Heh, whats going on?

#!/usr/bin/perl --
use strict;
use warnings;
use Config;

use Inline C => Config =>
  USING => 'ParseRegExp',
  CCFLAGSEX => q["-DMY_FORMAT=\"%.17e\""],
  BUILD_NOISY => 1,
  CLEAN_AFTER_BUILD => 0,
;

use Inline C =><<'EOC';

void* foo(void) {
    printf("***%s***\n", MY_FORMAT);
    printf(MY_FORMAT, sqrt(2.0));
    printf("\n%.17e\n", sqrt(2.0));
}

EOC

foo();

__END__
$ perl inline-eumm-define-sprintf-11104628.pl
validate Stage
Starting Build Preprocess Stage
get_maps Stage
Finished Build Preprocess Stage

Starting Build Parse Stage
Finished Build Parse Stage

Starting Build Glue 1 Stage
Finished Build Glue 1 Stage

Starting Build Glue 2 Stage
Finished Build Glue 2 Stage

Starting Build Glue 3 Stage
Finished Build Glue 3 Stage

Starting Build Compile Stage
  Starting "perl Makefile.PL" Stage
Generating a dmake-style Makefile
Writing Makefile for inline_eumm_define_sprintf_11104628_pl_5316
Writing MYMETA.yml and MYMETA.json
  Finished "perl Makefile.PL" Stage

  Starting "make" Stage
Running Mkbootstrap for inline_eumm_define_sprintf_11104628_pl_5316 ()
"D:\perl\bin\perl.exe" -MExtUtils::Command -e chmod -- 644 "inline_eum
+m_define_sprintf_11104628_pl_5316.bs"
"D:\perl\bin\perl.exe" -MExtUtils::Command::MM -e cp_nonempty -- inlin
+e_eumm_define_sprintf_11104628_pl_5316.
bs blib\arch\auto\inline_eumm_define_sprintf_11104628_pl_5316\inline_e
+umm_define_sprintf_11104628_pl_5316.bs 644
"D:\perl\bin\perl.exe" "D:\perl\site\lib\ExtUtils\xsubpp"  -typemap "D
+:\perl\lib\ExtUti
ls\typemap"  inline_eumm_define_sprintf_11104628_pl_5316.xs > inline_e
+umm_define_sprintf_11104628_pl_5316.xsc
"D:\perl\bin\perl.exe" -MExtUtils::Command -e mv -- inline_eumm_define
+_sprintf_11104628_pl_5316.xsc inline_eu
mm_define_sprintf_11104628_pl_5316.c
gcc -c           -s -O2 -DWIN32  -DPERL_RELOCATABLE_INCPUSH -DPERL_TEX
+TMODE_SCRIPTS -DUSE_SITECUSTOMIZE -
DPERL_IMPLICIT_CONTEXT -DPERL_IMPLICIT_SYS -fno-strict-aliasing -mms-b
+itfields -D_USE_32BIT_TIME_T "-DMY_FORMAT=\"%.17e\
"" -s -O2         -DVERSION=\"0.00\"    -DXS_VERSION=\"0.00\"  "-ID:\p
+erl\lib\CORE"   inline_eumm_define_spri
ntf_11104628_pl_5316.c
"D:\perl\bin\perl.exe" -MExtUtils::Mksymlists \
     -e "Mksymlists('NAME'=>\"inline_eumm_define_sprintf_11104628_pl_5
+316\", 'DLBASE' => 'inline_eumm_define_sprintf_111
04628_pl_5316', 'DL_FUNCS' => {  }, 'FUNCLIST' => [], 'IMPORTS' => {  
+}, 'DL_VARS' => []);"
g++ inline_eumm_define_sprintf_11104628_pl_5316.def -o blib\arch\auto\
+inline_eumm_define_sprintf_11104628_pl_5316\inline
_eumm_define_sprintf_11104628_pl_5316.dll -mdll -D_USE_32BIT_TIME_T -s
+ -L"D:\perl\lib\CORE" -L"D:\perl\mingw\mingw32\i686-w64-mingw32\lib" 
+inline_eumm_define_sprintf_11104628_pl_5316.o   "D:\perl\lib\CORE\li
bperl516.a" -lmoldname -lkernel32 -luser32 -lgdi32 -lwinspool -lcomdlg
+32 -ladvapi32 -lshell32 -lole32 -loleaut32 -lnetap
i32 -luuid -lws2_32 -lmpr -lwinmm -lversion -lodbc32 -lodbccp32 -lcomc
+tl32 -Wl,--enable-auto-image-base
"D:\perl\bin\perl.exe" -MExtUtils::Command -e chmod -- 755 blib\arch\a
+uto\inline_eumm_define_sprintf_11104628
_pl_5316\inline_eumm_define_sprintf_11104628_pl_5316.dll
  Finished "make" Stage

