::find_if(s.begin(), s.end(),
-                      std::not1(std::ptr_fun<int, int>(std::isspace))));
+                      [](unsigned char c) { return !std::isspace(c); }));
 }

 // trim from end (in place)
@@ -1771,7 +1771,7 @@ void
 rtrim(std::string &s)
 {
   s.erase(std::find_if(s.rbegin(), s.rend(),
-         std::not1(std::ptr_fun<int, int>(std::isspace))).base(), s.end());
+         [](unsigned char c) { return !std::isspace(c); }).base(), s.end());
 }

 // trim from both ends (in place)

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #1 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
N.B. the use of unsigned char as the lambda argument converts the incoming char
to unsigned char, avoiding the UB.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[fche at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

Frank Ch. Eigler <fche at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fche at redhat dot com
         Resolution|---                         |FIXED
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #2 from Frank Ch. Eigler <fche at redhat dot com> ---
Thanks, merged!

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[me at serhei dot io]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

Serhei Makarov <me at serhei dot io> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |me at serhei dot io

--- Comment #3 from Serhei Makarov <me at serhei dot io> ---
Note that commit f199d19 replaces the lambdas with a named function ptr

Someday we might want to give up building on older systems to use latest C++
features, but saving a couple of lines in ltrim()/rtrim() is probably not
sufficient reason to do it. stap continues to build just fine with this simple
tweak.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #4 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
My patch uses nothing newer than C++11, which is already required for building
GDB isn't?

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #5 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
Doh sorry, this wasn't a GDB bug! brainfart

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[me at serhei dot io]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #6 from Serhei Makarov <me at serhei dot io> ---
The lambdas were unsupported on a buildbot compiling stap with gcc 4.4.7 (rhel6
system compiler; I know). Since the codebase compiled just fine otherwise & the
fix was simple, from an upstream PoV I didn't see rtrim() as a sufficient
reason to deprecate this configuration.

If it was a complex piece of code that depends on lambdas as a simplifying
abstraction, and avoiding them would multiply the complexity, I'd probably make
the call otherwise.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #7 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
Makes perfect sense. The fact the two functions are right next to each other
and a named function can be defined immediately before them makes it an even
easier choice.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug dyninst/28790] Replace deprecated C++98 junk with lambdas

[me at serhei dot io]

https://sourceware.org/bugzilla/show_bug.cgi?id=28790

--- Comment #8 from Serhei Makarov <me at serhei dot io> ---
Ah, I misunderstood the import of your earlier comment and that it applies to
the GDB codebase in particular.

It does make me wonder when/if stap will have a feature that depends on C++11
features in a serious way....

-- 
You are receiving this mail because:
You are the assignee for the bug.