Menu
Search

Forums

Forum Index

Thread overview
[phobos] phobos commit, revision 2349
Jan 19, 2011
dsource.org
Jan 19, 2011
Brad Roberts
Jan 19, 2011
Brad Roberts
Jan 20, 2011
Brad Roberts
January 19, 2011
Gravatar of dsource.orgPosted by dsource.orgReply
dsource.org
Gravatar of dsource.org


Reply
phobos commit, revision 2349


user: andrei

msg:
Moved split from string to array, made one more pass through array

http://www.dsource.org/projects/phobos/changeset/2349

paths changed:
U   trunk/phobos/std/algorithm.d
U   trunk/phobos/std/array.d
U   trunk/phobos/std/getopt.d
U   trunk/phobos/std/range.d
U   trunk/phobos/std/string.d
January 18, 2011
Gravatar of Brad RobertsPosted by Brad Roberts
in reply to dsource.org		Reply
Brad Roberts
Gravatar of Brad Roberts
Posted in reply to dsource.org


Reply
On 1/18/2011 10:44 PM, dsource.org wrote:
> phobos commit, revision 2349
> 
> 
> user: andrei
> 
> msg:
> Moved split from string to array, made one more pass through array
> 
> http://www.dsource.org/projects/phobos/changeset/2349
> 
> paths changed:
> U   trunk/phobos/std/algorithm.d
> U   trunk/phobos/std/array.d
> U   trunk/phobos/std/getopt.d
> U   trunk/phobos/std/range.d
> U   trunk/phobos/std/string.d

Looks like this change set has exposed a new 64 bit code gen bug.  32 bit linux looks ok.  I'll narrow it down.

Later,
Brad
January 19, 2011
Gravatar of Brad RobertsPosted by Brad Roberts
in reply to Brad Roberts		Reply
Brad Roberts
Gravatar of Brad Roberts
Posted in reply to Brad Roberts


Reply
On 1/18/2011 11:52 PM, Brad Roberts wrote:
> On 1/18/2011 10:44 PM, dsource.org wrote:
>> phobos commit, revision 2349
>>
>>
>> user: andrei
>>
>> msg:
>> Moved split from string to array, made one more pass through array
>>
>> http://www.dsource.org/projects/phobos/changeset/2349
>>
>> paths changed:
>> U   trunk/phobos/std/algorithm.d
>> U   trunk/phobos/std/array.d
>> U   trunk/phobos/std/getopt.d
>> U   trunk/phobos/std/range.d
>> U   trunk/phobos/std/string.d
> 
> Looks like this change set has exposed a new 64 bit code gen bug.  32 bit linux looks ok.  I'll narrow it down.
> 
> Later,
> Brad

Still reducing.. the template heavy, highly static if overloaded code makes for painful reduction.  I'm pretty sure that it has to do with std.utf.decode(string, int).

If someone wants to pick up where I left off before I can get back to it tomorrow night:

  http://d.puremagic.com/bug-opt-array.d

Later,
Brad
January 20, 2011
Gravatar of Brad RobertsPosted by Brad Roberts
in reply to Brad Roberts		Reply
Brad Roberts
Gravatar of Brad Roberts
Posted in reply to Brad Roberts


Reply
I got this one reduced down to a nice small test case and handed it off to Walter to fix.  Hopefully we'll get the tests all back to passing shortly.

On 1/19/2011 2:27 AM, Brad Roberts wrote:
> On 1/18/2011 11:52 PM, Brad Roberts wrote:
>> On 1/18/2011 10:44 PM, dsource.org wrote:
>>> phobos commit, revision 2349
>>>
>>>
>>> user: andrei
>>>
>>> msg:
>>> Moved split from string to array, made one more pass through array
>>>
>>> http://www.dsource.org/projects/phobos/changeset/2349
>>>
>>> paths changed:
>>> U   trunk/phobos/std/algorithm.d
>>> U   trunk/phobos/std/array.d
>>> U   trunk/phobos/std/getopt.d
>>> U   trunk/phobos/std/range.d
>>> U   trunk/phobos/std/string.d
>>
>> Looks like this change set has exposed a new 64 bit code gen bug.  32 bit linux looks ok.  I'll narrow it down.
>>
>> Later,
>> Brad
> 
> Still reducing.. the template heavy, highly static if overloaded code makes for painful reduction.  I'm pretty sure that it has to do with std.utf.decode(string, int).
> 
> If someone wants to pick up where I left off before I can get back to it tomorrow night:
> 
>   http://d.puremagic.com/bug-opt-array.d
> 
> Later,
> Brad
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos