[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Had anybody problems with clisp-1996-07-22 before me?

I am just wondering. I am currently hacking using clisp-1996-07-22 and
it seems to have some problems:

 - It seems to have a serve memory leak. Well, yesterday I was
   wondering why a lisp session only five hours old had according to
   ps(1) grew to 30MB. Even after I explicitly kicked out my own data
   structures clisp reclaimed only about 1MB (after heavy trashing of
   course -- the full garbage collection lasted longer than one
   cigarette so the 30MB was not imaginary). The 1MB was approx the
   size of my data structure.
   Then I took a brute force attack -- uninterning *all* symbols from
   *all* my packages (and user), after another trashing the core size
   reduced only by a couple of kilo words.
   Then I 'rebooted' CLISP.

   That the X server grows to 24MB easily is known, but CLISP?
   Come on, tell me that CLISP o.k! Or do we start to write bad code
   too, because C++ hackers do? :-)

 - Sometimes CLISP dies badly with no appearing reason. Even
   evaluating a symbol from the REP could do that. This could be
   pretty nice my own code, but I am clueless here.

I would love to know if other people had similiar problems with 07-22.
On the weekend I'll try to get a reproducible example for the memory
leak. I have still the malloc implementation under suspect.  [I have
an example -- my current project -- but it is pretty much too large
(~3000 lines) to say anything useful.]

Anyhow, in the meantime I'll switch back to 05-30 to get work done. 
I am not in the mood now to do C hacking.

Wish you a good hack,

PS. No, I am not telling what my current project is, because all of you
    would laugh. Anybody want to try a guess? :-)