From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 8291 invoked by alias); 26 Apr 2004 02:14:13 -0000 Mailing-List: contact binutils-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: binutils-owner@sources.redhat.com Received: (qmail 8007 invoked from network); 26 Apr 2004 02:14:10 -0000 Received: from unknown (HELO yosemite.airs.com) (209.128.65.135) by sources.redhat.com with SMTP; 26 Apr 2004 02:14:10 -0000 Received: (qmail 27448 invoked by uid 10); 26 Apr 2004 02:14:09 -0000 Received: (qmail 9449 invoked by uid 500); 26 Apr 2004 02:13:58 -0000 Mail-Followup-To: hp@bitrange.com, schwab@suse.de, binutils@sources.redhat.com, zack@codesourcery.com From: Ian Lance Taylor To: Zack Weinberg Cc: hp@bitrange.com, schwab@suse.de, binutils@sources.redhat.com Subject: Re: demand_empty_rest_of_line and ignore_rest_of_line References: <40587E7B.10705@codesourcery.com> <405979E9.9030805@codesourcery.com> <87brlhdk47.fsf@codesourcery.com> <87d65va8gd.fsf@egil.codesourcery.com> Date: Mon, 26 Apr 2004 02:46:00 -0000 In-Reply-To: <87d65va8gd.fsf@egil.codesourcery.com> Message-ID: User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-SW-Source: 2004-04/txt/msg00685.txt.bz2 Zack Weinberg writes: > The denser packing, and the lack of a nicely lined up arguments field, > mean substantially more mental effort just to see what it's doing. I'd say you're making a mountain out of a molehill. You're condemning the entire world to have slower compiles because you don't want to work a little bit harder to read assembler code, nor to write a sed script to do beautification, nor to use the -fverbose-asm option. If there is really a 1% speedup with #NO_APP, then it might even save you time overall due to faster compiles, even at the cost of spending more time in the relatively rare cases that you have to look at i386 assembler output. I'm sure it would save time for everybody else. For that matter, this patch to gas would accept your example input in #NO_APP mode. Ian Index: tc-i386.c =================================================================== RCS file: /cvs/src/src/gas/config/tc-i386.c,v retrieving revision 1.150 diff -u -r1.150 tc-i386.c --- tc-i386.c 12 Mar 2004 10:14:29 -0000 1.150 +++ tc-i386.c 26 Apr 2004 02:05:49 -0000 @@ -237,6 +237,7 @@ #define is_operand_char(x) (operand_chars[(unsigned char) x]) #define is_register_char(x) (register_chars[(unsigned char) x]) #define is_space_char(x) ((x) == ' ') +#define is_space_or_tab_char(x) ((x) == ' ' || (x) == '\t') #define is_identifier_char(x) (identifier_chars[(unsigned char) x]) #define is_digit_char(x) (digit_chars[(unsigned char) x]) @@ -1506,7 +1507,7 @@ } l++; } - if (!is_space_char (*l) + if (!is_space_or_tab_char (*l) && *l != END_OF_INSN && *l != PREFIX_SEPARATOR && *l != ',') @@ -1528,7 +1529,7 @@ current_templates = hash_find (op_hash, mnemonic); if (*l != END_OF_INSN - && (!is_space_char (*l) || l[1] != END_OF_INSN) + && (!is_space_or_tab_char (*l) || l[1] != END_OF_INSN) && current_templates && (current_templates->start->opcode_modifier & IsPrefix)) { @@ -1673,7 +1674,7 @@ while (*l != END_OF_INSN) { /* Skip optional white space before operand. */ - if (is_space_char (*l)) + if (is_space_or_tab_char (*l)) ++l; if (!is_operand_char (*l) && *l != END_OF_INSN) {