Re: c/8927: Gcc give error for wrong line of C code.

Re: c/8927: Gcc give error for wrong line of C code.

[bangerth]

Synopsis: Gcc give error for wrong line of C code.

State-Changed-From-To: open->analyzed
State-Changed-By: bangerth
State-Changed-When: Fri Dec 20 20:54:04 2002
State-Changed-Why:
    Confirmed. This is the same problem which is reported in
    another report (which I am unable to find at present,
    annoyingly) where you had a token that creates an
    error and immediately afterwards a #include. The error
    was reported to be inside the include file.
    
    Errors should be reported for the position of a token,
    not for the position of the next one.
    
    W.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=8927

Re: c/8927: Gcc give error for wrong line of C code.

[Joseph S. Myers]

The following reply was made to PR c/8927; it has been noted by GNATS.

From: "Joseph S. Myers" <jsm28@cam.ac.uk>
To: Neil Booth <neil@daikokuya.co.uk>
Cc: Steven Bosscher <s.bosscher@student.tudelft.nl>,  <gcc-gnats@gcc.gnu.org>, 
     <gcc-bugs@gcc.gnu.org>,  <gp@qnx.com>, 
    Wolfgang Bangerth <bangerth@ticam.utexas.edu>
Subject: Re: c/8927: Gcc give error for wrong line of C code.
Date: Sun, 16 Feb 2003 00:09:54 +0000 (GMT)

 On Sat, 15 Feb 2003, Neil Booth wrote:
 
 > Too bad the front end separates them.  This whole area is currently
 > quite messy, and with bison unprofitably hard to fix.  The C++ front
 
 I'll point out again that Bison has built-in support for syntax
 productions to have a location as well as a value, although no GCC front
 ends currently use it.  (It would at least be reasonable for example front
 ends such as treelang to make use of this Bison feature.)
 
 -- 
 Joseph S. Myers
 jsm28@cam.ac.uk
 

Re: c/8927: Gcc give error for wrong line of C code.

[Neil Booth]

The following reply was made to PR c/8927; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.co.uk>
To: "Joseph S. Myers" <jsm28@cam.ac.uk>
Cc: Steven Bosscher <s.bosscher@student.tudelft.nl>, gcc-gnats@gcc.gnu.org,
	gcc-bugs@gcc.gnu.org, gp@qnx.com,
	Wolfgang Bangerth <bangerth@ticam.utexas.edu>
Subject: Re: c/8927: Gcc give error for wrong line of C code.
Date: Sun, 16 Feb 2003 00:14:06 +0000

 Joseph S. Myers wrote:-
 
 > On Sat, 15 Feb 2003, Neil Booth wrote:
 > 
 > > Too bad the front end separates them.  This whole area is currently
 > > quite messy, and with bison unprofitably hard to fix.  The C++ front
 > 
 > I'll point out again that Bison has built-in support for syntax
 > productions to have a location as well as a value, although no GCC front
 > ends currently use it.  (It would at least be reasonable for example front
 > ends such as treelang to make use of this Bison feature.)
 
 Hence the use of "unprofitably", which you even referred to in your
 original mail.  We don't want things slower or harder to understand
 than they already are.
 
 Neil.

Re: c/8927: Gcc give error for wrong line of C code.

[Neil Booth]

The following reply was made to PR c/8927; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.co.uk>
To: Steven Bosscher <s.bosscher@student.tudelft.nl>
Cc: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gp@qnx.com,
	nobody@gcc.gnu.org, gcc-prs@gcc.gnu.org,
	Wolfgang Bangerth <bangerth@ticam.utexas.edu>
Subject: Re: c/8927: Gcc give error for wrong line of C code.
Date: Sat, 15 Feb 2003 23:57:37 +0000

 Steven Bosscher wrote:-
 
 > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=8927
 > 
 > The same problem happens for g++, so this is not a C bug.
 > 
 > c_lex() and c-parse.in:_yylex() return the right line number for the
 > token.  But bison looks ahead a bit and that causes the line number to
 > advance as well.  This is also why saving the line number in the
 > "primary" production doesn't help: Before we parse primary:IDENTIFIER,
 > we've already advanced the line number.
 > 
 > So my guess is that this is a cpplib bug: It should not advance the line
 > number until the previous token is accepted.  Dunno how to attack this
 > bug.
 
 CPP doesn't control the line number; it attaches line numbers to tokens,
 and is therefore never wrong 8-)
 
 Too bad the front end separates them.  This whole area is currently
 quite messy, and with bison unprofitably hard to fix.  The C++ front
 end should be able to do a near-perfect job now; though I can imagine
 the current implementation of c-lex.c (which isn't appropriate to
 RD parsers) getting in the way.
 
 Neil.

Re: c/8927: Gcc give error for wrong line of C code.

[Steven Bosscher]

The following reply was made to PR c/8927; it has been noted by GNATS.

From: Steven Bosscher <s.bosscher@student.tudelft.nl>
To: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gp@qnx.com,
	nobody@gcc.gnu.org, gcc-prs@gcc.gnu.org
Cc: Wolfgang Bangerth <bangerth@ticam.utexas.edu>
Subject: Re: c/8927: Gcc give error for wrong line of C code.
Date: 15 Feb 2003 22:56:46 +0100

 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=8927
 
 The same problem happens for g++, so this is not a C bug.
 
 c_lex() and c-parse.in:_yylex() return the right line number for the
 token.  But bison looks ahead a bit and that causes the line number to
 advance as well.  This is also why saving the line number in the
 "primary" production doesn't help: Before we parse primary:IDENTIFIER,
 we've already advanced the line number.
 
 So my guess is that this is a cpplib bug: It should not advance the line
 number until the previous token is accepted.  Dunno how to attack this
 bug.
 
 Greetz
 Steven