  Starting "make install" Stage
"D:\perl\bin\perl.exe" -MExtUtils::Command::MM -e cp_nonempty -- inlin
+e_eumm_define_sprintf_11104628_pl_5316.
bs blib\arch\auto\inline_eumm_define_sprintf_11104628_pl_5316\inline_e
+umm_define_sprintf_11104628_pl_5316.bs 644
Files found in blib\arch: installing files in blib\lib into architectu
+re dependent library tree
Installing _Inline\lib\auto\inline_eumm_define_sprintf_11104628_pl_531
+6\inline_eumm_define_sprintf_11104628_
pl_5316.dll
  Finished "make install" Stage

  Starting Cleaning Up Stage
  Finished Cleaning Up Stage

Finished Build Compile Stage

***%.17e***
1.41421356237309510e+000
1.41421356237309510e+000
[download]

Heh, whats going on?

Good question ... I don't have an answer to that ;-)

Your code still results in the same error for me: fatal error: EXTERN.h: No such file or directory.
Looks like the "-iquote" that my Inline-C-0.81 invokes is stuffing things up.

UPDATE: No, it has nothing to do with "-iquote". Turns out that when I switched to I-C-0.76 I also switched to EU-MM-7.04, and the problem I'm facing is due solely to a change that occurs in EU::MM some time between 7.0401 (which is fine) and 7.1001 (which breaks Inline::C). I'll investigate further when I get a chance and file a bug report if warranted.

What version of Inline::C are you running ? When I switch to Inline-C-0.76, your script works fine - as does mine if I replace the incorrect SV * foo(... with void foo(....

But now I'm puzzled as to the output we get for printf("***%s***\n", MY_FORMAT);
The quotes either side of %.17e are apparently missing from MY_FORMAT. So how does printf(MY_FORMAT, sqrt(2.0)) even compile ?
Can you explain that ?

BTW, seems that my problem with finding a correct rendition for the DEFINE in the Makefile.PL is dependent upon the version of EU::MM.
With EU-MM-7.04, this is fine:

DEFINE => q["-DMY_FORMAT=\\"%.17\\""];
[download]

But with EU-MM-7.34 (which is what I've been using) the crucial percentage symbol is still being omitted.

Cheers,
Rob

But now I'm puzzled as to the output we get for printf("***%s***\n", MY_FORMAT); The quotes either side of %.17e are apparently missing from MY_FORMAT. So how does printf(MY_FORMAT, sqrt(2.0)) even compile ?

C syntax and preprocessing rules. MY_FORMAT is substituted by the C preprocessor, so the compiler sees printf("***%s***\n", "%.17e");, strips the quotes off of the string constant like any other, and the output is ***%.17e*** if everything works correctly. The % character is also special in the DOS shell, which uses it for variable substitution: %PATH% on DOS is $PATH on *nix. The trailing % is required, but I would not be surprised if bugs in the shell like to eat random % characters.

You will be much better off passing a number instead of a string constant if you want this to compile on Windows.

Inline 0.80
Inline::C 0.76
ExtUtils::MakeMaker 7.20
[download]