CRAN warning "Found 'sprintf', possibly from 'sprintf' (C)" using BH about bh

[mclements]

Well done! (I was concerned that was going to be a much larger task.)

[eddelbuettel]

I am aware. I have no fix. I said I have no plan to edit BH manually:

edd@rob:~/git/bh(master)$ ag -c sprint inst/include/ 
inst/include/boost/asio/detail/impl/handler_tracking.ipp:3
inst/include/boost/asio/detail/impl/win_static_mutex.ipp:1
inst/include/boost/asio/detail/impl/socket_ops.ipp:11
inst/include/boost/asio/ip/impl/network_v4.ipp:3
inst/include/boost/asio/ip/impl/network_v6.ipp:3
inst/include/boost/asio/impl/connect_pipe.ipp:1
inst/include/boost/assert/source_location.hpp:1
inst/include/boost/lexical_cast/detail/converter_lexical_streams.hpp:6
inst/include/boost/test/utils/setcolor.hpp:2
inst/include/boost/test/impl/debug.ipp:2
inst/include/boost/test/impl/test_tools.ipp:2
inst/include/boost/compute/detail/parameter_cache.hpp:1
inst/include/boost/interprocess/detail/win32_api.hpp:1
inst/include/boost/xpressive/traits/detail/c_ctype.hpp:2
inst/include/boost/spirit/home/support/char_encoding/ascii.hpp:2
inst/include/boost/spirit/home/support/char_encoding/standard.hpp:2
inst/include/boost/spirit/home/support/char_encoding/iso8859_1.hpp:2
inst/include/boost/spirit/home/support/char_encoding/unicode.hpp:1
inst/include/boost/spirit/home/support/char_encoding/standard_wide.hpp:1
inst/include/boost/spirit/home/support/char_class.hpp:1
inst/include/boost/spirit/home/x3/support/traits/print_attribute.hpp:1
inst/include/boost/spirit/home/x3/support/traits/print_token.hpp:1
inst/include/boost/spirit/home/classic/core/primitives/impl/primitives.ipp:4
inst/include/boost/spirit/home/classic/core/primitives/primitives.hpp:1
inst/include/boost/regex/v4/c_regex_traits.hpp:1
inst/include/boost/regex/v4/regex_workaround.hpp:1
inst/include/boost/regex/v5/c_regex_traits.hpp:1
inst/include/boost/numeric/odeint/integrate/max_step_checker.hpp:2
inst/include/boost/core/lightweight_test.hpp:2
inst/include/boost/core/type_name.hpp:1
edd@rob:~/git/bh(master)$ 

That is from using the release candidate 1.81.0 described here: #88

[eddelbuettel]

A closer look reveals that a few of them are in fact sprintf_s (also safe) or preceded (in the case of Asio) by an #ifdef for snprintf so maybe I can get to this.

I am a little fearful of fat-fingering it but CRAN leaves me little choice I guess.

[eddelbuettel]

Some notes to self: Using ag from emacs showing the matches reveals some are in comment. sprintf_s is also fine

-*- mode: ag; default-directory: "~/git/bh/inst/include/boost/" -*-
Ag started at Thu Dec 29 11:24:15

ag --literal --group --line-number --column --color --color-match 30\;43 --color-path 1\;32 --smart-case --stats -- sprintf .
File: asio/detail/impl/handler_tracking.ipp
363:31:  using namespace std; // For sprintf (or equivalent).
372:17:  int length = vsprintf_s(line, sizeof(line), format, args);
374:17:  int length = vsprintf(line, format, args);

File: asio/detail/impl/win_static_mutex.ipp
43:31:  using namespace std; // For sprintf.

