c++/10244: [3.4 regression] In error messages for templates, template type is no longer displayed

c++/10244: [3.4 regression] In error messages for templates, template type is no longer displayed

[bangerth]

>Number:         10244
>Category:       c++
>Synopsis:       [3.4 regression] In error messages for templates, template type is no longer displayed
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Mar 27 15:56:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     Wolfgang Bangerth
>Release:        unknown-1.0
>Organization:
>Environment:
present mainline
>Description:
For a code like this:
--------------------------
template <class T> void f(T) { BOGUS; }
int main() { f (1); }
--------------------------
we would previously get an error message that states
the type of T:

g/x> /home/bangerth/bin/gcc-3.3-pre/bin/c++ -c compiler_bug.cc
compiler_bug.cc: In function `void f(T) [with T = int]':
compiler_bug.cc:9:   instantiated from here
compiler_bug.cc:4: error: `BOGUS' undeclared (first use this function)
compiler_bug.cc:4: error: (Each undeclared identifier is reported only once for
   each function it appears in.)


With present mainline, this "[with T=int]" is missing:
g/x> /home/bangerth/bin/gcc-3.4-pre/bin/c++ -c compiler_bug.cc
compiler_bug.cc: In function `void f(T)':
compiler_bug.cc:4: error: `BOGUS' has not been declared

I think this makes finding errors harder to find, and
consider this a regression.

W.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:

Re: c++/10244: [3.4 regression] In error messages for templates, template type is no longer displayed

[reichelt]

Synopsis: [3.4 regression] In error messages for templates, template type is no longer displayed

State-Changed-From-To: open->closed
State-Changed-By: reichelt
State-Changed-When: Tue Apr 29 20:46:04 2003
State-Changed-Why:
    Not a bug, IMHO.
    
    It's a consequence of improved two-stage name-lookup.
    Since BOGUS is not a dependant name, the parser can see that this is a bug,
    while parsing the template (a not while instantiating it). As you can see,
    not only the "[with T = int]" is missing, but also the
    "compiler_bug.cc:9:   instantiated from here" line. There is in fact no
    value for T when the error message is issued.
    
    
    On the other hand, if I change the code to
    
    ------------------------------------------
    template <class T> void f(T) { T::BOGUS; }
    int main() { f (1); }
    ------------------------------------------
    
    the compiler has to wait for the instantiation to be able to see that
    BOGUS is really bogus. In this case I get:
    
    bogus.cc: In function `void f(T) [with T = int]':
    bogus.cc:2:   instantiated from here
    bogus.cc:1: error: `int' is not an aggregate type
    
    which does make sense to me.
    
    Regards,
    Volker

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10244