Ping! Re: [patch, libgfortran] PR37754 [4.4 Regression] READ I/O Performance regression from 4.3 to 4.4/4.5

[Jerry DeLisle]

Jerry DeLisle wrote:
> Jack Howarth wrote:
>> On Mon, May 25, 2009 at 10:23:15PM +0300, Janne Blomqvist wrote:
>>> Jerry DeLisle wrote:
>>>> Jerry DeLisle wrote:
>>>>> The attached patch is a backport of the I/O performance patches 
>>>>> and  associated fixes from trunk.
>>>>>
>>>>> Regression tested on x86-64.
>>>>>
>>>>> OK for 4.4.  This will close out the PR.
>>>> I am staged and ready to pull the trigger on this.  Need an 
>>>> approval. :)
>>> Ok.
>>>
>>> I'm not completely comfortable with this, it's a pretty large set of  
>>> changes. However, they seem to have worked surprisingly well on 
>>> trunk.  Or perhaps people are using trunk less now that the released 
>>> versions  are good enough for most purposes..
>>>
>>> -- 
>>> Janne Blomqvist
>>
>> Has anyone actually tried to bootstrap gcc 4.4 with the proposed patch?
>> I had a rather cryptic error from make when trying to build it against
>> Saturday's gcc 4.4 branch on x86_64-apple-darwin9. Without the proposed
>> patch, gcc 4.4 branch built fine.
>>              Jack
>>
> Please accept my apology.  I missed some hunks related to xtoa and itoa 
> functions that were moved or renamed to gfc_xtoa and gfc_itoa.
> 
> The attached patch corrects all this and is what I will commit.  I 
> regression tested and NIST tested again just to be sure.
> 
> Jerry
> 
Sending        ChangeLog
Sending        io/fbuf.c
Sending        io/file_pos.c
Sending        io/format.c
Sending        io/intrinsics.c
Sending        io/io.h
Sending        io/list_read.c
Sending        io/open.c
Sending        io/read.c
Sending        io/transfer.c
Sending        io/unit.c
Sending        io/unix.c
Sending        io/write.c
Sending        io/write_float.def
Sending        libgfortran.h
Sending        runtime/backtrace.c
Sending        runtime/error.c
Transmitting file data .................
Committed revision 147887.

The deed is done.

Jerry