home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
GEMini Atari
/
GEMini_Atari_CD-ROM_Walnut_Creek_December_1993.iso
/
files
/
utility
/
bison
/
bison.bug
< prev
next >
Wrap
Text File
|
1988-07-10
|
5KB
|
139 lines
BABYL OPTIONS:
Version: 5
Labels:
Note: This is the header of an rmail file.
Note: If you are seeing it in rmail,
Note: it means the file has no messages in it.
1,forwarded,,
Summary-line: 1-Apr tower #["ASC::GUTHERY%slb-test.csnet"@RELAY.CS.NET: Bison Bug?]
Received: by PREP.AI.MIT.EDU; Wed, 1 Apr 87 20:10:49 EST
Date: Wed, 1 Apr 87 20:10:49 EST
From: tower (Leonard H. Tower Jr.)
Message-Id: <8704020110.AA00495@prep.ai.mit.edu>
To: rms
Subject: ["ASC::GUTHERY%slb-test.csnet"@RELAY.CS.NET: Bison Bug?]
Reply-To: tower@prep.ai.mit.edu
Organization: Project GNU, Free Software Foundation,
1000 Mass. Ave., Cambridge, MA 02138, USA +1 (617) 876-3296
Home: 36 Porter Street, Somerville, MA 02143, USA +1 (617) 623-7739
Status: O
*** EOOH ***
Date: Wed, 1 Apr 87 20:10:49 EST
From: tower (Leonard H. Tower Jr.)
To: rms
Subject: ["ASC::GUTHERY%slb-test.csnet"@RELAY.CS.NET: Bison Bug?]
Reply-To: tower@prep.ai.mit.edu
Organization: Project GNU, Free Software Foundation,
1000 Mass. Ave., Cambridge, MA 02138, USA +1 (617) 876-3296
Home: 36 Porter Street, Somerville, MA 02143, USA +1 (617) 623-7739
Date: Wed, 1 Apr 87 14:12 EDT
From: "guthery%asc%slb-doll.csnet@relay.cs.net" <"ASC::GUTHERY%slb-test.csnet"@RELAY.CS.NET>
To: gnu@PREP.AI.MIT.EDU
Subject: Bison Bug?
X-Vms-To: SDR::IN%"gnu@prep.ai.mit.edu"
In the build_relations function in the lalr.c file in GNU Bison,
the code
length--;
done = 0;
while(!done) ...
should, I believe, be replaced with
length--;
done = 0;
if(length) while(!done) ...
otherwise things like
list: /* empty */ | item | list item;
produce an internal error in map_goto.
Cheers, Scott
1,,
Return-Path: <rms@wheaties.ai.mit.edu>
Received: by frosted-flakes.ai.mit.edu; Wed, 23 Sep 87 12:49:33 EDT
Date: Wed, 23 Sep 87 12:49:33 EDT
From: rms@wheaties.ai.mit.edu (Richard Stallman)
Message-Id: <8709231649.AA18034@frosted-flakes.ai.mit.edu>
To: rms
Subject: Bison
*** EOOH ***
Return-Path: <rms@wheaties.ai.mit.edu>
Date: Wed, 23 Sep 87 12:49:33 EDT
From: rms@wheaties.ai.mit.edu (Richard Stallman)
To: rms
Subject: Bison
Bison should have a feature to specify a prefix for all the names.
1,answered,,
Received: from prep.ai.mit.edu by wheaties.ai.mit.edu; Thu, 31 Dec 87 00:31:49 EST
Received: by PREP.AI.MIT.EDU; Thu, 31 Dec 87 00:26:53 EST
Received: from balance.aca.mcc.com by MCC.COM with TCP; Wed 30 Dec 87 23:32:14-CST
Date: Wed, 30 Dec 87 23:33:06 cst
From: tiemann@mcc.com (Michael Tiemann)
Posted-Date: Wed, 30 Dec 87 23:33:06 cst
Message-Id: <8712310533.AA17290@balance.aca.mcc.com>
Received: by balance.aca.mcc.com (4.12/ACA-V2.1)
id AA17290; Wed, 30 Dec 87 23:33:06 cst
To: rms@prep.ai.mit.edu
Subject: -fomit-frame-pointer (and YYEMPTY)
Status: RO
*** EOOH ***
Date: Wed, 30 Dec 87 23:33:06 cst
From: tiemann@mcc.com (Michael Tiemann)
Posted-Date: Wed, 30 Dec 87 23:33:06 cst
To: rms@prep.ai.mit.edu
Subject: -fomit-frame-pointer (and YYEMPTY)
YYEMPTY is true iff there is no current lookahead symbol. Such a case
arises after a shift, before the next call to yylex. Note that often,
reductions can be performed in the absense of any lookahead symbols.
By not reading the next symbol immediately after shifting, it is
possible for the parser to change the state of the lexical analyzer
before it goes after the next input symbol.
On the -fomit-frame-pointer front, I could not reproduce the bug
(which I only observed as a segmentation violation). When I could not
reproduce it on SB-Prolog, I recompiled GCC with itself, using
-fomit-frame-pointer. It did not blink. The total speed improvement
was about 1-2%. However, someone is running a large experiment on the
Balance which is keeping it from giving me "clean" readings. Here are
the total times for each complete make:
gcc -O: 3026.3u 757.1s 1:05:37 96% 3396+3848io 639pf+0w
gcc -O -fomit-frame-pointer: 2981.8u 749.0s 1:04:41 96% 3415+3810io 628pf+0w
These include the assembly and link times. Here are the two compiling
the file expr.c:
-O -fomit-frame-pointer -O alone
time in parse: 47.660000 time in parse: 47.770000
time in jump: 5.590000 time in jump: 5.830000
time in cse: 13.860000 time in cse: 14.120000
time in loop: 1.240000 time in loop: 1.230000
time in flow: 15.940000 time in flow: 15.980000
time in combine: 4.320000 time in combine: 4.520000
time in local-alloc: 11.870000 time in local-alloc: 12.150000
time in global-alloc: 15.310000 time in global-alloc: 15.840000
time in final: 8.850000 time in final: 9.410000
time in varconst: 0.510000 time in varconst: 0.690000
time in symout: 0.000000 time in symout: 0.000000
time in dump: 0.000000 time in dump: 0.000000
Michael