Re: [patch/testsuite/cp] local.exp: accommodate gcc abi 2

Re: [patch/testsuite/cp] local.exp: accommodate gcc abi 2

[Michael Elizabeth Chastain]

ac> Is this "The New ABI" or a yet to be named but for the moment would be 
ac> best refered to as "The New New ABI"?

The former.

It's minor version 2 of "The New ABI", also known as "The IA-64 ABI".
It contains small bug fixes to g++ to make it more compatible with
the spec for "The New ABI".

  gcc 2.95.3	old abi
  gcc 3.3.2	new abi, version 1
  gcc 3.4.0	new abi, version 2

Here's an example of the differences between "new abi, version 1"
and "new abi, version 2": suppose that a class has a virtual destructor,
but the user doesn't specify the virtual destructor, so the compiler
synthesizes it.  With "new abi, version 1", the synthesized virtual
destructor appears in the vtable *before* the user's own virtual
methods.  With "new abi, version 2", the synthesized virtual destructor
appears in the vtable *after* the user's own virtual methods.

There are about 5-10 tweaks like this in gcc 3.4.

So it's not "New New ABI", fortunately for us.  It's just bug fixes
to "New ABI".  It's the same mangling scheme and same data structures,
but gcc puts slightly different data into them.  Different enough that
g++ 3.3.X and g++ 3.4.X are sometimes not link-compatible, but similar
enough that gdb will not need a new module.

gcc 3.4 has a flag "-fabi-version" to select between "New Abi,
version 1" and "New Abi, version 2".  "New Abi, version 2" is the
default.

Michael C

Re: [patch/testsuite/cp] local.exp: accommodate gcc abi 2

[Andrew Cagney]

> ac> Is this "The New ABI" or a yet to be named but for the moment would be 
> ac> best refered to as "The New New ABI"?
> 
> The former.
> 
> It's minor version 2 of "The New ABI", also known as "The IA-64 ABI".
> It contains small bug fixes to g++ to make it more compatible with
> the spec for "The New ABI".
> 
>   gcc 2.95.3	old abi

As daniel eluded to, there was an even earlier ABI (I've no idea of its 
history, I guess it was 1.x related).  "The New ABI" is really the third 
ABI.

>   gcc 3.3.2	new abi, version 1
3.1
>   gcc 3.4.0	new abi, version 2
3.2
> 
> Here's an example of the differences between "new abi, version 1"
> and "new abi, version 2": suppose that a class has a virtual destructor,
> but the user doesn't specify the virtual destructor, so the compiler
> synthesizes it.  With "new abi, version 1", the synthesized virtual
> destructor appears in the vtable *before* the user's own virtual
> methods.  With "new abi, version 2", the synthesized virtual destructor
> appears in the vtable *after* the user's own virtual methods.
> 
> There are about 5-10 tweaks like this in gcc 3.4.
> 
> So it's not "New New ABI", fortunately for us.  It's just bug fixes
> to "New ABI".  It's the same mangling scheme and same data structures,
> but gcc puts slightly different data into them.  Different enough that
> g++ 3.3.X and g++ 3.4.X are sometimes not link-compatible, but similar
> enough that gdb will not need a new module.
> 
> gcc 3.4 has a flag "-fabi-version" to select between "New Abi,
> version 1" and "New Abi, version 2".  "New Abi, version 2" is the
> default.

So "abi-version" is the minor version :-/

"logical" :-^

Andrew

Re: [patch/testsuite/cp] local.exp: accommodate gcc abi 2

[Daniel Jacobowitz]

On Mon, Jan 05, 2004 at 11:01:47AM -0500, Andrew Cagney wrote:
> >This expands local.exp to accommodate gcc HEAD -gstabs+ with the new abi.
> >As with several other files, I just cloned the pattern for
> >gcc 3.3.2 -gstabs+ abi-1 and re-ordered the fields.
> >
> 
> Michael, humor me :-)
> 
> Is this "The New ABI" or a yet to be named but for the moment would be 
> best refered to as "The New New ABI"?
> 
> Several years ago GCC added what they should add a new ABI and so called 
> it "The New ABI" (...).  I believe its motivation and specification both 
> at least in part originated with the IA-64.  I thought "The New ABI" 
> became known as "ABI 2"?

