From edi at agharta.de Fri Feb 6 13:21:14 2004 From: edi at agharta.de (Edi Weitz) Date: Fri, 06 Feb 2004 14:21:14 +0100 Subject: [cl-ppcre-devel] Re: install fails - Allegro 4.3 In-Reply-To: <200402061132.i16BWvsG012556@griffin.cs.nyu.edu> (Roman Yangarber's message of "Fri, 6 Feb 2004 06:32:57 -0500 (EST)") References: <200402061132.i16BWvsG012556@griffin.cs.nyu.edu> Message-ID: On Fri, 6 Feb 2004 06:32:57 -0500 (EST), Roman Yangarber wrote: > i could not install the version i downoladed from your page > yesterday. it crashes, the message given below. > > in the next section i give the trace of the crash. in the following > section i show how i hacked around it. Hi Roman! Could you please use the mailing list at for bug reports. That way other people can read them, too, and maybe help out. To me this looks like a bug in AllegroCL 4.3. I'd suggest that you ask Franz support if they have a patch for it. Could you run all tests successfully after changing the threshold? Cheers, Edi. > ______________________________________________________________________________ > USER(3): (mk:compile-system "cl-ppcre") > ; Loading /home/Data/Software/cl-ppcre/cl-ppcre.system > ;;; Compiling file /home/Data/Software/cl-ppcre/packages.lisp > ;;; Writing fasl file /home/Data/Software/cl-ppcre/packages.fasl > ;;; Fasl write complete > > < compiles and loads everything up to...> > > ;;; Writing fasl file /home/Data/Software/cl-ppcre/api.fasl > ;;; Fasl write complete > ; Fast loading /home/Data/Software/cl-ppcre/api.fasl > Error: %puthash internal error: full hash table: > #(#\p #\a #\b #\c #\d #\e #\f #\g #\h #\i #\j #\k #\l #\m #\n #\o) 1 1 > > Restart actions (select using :continue): > 0: retry the load of /home/Data/Software/cl-ppcre/api.fasl > 1: skip loading /home/Data/Software/cl-ppcre/api.fasl > > [changing package from "COMMON-LISP-USER" to "CL-PPCRE"] > [1] PPCRE(4): :zo > Evaluation stack: > > (ERROR "%puthash internal error: full hash table: ~s ~s ~s" #(# # # ...) > ...) > ->(ADD-RANGE-TO-HASH # #\a ...) > (CONVERT-CHAR-CLASS-TO-HASH (# # # ...)) > (CONVERT-AUX (NIL NIL NIL)) > (CONVERT-AUX (:GROUP #)) > (CONVERT (:GROUP #)) > ((METHOD CREATE-SCANNER (T)) (:GROUP #) :CASE-INSENSITIVE-MODE ...) > (CREATE-SCANNER (:GROUP #) :CASE-INSENSITIVE-MODE ...) > ((METHOD CREATE-SCANNER (STRING)) :UNKNOWN) > ______________________________________________________________________________ > my patch/fix/hack around: > > i had to modify the settings for the make-hash-table call in convert.lisp: > > (defun convert-char-class-to-hash (list) ... > > (loop with hash = (make-hash-table :size (ceiling (expt *regex-char-code-limit* (/ 1 4))) > :rehash-size (float (expt *regex-char-code-limit* (/ 1 4))) > :rehash-threshold 0.5) > > ... ) > > changing the :size and :rehash-size didn't work. lowering :rehash-threshold > from 1.0 to .5 seemed to help (for now?!) > > it seemed like the hash table is full and is not getting resized properly. > (i don't really understand how hash-tables work in lisp, so i'm not confident > at all that this is the right thing to do.) From edi at agharta.de Tue Feb 17 00:35:53 2004 From: edi at agharta.de (Edi Weitz) Date: Tue, 17 Feb 2004 01:35:53 +0100 Subject: [cl-ppcre-devel] New release 0.7.4 Message-ID: Hi! A new release is available from . Here's the relevant part from the changelog: Version 0.7.4 2004-02-16 Fixed wrong call to SIGNAL-PPCRE-SIGNAL-ERROR in lexer.lisp (caught by Peter Graves) Added :CL-PPCRE to *FEATURES* (for CL-INTERPOL) Compiler macro for SPLIT Have fun, Edi.