Re: [fpc-devel] Apache header translation Patch
You can't get rid of that warning, and it should actually even be worded more strongly. It is simply not possible to access C-style varargs parameters at the callee side in FPC (you may be able to hak using pointers which will work on some platforms, but it will crash and burn in most cases). This is the c prototype ( type ) translated, actualy it is a array of PCHAR. code taken from apache source const char *(*take_argv) (cmd_parms *parms, void *mconfig, int argc, char *const argv[]); /code Is the following translation valid ? take_argv_t = function (parms: Pcmd_parms; mconfig: Pointer; argc: cint; argv: array of PChar): PChar; cdecl; -- Inoussa O. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] Apache header translation Patch
On 12 Nov 2007, at 10:32, Inoussa OUEDRAOGO wrote: This is the c prototype ( type ) translated, actualy it is a array of PCHAR. code taken from apache source const char *(*take_argv) (cmd_parms *parms, void *mconfig, int argc, char *const argv[]); /code Is the following translation valid ? take_argv_t = function (parms: Pcmd_parms; mconfig: Pointer; argc: cint; argv: array of PChar): PChar; cdecl; I think it's better to do this: take_argv_t = function (parms: Pcmd_parms; mconfig: Pointer; argc:cint; argv: PPChar): PChar; cdecl; You can still index it as an array in FPC, but it's clearer what the actual type is that gets passed (an array of pchar is the same as a double pointer to char in C). I was indeed wrong about the varargs stuff I said earlier, for some reason I had read array of pchar as array of const. Jonas ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] Apache header translation Patch
More importantly then supressing it, which I consider something unimportant, would be knowing what it means, and if there is some kind of potential problem it indicates that we are overlooking. Agree, clarified by Jonas's response . By the way, have you been able to make a module on ( Linux + Apache 2.2.x ) ? I tried on - Ubuntu + Apache 2.2.3/2.2.4 - Fedora + Apache 2.2.4 without succes. Even worse, in the handler proc, r^.handler ( used to test if this request should be handled by the current module ) always holds some weird string,for a very basic module. On windows it does work. -- Inoussa O. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] Apache header translation Patch
double pointer to char in C). I was indeed wrong about the varargs stuff I said earlier, for some reason I had read array of pchar as array of const. Ok, Thanks. By the way, h2pas should also be corrected, because it translates code langage=c void f(int a, char* p[]); /code to code langage=pascal procedure f(a:longint; p:array of pchar);cdecl;external; /code Please file a bug report with examples, so that it doesn't get forgotten. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] Apache header translation Patch
done : http://www.freepascal.org/mantis/view.php?id=10156 2007/11/12, Marco van de Voort [EMAIL PROTECTED]: double pointer to char in C). I was indeed wrong about the varargs stuff I said earlier, for some reason I had read array of pchar as array of const. Ok, Thanks. By the way, h2pas should also be corrected, because it translates code langage=c void f(int a, char* p[]); /code to code langage=pascal procedure f(a:longint; p:array of pchar);cdecl;external; /code Please file a bug report with examples, so that it doesn't get forgotten. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel -- Inoussa O. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] fpmkunit
After reconsidering the -n option. It should depend on the -f / --config= option. Darius Darius Blaszijk wrote: I've been playing around with fpmake again and came to the conclusion that it exhibits some strange behavior (at least I think so) The -v option gave me the following fFCl command that was executed; fpc -n -FUunits\i386-win32\ -Fu..\..\src\ -Fuunits\i386-win32\* myexample.pp 1. What strikes me is that the command line has the option -n in it. It appears that this command line option is hard coded in TBuildEngine.GetCompilerCommand. I strongly believe it should become a property named DefaultConfigFile which is by default set to True ie no '-n'. 2. -Fuunits\i386-win32\* I did not explicitly added this path. fpmake just takes this path by default. Is this desired behavior? I tend to believe not. If agreed upon, I will prepare a patch. Darius ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
[fpc-devel] fpmkunit patch
The following patch fixes a broken TTarget.Mode. Darius Index: src/fpmkunit.pp === --- src/fpmkunit.pp (revision 9207) +++ src/fpmkunit.pp (working copy) @@ -1206,6 +1208,7 @@ begin Result:=LowerCase(GetenumName(TypeInfo(TCompilerMode),Ord(Mode))); + Result := Copy(Result, 3, Length(Result) - 2); end; Function StringToMode(const S : String) : TCompilerMode;___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel