Home > Archive > Tcl > May 2004 > compiling with TCL_MEM_DEBUG
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
compiling with TCL_MEM_DEBUG
|
|
| sheila miguez herndon 2004-05-12, 9:13 pm |
| In the realm of ftupid questions: I downloaded the source for tcl and tk,
and compiled with TCL_MEM_DEBUG on. I checked in Tcl_Main and Tk_Main
respectively. Tk_Main supposedly calls Tcl_InitMemory if TCL_MEM_DEBUG is
defined, and Tcl_Main seems to call it regardless. Considering that, why
isn't "memory" recognized as a command when I start either a tclsh or
wish shell? I'm  .
I have a wretched bug I'm trying to demolish, but I can't discuss it here
due to NDA stuff. :/
I need logging. Wish crashes due to something in my extension. I'm
betting it's a memory problem because I found some and plugged them, yet
I can still get wish to crash. I'm now reading all the Tcl C command
stuff to see what I can find.
I notice that whoever wrote the code writes to pointers that were
returned from Tcl_GetStringFromObj, and this is not supposed to be done
according to the man page (since our use doesn't fall into the
exception):
"It is passed back as a writable pointer so that extension author
creating their own Tcl_ObjType will be able to modify the string
representation within the Tcl_UpdateStringProc of their Tcl_ObjType.
Except for that limited purpose, the pointer returned by
Tcl_GetStringFromObj or Tcl_GetString should be treated as read-only. It
is recommended that this pointer be assigned to a (CONST char *)
variable."
--
sheila
| |
| Don Porter 2004-05-12, 9:13 pm |
| sheila miguez herndon wrote:
> In the realm of ftupid questions: I downloaded the source for tcl and tk,
> and compiled with TCL_MEM_DEBUG on. I checked in Tcl_Main and Tk_Main
> respectively. Tk_Main supposedly calls Tcl_InitMemory if TCL_MEM_DEBUG is
> defined, and Tcl_Main seems to call it regardless. Considering that, why
> isn't "memory" recognized as a command when I start either a tclsh or
> wish shell? I'm .
Perhaps your tclsh and wish programs are loading an installed
libtcl8.4.so that was not configured with TCL_MEM_DEBUG rather
than the one you configured for mem debugging, but did not install?
ldd should be able to tell you for sure.
I find that the --disable-shared configure option is my friend.
--
| Don Porter Mathematical and Computational Sciences Division |
| donald.porter@nist.gov Information Technology Laboratory |
| http://math.nist.gov/~DPorter/ NIST |
|_______________________________________
_______________________________|
| |
| sheila miguez herndon 2004-05-18, 3:43 pm |
| R. Timothy Edwards wrote in news:c7tb1c$r4j$1@aplcore.jhuapl.edu:
> Note that if the target app extends Tcl/Tk, you need to make
> sure that any use of malloc() & friends is recast to Tcl_Alloc()
> et al.; this can be done with a #define. A little more subtle is
> the use of strdup(), which stumped me for a while recently, as it
> makes an internall call to malloc() and therefore has to be
> rewritten to be used with TCL_MEM_DEBUG.
>
Shouldn't one be using ckalloc() and ckfree()?
--
sheila
|
|
|
|
|