diff mbox

[go] : Restore bootstrap

Message ID CAFULd4a2Y-Gvnm9XROQdfPqFsG0ZcdOBL8ic2PsjkGv5+hjW3Q@mail.gmail.com
State New
Headers show

Commit Message

Uros Bizjak July 24, 2014, 5:57 p.m. UTC
Hello!

2014-07-24  Uros Bizjak  <ubizjak@gmail.com>

    * go/go-gcc.cc (Gcc_backend::global_variable_set_init): Rename
    symtab_get_node to symtab_node::get.

Tested on x86_64-linux-gnu and committed to mainline SVN.

Uros.

Comments

Martin Liška July 24, 2014, 7:21 p.m. UTC | #1
On 07/24/2014 07:57 PM, Uros Bizjak wrote:
> Hello!
>
> 2014-07-24  Uros Bizjak  <ubizjak@gmail.com>
>
>      * go/go-gcc.cc (Gcc_backend::global_variable_set_init): Rename
>      symtab_get_node to symtab_node::get.
>
> Tested on x86_64-linux-gnu and committed to mainline SVN.
>
> Uros.
Hello,
     thank you for your fix. As I see my configure output:

$ ../configure --disable-bootstrap --enable-checking=release 
--enable-languages=all --disable-multilib

The following languages will be built: c,c++,fortran,java,lto,objc
*** This configuration is not supported in the following subdirectories:
      gnattools target-libada target-libgo target-libbacktrace
     (Any other directories should still work fine.)

I thought with 'all' I cover all front-ends, is it really intended 
behavior that go is disable by 'all'?

Thank you,
Martin
Ian Lance Taylor July 25, 2014, 2:51 a.m. UTC | #2
On Thu, Jul 24, 2014 at 12:21 PM, Martin Liška <mliska@suse.cz> wrote:
> On 07/24/2014 07:57 PM, Uros Bizjak wrote:
>>
>> Hello!
>>
>> 2014-07-24  Uros Bizjak  <ubizjak@gmail.com>
>>
>>      * go/go-gcc.cc (Gcc_backend::global_variable_set_init): Rename
>>      symtab_get_node to symtab_node::get.
>>
>> Tested on x86_64-linux-gnu and committed to mainline SVN.
>>
>> Uros.
>
> Hello,
>     thank you for your fix. As I see my configure output:
>
> $ ../configure --disable-bootstrap --enable-checking=release
> --enable-languages=all --disable-multilib
>
> The following languages will be built: c,c++,fortran,java,lto,objc
> *** This configuration is not supported in the following subdirectories:
>      gnattools target-libada target-libgo target-libbacktrace
>     (Any other directories should still work fine.)
>
> I thought with 'all' I cover all front-ends, is it really intended behavior
> that go is disable by 'all'?

Yes, --enable-languages=all is the default, so it only builds the
frontends that are enabled by default, so it does not build the Go or
Ada frontends.  I know it doesn't make much sense.

Ian
Martin Liška July 25, 2014, 6:35 a.m. UTC | #3
On 07/25/2014 03:51 AM, Ian Lance Taylor wrote:
> On Thu, Jul 24, 2014 at 12:21 PM, Martin Liška <mliska@suse.cz> wrote:
>> On 07/24/2014 07:57 PM, Uros Bizjak wrote:
>>> Hello!
>>>
>>> 2014-07-24  Uros Bizjak  <ubizjak@gmail.com>
>>>
>>>       * go/go-gcc.cc (Gcc_backend::global_variable_set_init): Rename
>>>       symtab_get_node to symtab_node::get.
>>>
>>> Tested on x86_64-linux-gnu and committed to mainline SVN.
>>>
>>> Uros.
>> Hello,
>>      thank you for your fix. As I see my configure output:
>>
>> $ ../configure --disable-bootstrap --enable-checking=release
>> --enable-languages=all --disable-multilib
>>
>> The following languages will be built: c,c++,fortran,java,lto,objc
>> *** This configuration is not supported in the following subdirectories:
>>       gnattools target-libada target-libgo target-libbacktrace
>>      (Any other directories should still work fine.)
>>
>> I thought with 'all' I cover all front-ends, is it really intended behavior
>> that go is disable by 'all'?
> Yes, --enable-languages=all is the default, so it only builds the
> frontends that are enabled by default, so it does not build the Go or
> Ada frontends.  I know it doesn't make much sense.
>
> Ian
I would suggest to replace currently used 'all' to 'default'. And 'all' can be really used for all possible frontends we have :) I know it changes the behavior, but I hope I makes sense?

Thank you,
Martin
Gerald Pfeifer July 28, 2014, 10:29 a.m. UTC | #4
On Fri, 25 Jul 2014, Martin Liška wrote:
>> Yes, --enable-languages=all is the default, so it only builds the
>> frontends that are enabled by default, so it does not build the Go or
>> Ada frontends.  I know it doesn't make much sense.
> I would suggest to replace currently used 'all' to 'default'. And 'all' 
> can be really used for all possible frontends we have :) I know it 
> changes the behavior, but I hope I makes sense?

I'd be in favor of that (assuming you mean "all that we have and
that are supported for the host/target combination in questions).

Gerald
Mike Stump July 28, 2014, 4:24 p.m. UTC | #5
On Jul 28, 2014, at 3:29 AM, Gerald Pfeifer <gerald@pfeifer.com> wrote:
> On Fri, 25 Jul 2014, Martin Liška wrote:
>>> Yes, --enable-languages=all is the default, so it only builds the
>>> frontends that are enabled by default, so it does not build the Go or
>>> Ada frontends.  I know it doesn't make much sense.
>> I would suggest to replace currently used 'all' to 'default'. And 'all' 
>> can be really used for all possible frontends we have :) I know it 
>> changes the behavior, but I hope I makes sense?
> 
> I'd be in favor of that (assuming you mean "all that we have and
> that are supported for the host/target combination in questions).

When I want to do all, I actually do want to do all.  I think as long as we retain a spelling for the trimmed list, (aka default or most or some spelling), I think it would be an improvement.  A second option would be a noimeanreallyall spelling to get all and leave all alone.  I don’t favor that.
diff mbox

Patch

Index: go-gcc.cc
===================================================================
--- go-gcc.cc	(revision 213019)
+++ go-gcc.cc	(working copy)
@@ -2374,8 +2374,8 @@  Gcc_backend::global_variable_set_init(Bvariable* v
 
   // If this variable goes in a unique section, it may need to go into
   // a different one now that DECL_INITIAL is set.
-  if (symtab_get_node(var_decl)
-      && symtab_get_node(var_decl)->implicit_section)
+  if (symtab_node::get(var_decl)
+      && symtab_node::get(var_decl)->implicit_section)
     {
       set_decl_section_name (var_decl, NULL);
       resolve_unique_section (var_decl,