[Tcl-bugs] [ tcl-Bugs-1090869 ] Tcl_GetInt Returns Illegal Overlfow On AMD64

[SourceForge.net <no...@so...>]

Bugs item #1090869, was opened at 2004-12-24 10:43
Message generated for change (Comment added) made by dgp
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=1090869&group_id=10894

Category: 44. Parsing and Eval
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: sanjeev singh (sanjeev_singh)
Assigned to: Kevin B KENNY (kennykb)
Summary: Tcl_GetInt Returns Illegal Overlfow On AMD64

Initial Comment:
Hi,
I am running on Opteron ie Amd 64 bit mc. While porting 
my 32 bit code I am running into this problem with 
reading 32 bit numbers that have bit 31 set.

Basically the code below fails if the user passed 
0x8000000 as the arguments.

int tb_test(ClientData dummy, Tcl_Interp *interp, int 
argc, char **argv) 
{
  int code = TCL_OK;  
  int value[2];

  Tcl_GetInt(interp, argv[1], &value[0]);
  Tcl_GetInt(interp, argv[2],  &value[1]);

  printf("VALUE0=0x%x\n", value[0]);
  printf("VALUE1=0x%x\n", value[1]);

  return code;
}

When run in tclsh this gives
% myCmd 0x7fffffff 0x80000000
VALUE0=0x7fffffff
VALUE1=0x0
integer value too large to represent
% 

The problem seems to be in file tclGet.c in the following 
line-
 if ((errno == ERANGE) || (((long)(int) i) != i)) {

It should be instead
 if ((errno == ERANGE) || (((long)(unsigned int) i) != i)) {

-sanjeev


----------------------------------------------------------------------

>Comment By: Don Porter (dgp)
Date: 2005-04-19 13:33

Message:
Logged In: YES 
user_id=80530


That proposed patch fails 54 tests
in the test suite.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=1090869&group_id=10894