It's not a new ABI in the sense that you mean.  It's a new ABI version
- gcc now has a command line switch for incompatible bug-fixes to the
existing ABI.  The only change between -fabi-version=1 and
-fabi-version=2 that will affect GDB is probably this one, which moves
virtual destructors to another location in the vtable.

This is still the "Itanium C++ ABI".  It's just a little closer to the
published document than it used to be :)  I don't remember it being
called ABI 2 - if anything, it was called the v3 ABI, for GCC 3.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch/testsuite/cp] local.exp: accommodate gcc abi 2

[Andrew Cagney]

> This expands local.exp to accommodate gcc HEAD -gstabs+ with the new abi.
> As with several other files, I just cloned the pattern for
> gcc 3.3.2 -gstabs+ abi-1 and re-ordered the fields.
> 

Michael, humor me :-)

Is this "The New ABI" or a yet to be named but for the moment would be 
best refered to as "The New New ABI"?

Several years ago GCC added what they should add a new ABI and so called 
it "The New ABI" (...).  I believe its motivation and specification both 
at least in part originated with the IA-64.  I thought "The New ABI" 
became known as "ABI 2"?

Andrew

[patch/testsuite/cp] local.exp: accommodate gcc abi 2

[Michael Elizabeth Chastain]

This expands local.exp to accommodate gcc HEAD -gstabs+ with the new abi.
As with several other files, I just cloned the pattern for
gcc 3.3.2 -gstabs+ abi-1 and re-ordered the fields.

Tested with:

  native i686-pc-linux-gnu
  gcc 2.95.3 3.3.2 HEAD
  dwarf2, stabs+
  tcl 8.4.5, expect 5.39, dejagnu 1.4.3
  tcl + expect + dejagnu from sourceware

I am committing this now.

Michael C

2004-01-02  Michael Chastain  <mec.gnu@mindspring.com>

	* gdb.cp/local.exp: Accept gcc abi 2.

Index: local.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.cp/local.exp,v
retrieving revision 1.1
diff -c -3 -p -r1.1 local.exp
*** local.exp	23 Aug 2003 03:55:59 -0000	1.1
--- local.exp	3 Jan 2004 01:40:16 -0000
***************
*** 1,4 ****
! # Copyright 1998, 1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
  
  # This program is free software; you can redistribute it and/or modify
  # it under the terms of the GNU General Public License as published by
--- 1,5 ----
! # Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004
! # Free Software Foundation, Inc.
  
  # This program is free software; you can redistribute it and/or modify
  # it under the terms of the GNU General Public License as published by
*************** gdb_test "up" ".*foobar.*" "up from mark
*** 96,101 ****
--- 97,109 ----
  #
  # chastain 2002-04-08
  
+ # pattern #5
+ # PASS
+ #   stabs+
+ #     gcc HEAD abi-2
+ #
+ # chastain 2004-01-02
+ 
  set sep "(\[.\]|___)\[0-9\]"
  
  send_gdb "ptype Local\n"
*************** gdb_expect {
*** 106,111 ****
--- 114,120 ----
    }
    -re "type = class Local \{\[\r\n\t \]*public:\[\r\n\t \]*int loc1;\[\r\n\t \]*Local & operator *=\\((foobar__Fi${sep}::|)Local const *&\\);\[\r\n\t \]*Local\\((foobar__Fi${sep}::|)Local const *&\\);\[\r\n\t \]*Local\\((void|)\\);\[\r\n\t \]*char loc_foo\\(char\\);\[\r\n\t \]*\}.*$gdb_prompt $" { pass "ptype Local" }
    -re "type = class Local \{\r\n\[\t \]*public:\r\n\[\t \]*int loc1;\r\n\r\n\[\t \]*char loc_foo\\(char\\);\r\n\[\t \]*\\(Local at.*local\\.cc:\[0-9\]*\\)\r\n\}.*$gdb_prompt $" { xpass "ptype Local (old aCC)" }
+   -re "type = class Local \{\[\r\n\t \]*public:\[\r\n\t \]*int loc1;\[\r\n\t \]*char loc_foo\\(char\\);\[\r\n\t \]*Local & operator *=\\((foobar__Fi${sep}::|)Local const *&\\);\[\r\n\t \]*Local\\((foobar__Fi${sep}::|)Local const *&\\);\[\r\n\t \]*Local\\((void|)\\);\[\r\n\t \]*\}.*$gdb_prompt $" { pass "ptype Local" }
    -re ".*$gdb_prompt $"   {  fail "ptype Local" }
    timeout             { fail "(timeout) ptype Local" }
  }
*************** gdb_expect {
*** 131,136 ****
--- 140,149 ----
          # this right.
          kfail gdb/825 "Local out of scope"
      }
+     -re "ptype Local${eol}type = class Local {${eol}  public:${eol}    int loc1;${eol}    char loc_foo\\(char\\);${eol}.*${eol}}${eol}${gdb_prompt} " {
+ 	# gcc 3.X abi-2 -gstabs+
+         kfail gdb/825 "Local out of scope"
+     }
      -re ".*${gdb_prompt} $" {
          fail "Local out of scope"
      }
*************** gdb_expect {
*** 168,179 ****
--- 181,200 ----
  #
  # chastain 2002-05-27
  
