sh(1) calls abort(3) on certain input (backslash followed by the character with ASCII code 130). How-To-Repeat: sh -c 'echo -e echo "\\\\\0202"' |sh Abort trap - core dumped
Try this patch. Index: expand.c =================================================================== RCS file: /home/ncvs/src/bin/sh/expand.c,v retrieving revision 1.32 diff -u -r1.32 expand.c --- expand.c 2000/05/15 12:33:17 1.32 +++ expand.c 2001/03/04 23:52:35 @@ -315,7 +315,7 @@ goto lose; *p = c; while ((c = *home++) != '\0') { - if (quotes && c >= 0 && SQSYNTAX[(int)c] == CCTL) + if (quotes && SQSYNTAX[(int)c] == CCTL) STPUTC(CTLESC, expdest); STPUTC(c, expdest); } @@ -478,7 +478,7 @@ } lastc = *p++; if (lastc != '\0') { - if (quotes && lastc >= 0 && syntax[(int)lastc] == CCTL) + if (quotes && syntax[(int)lastc] == CCTL) STPUTC(CTLESC, dest); STPUTC(lastc, dest); } @@ -694,7 +694,7 @@ } else { while (*val) { - if (quotes && *val >= 0 && + if (quotes && syntax[(int)*val] == CCTL) STPUTC(CTLESC, expdest); STPUTC(*val++, expdest); @@ -866,7 +866,7 @@ if (allow_split) { \ syntax = quoted? DQSYNTAX : BASESYNTAX; \ while (*p) { \ - if (*p >= 0 && syntax[(int)*p] == CCTL) \ + if (syntax[(int)*p] == CCTL) \ STPUTC(CTLESC, expdest); \ STPUTC(*p++, expdest); \ } \ Index: mksyntax.c =================================================================== RCS file: /home/ncvs/src/bin/sh/mksyntax.c,v retrieving revision 1.15 diff -u -r1.15 mksyntax.c --- mksyntax.c 2001/06/19 15:41:57 1.15 +++ mksyntax.c 2001/06/21 21:54:20 @@ -352,9 +352,9 @@ static char *macro[] = { "#define is_digit(c)\t((c >= 0 && is_type+SYNBASE)[c] & ISDIGIT)", - "#define is_alpha(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && isalpha((unsigned char) (c)))", - "#define is_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && ((c) == '_' || isalpha((unsigned char) (c))))", - "#define is_in_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && ((c) == '_' || isalnum((unsigned char) (c))))", + "#define is_alpha(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLQUOTEMARK) && isalpha((unsigned char) (c)))", + "#define is_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLQUOTEMARK) && ((c) == '_' || isalpha((unsigned char) (c))))", + "#define is_in_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLQUOTEMARK) && ((c) == '_' || isalnum((unsigned char) (c))))", "#define is_special(c)\t((is_type+SYNBASE)[c] & (ISSPECL|ISDIGIT))", NULL }; Index: parser.c =================================================================== RCS file: /home/ncvs/src/bin/sh/parser.c,v retrieving revision 1.36 diff -u -r1.36 parser.c --- parser.c 2001/04/09 12:46:19 1.36 +++ parser.c 2001/04/17 18:27:52 @@ -977,7 +977,7 @@ c != '`' && c != '$' && (c != '"' || eofmark != NULL)) USTPUTC('\\', out); - if (c >= 0 && SQSYNTAX[c] == CCTL) + if (SQSYNTAX[c] == CCTL) USTPUTC(CTLESC, out); else if (eofmark == NULL) USTPUTC(CTLQUOTEMARK, out); @@ -1496,7 +1496,7 @@ continue; if (c == CTLESC) p++; - else if (c >= 0 && BASESYNTAX[(int)c] == CCTL) + else if (BASESYNTAX[(int)c] == CCTL) return 0; } return 1; - Tor Egge
Nope, doesn't work. Your patch only hides the real bug, and simply backs out changes Martin did in parser.c,v 1.27, etc. > Index: parser.c > =================================================================== > RCS file: /home/ncvs/src/bin/sh/parser.c,v > retrieving revision 1.36 > diff -u -r1.36 parser.c > --- parser.c 2001/04/09 12:46:19 1.36 > +++ parser.c 2001/04/17 18:27:52 > @@ -977,7 +977,7 @@ > c != '`' && c != '$' && > (c != '"' || eofmark != NULL)) > USTPUTC('\\', out); > - if (c >= 0 && SQSYNTAX[c] == CCTL) > + if (SQSYNTAX[c] == CCTL) > USTPUTC(CTLESC, out); > else if (eofmark == NULL) > USTPUTC(CTLQUOTEMARK, out); With my test, `c' will be -126 here, and SQSYNTAX[-126] is obviously the wrong thing. Funny though, I've managed to make almost identical patch before I sent a PR. :-) Then realized that it's bogus, and the actual problem is that our sh(1) is not 8-bit clean (apparently). Cheers, -- Ruslan Ermilov Oracle Developer/DBA, ru@sunbay.com Sunbay Software AG, ru@FreeBSD.org FreeBSD committer, +380.652.512.251 Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age
> With my test, `c' will be -126 here, and SQSYNTAX[-126] is obviously t= he > wrong thing. It is obviously the right thing. From=20the generated file syntax.h: #define SYNBASE 129 #define SQSYNTAX (sqsyntax + SYNBASE) i.e. SQSYNTAX[-126] becomes (sqsyntax + 129)[-126] which is sqsyntax[3].= - Tor Egge
Ouch, then this definitely works for me, at least this doesn't abort(3)'s anymore. On Fri, Jun 29, 2001 at 02:14:49PM +0200, Tor.Egge@fast.no wrote: > > > > With my test, `c' will be -126 here, and SQSYNTAX[-126] is obviously the > > wrong thing. > > It is obviously the right thing. > > >From the generated file syntax.h: > > > #define SYNBASE 129 > #define SQSYNTAX (sqsyntax + SYNBASE) > > i.e. SQSYNTAX[-126] becomes (sqsyntax + 129)[-126] which is sqsyntax[3]. > > - Tor Egge -- Ruslan Ermilov Oracle Developer/DBA, ru@sunbay.com Sunbay Software AG, ru@FreeBSD.org FreeBSD committer, +380.652.512.251 Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age
Responsible Changed From-To: freebsd-bugs->cracauer Martin is the sh maintainer
State Changed From-To: open->closed This has been fixed in -current and -stable for a while now.