This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Merge Ignore and Deprecated in .opt files.


On 08/18/2018 12:24 AM, Segher Boessenkool wrote:
> Hi!
> 
> On Thu, Aug 16, 2018 at 11:18:15AM +0200, Martin Liška wrote:
>> On 08/15/2018 06:38 PM, Joseph Myers wrote:
>>> On Wed, 15 Aug 2018, Martin Liška wrote:
>>>
>>>> Ok, so you have very similar opinion as Jakub. Thus I'm sending new 
>>>> version that preserves status quo, it only does:
>>>
>>> This is removing RejectNegative from some Deprecated options.  Won't that 
>>> result in the -fno-* variants of those options starting to be accepted, 
>>> when they never were accepted when the positive versions of the options 
>>> did something useful?
>>>
>>
>> That's not intended, I fixed that. It the patch acceptable in form in which it is?
>>
>> Thanks,
>> Martin
>>
> 
>> >From 0a7d5cd6cd6ca0586a350b95cd8f6ded095ba9c8 Mon Sep 17 00:00:00 2001
>> From: marxin <mliska@suse.cz>
>> Date: Wed, 18 Jul 2018 13:40:24 +0200
>> Subject: [PATCH] Merge Ignore and Deprecated in .opt files.
>>
>> gcc/ChangeLog:
>>
>> 2018-07-18  Martin Liska  <mliska@suse.cz>
>>
>> 	* common.opt: Remove Warn, Init and Report for options with
>>         Ignore/Deprecated flag. Warning is done automatically for
>>         Deprecated flags.
> 
> Too much indent.  Two spaces after a full stop.

Thanks.

> 
> Removing Init is *wrong* as far as I see; it changes things, anyway.
> Could you not have done this as a separate patch?

It's already in, but it should be fine. Note that I added check into
opt-functions.awk where I can't use Var with Deprecated or Ignore attribute.
Thus removal of Init should be fine. Or am I wrong?

> 
>> 	* config/rs6000/rs6000.opt: Likewise.
> 
>> diff --git a/gcc/config/rs6000/rs6000.opt b/gcc/config/rs6000/rs6000.opt
>> index 25a4883b161..b07f7f7e833 100644
>> --- a/gcc/config/rs6000/rs6000.opt
>> +++ b/gcc/config/rs6000/rs6000.opt
>> @@ -483,8 +483,9 @@ mcrypto
>>  Target Report Mask(CRYPTO) Var(rs6000_isa_flags)
>>  Use ISA 2.07 Category:Vector.AES and Category:Vector.SHA2 instructions.
>>  
>> +; We can't use Ignore flag because DIRECT_MOVE mask is still used.
>>  mdirect-move
>> -Target Undocumented Mask(DIRECT_MOVE) Var(rs6000_isa_flags) Ignore Warn(%qs is deprecated)
>> +Target Undocumented Mask(DIRECT_MOVE) Var(rs6000_isa_flags) Warn(%qs is deprecated)
> 
> This is not described in the changelog.
> 
> I don't understand what it means either; did you change the semantics of
> the "Ignore" flag?  It worked just fine before, and I don't know if it
> still does :-/
> 
> [ Please cc: the rs6000 maintainers when you change rs6000 code.  Thanks! ].

Sorry for that I forgot to add you to comment about. It's again related to sanity check
of Ignore attribute. This case is bit problematic, so I tend to revert it and remove
sanity check about 'Warn' attribute of an optioned when combined with 'Ignore'.

Martin

> 
> 
> Segher
> 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]