|tcl and conservative GC? kurt_stephens@black_guard.swissbank.com (1994-07-01)|
|Re: tcl and conservative GC? email@example.com (1994-07-02)|
|From:||firstname.lastname@example.org (Chris N Anderson)|
|Organization:||The World Public Access UNIX, Brookline, MA|
|Date:||Sat, 2 Jul 1994 20:45:08 GMT|
"Standard" Tcl/Tk/TclX stuff does not encode pointers in strings in the
manner you suggest. Rather, unique identifiers are generated which are
placed in hash tables that are used to translate the "pointer" string to
an actual C pointer; thus, a conservative GC should have no problem - the
real pointer is in the hash table.
Any packages that might encode pointers into strings without this method
have no means to protect against invalid pointer-strings that could crash
the program. Of course, CGC will not work with such packages, but I'm not
personally aware of any.
Return to the
Search the comp.compilers archives again.