File: asio/detail/impl/socket_ops.ipp
2416:31:  using namespace std; // For sprintf.
2420:5:    sprintf_s(dest, length, "%u.%u.%u.%u",
2432:14:        n += sprintf_s(dest + n, length - n, ":%s", b < 16 ? "" : ":"), ++z;
2436:14:        n += sprintf_s(dest + n, length - n, "%s%x", b ? ":" : "",
2442:12:      n += sprintf_s(dest + n, length - n, "%%%lu", scope_id);
2514:44:    using namespace std; // For strcat and sprintf.
2526:7:      sprintf(if_name + 1, "%lu", scope_id);
3638:7:      sprintf_s(serv, servlen, "%u", ntohs(port));
3640:7:      sprintf(serv, "%u", ntohs(port));
3663:9:        sprintf_s(serv, servlen, "%u", ntohs(port));
3665:9:        sprintf(serv, "%u", ntohs(port));

File: asio/ip/impl/network_v6.ipp
99:31:  using namespace std; // For sprintf.
105:3:  sprintf_s(prefix_len, sizeof(prefix_len), "/%u", prefix_length_);
107:3:  sprintf(prefix_len, "/%u", prefix_length_);

File: asio/ip/impl/network_v4.ipp
130:31:  using namespace std; // For sprintf.
136:3:  sprintf_s(prefix_len, sizeof(prefix_len), "/%u", prefix_length_);
138:3:  sprintf(prefix_len, "/%u", prefix_length_);

File: asio/impl/connect_pipe.ipp
49:31:  using namespace std; // For sprintf and memcmp.

File: assert/source_location.hpp
79:58:# define BOOST_ASSERT_SNPRINTF(buffer, format, arg) std::sprintf(buffer, format, arg)

File: lexical_cast/detail/converter_lexical_streams.hpp
283:21:                    sprintf_s(begin, CharacterBufferSize,
285:21:                    sprintf(begin,
295:21:                    sprintf_s(begin, CharacterBufferSize,
297:21:                    sprintf(begin,
308:21:                    sprintf_s(begin, CharacterBufferSize,
310:21:                    sprintf(begin,

File: test/utils/setcolor.hpp
92:31:        m_command_size = std::sprintf( m_control_command, "%c[%c;3%c;4%cm",
103:31:        m_command_size = std::sprintf(m_control_command, "%c[%c;3%c;4%cm",

File: test/impl/debug.ipp
45:41:namespace std { using ::memset; using ::sprintf; }
883:10:    std::sprintf( cmd_line, format, ::GetCurrentProcessId(), dbg_init_done_ev );

File: compute/detail/parameter_cache.hpp
123:37:            #define DETAIL_SNPRINTF sprintf_s

File: interprocess/detail/win32_api.hpp
1801:24:                  std::sprintf(&stamp_str[0], "%u", ((unsigned int)pTypedRecord->TimeGenerated));

File: regex/v4/regex_workaround.hpp
53:12:   using ::sprintf; using ::strcpy; using ::strcat; using ::strlen;

File: numeric/odeint/integrate/max_step_checker.hpp
72:18:            std::sprintf(error_msg, "Max number of iterations exceeded (%d).", m_max_steps);
104:18:            std::sprintf(error_msg, "Max number of iterations exceeded (%d). A new step size was not found.", m_max_steps);

File: core/lightweight_test.hpp
214:14:        std::sprintf( buffer, "\\x%02X", static_cast<unsigned char>( v ) );

File: core/type_name.hpp
191:10:    std::sprintf( buffer, "%lu", static_cast< unsigned long >( n ) );
40 matches
16 files contained matches
12575 files searched
129671692 bytes searched
0.101163 seconds

Ag finished at Thu Dec 29 11:24:15

lexical_cast seems gnarlier but points to a new (portable) boost::core:snprintf in this Boost Core PR which may get combined with this PR in lexical_cast.

Several other ones are easier.

[eddelbuettel]

Thanks for the gentle prodding. I created a new branch and the changeset is actually small and reasonable (apart from the issue mentioned in the previous comment, for which I found the proper treatment in that PR.

Testing this now, and with some luck this should be on CRAN soon after they reopen (and we take care of process in #88, i.e. let a small number of packages update) after which your issue will be taken care of.