+ # Pattern 5:
+ # PASS
+ #   stabs+
+ #     HEAD
+ #
+ # chastain 2004-01-02
+ 
  send_gdb "ptype InnerLocal\n"
  gdb_expect {
    -re "type = class InnerLocal \{\[\r\n\t \]*public:\[\r\n\t \]*char ilc;\[\r\n\t \]*int \\* *ip;\[\r\n\t \]*(InnerLocal::|)NestedInnerLocal nest1;\[\r\n\t \]*int il_foo\\((unsigned char const|const unsigned char) *&\\);\[\r\n\t \]*\}.*$gdb_prompt $" { pass "ptype InnerLocal (pattern 1)" }
    -re "type = class InnerLocal \{\[\r\n\t \]*public:\[\r\n\t \]*char ilc;\[\r\n\t \]*int \\* *ip;\[\r\n\t \]*(InnerLocal::|)NestedInnerLocal nest1;\[\r\n\t \]*InnerLocal *& operator *=\\((main${sep}::|)InnerLocal const *&\\);\[\r\n\t \]*InnerLocal\\((main${sep}::|)InnerLocal const *&\\);\[\r\n\t \]*InnerLocal\\((void|)\\);\[\r\n\t \]*int il_foo\\(unsigned char const *&\\);\[\r\n\t \]*\}.*$gdb_prompt $" { pass "ptype InnerLocal (pattern 2)" }
    -re "type = class InnerLocal \{\r\n\[\t \]*public:\r\n\[\t \]*char ilc;\r\n\[\t \]*int \\*ip;\r\n\[\t \]*InnerLocal::NestedInnerLocal nest1;\r\n\r\n\[\t \]*.int il_foo\\(unsigned char const &\\);\r\n\[\t \]*\}\[\t \]*\\(Local at.*local\\.cc:36\\).*$gdb_prompt $" { pass "ptype InnerLocal (old HP aCC)" }
    -re "type = class InnerLocal \{\r\n\[\t \]*public:\r\n\[\t \]*char ilc;\r\n\[\t \]*int \\*ip;\r\n\[\t \]*class InnerLocal4::NestedInnerLocal nest1;\r\n\r\n\[\t \]*int il_foo\\(unsigned char const &\\);\r\n\[\t \]*\\(Local at.*local\.cc:\[0-9\]+\\)\r\n\}.*$gdb_prompt $" { pass "ptype InnerLocal (old HP aCC)" }
+   -re "type = class InnerLocal \{\[\r\n\t \]*public:\[\r\n\t \]*char ilc;\[\r\n\t \]*int \\* *ip;\[\r\n\t \]*(InnerLocal::|)NestedInnerLocal nest1;\[\r\n\t \]* int il_foo\\(unsigned char const *&\\);\[\r\n\t \]*InnerLocal *& operator *=\\((main${sep}::|)InnerLocal const *&\\);\[\r\n\t \]*InnerLocal\\((main${sep}::|)InnerLocal const *&\\);\[\r\n\t \]*InnerLocal\\((void|)\\);\[\r\n\t \]*\}.*$gdb_prompt $" { pass "ptype InnerLocal (pattern 5)" }
    -re ".*$gdb_prompt $"   {  fail "ptype InnerLocal" }
    timeout             { fail "(timeout) ptype InnerLocal" }
  }