Re: [ft] freetype2 and Win64

[mpsuzuki]

I'm quite sorry for my long long silence.

A) bdflib.c

commit 704f4d75cdfae502cc0b70e3413c26ad996b1ac2
Author: suzuki toshiya <address@hidden>
Date:   Sun Sep 13 00:50:14 2009 +0900

    [BDF] Modify hash API to take size_t value instead of void *.

B) t1decode.c & cffgload.c

commit 0729bc9f5223aeefa6d0bf8573a2ba0634cc1ca6
Author: suzuki toshiya <address@hidden>
Date:   Thu Sep 10 16:09:55 2009 +0900

    [Win64] Improve the computation of random seed from stack address.

C) ftcbasic.c & ftccmap.c

It is not fixed yet. I will try to fix within this
month.

# During the consideration of this issue, other
# cache vs optimizer bug was found and I slipped
# to return this original task.

Yet I don't know serious requirement of 64-bit
hash in FreeType cache. I will insert some hooks
to prevent unexpected overflows by interplaying
with 64-bit pointer and 32-bit integers.

On Fri, 8 Oct 2010 08:58:33 -0400
NightStrike <address@hidden> wrote:

>Ping?
>
>On Thu, Jan 28, 2010 at 4:22 PM, NightStrike <address@hidden> wrote:
>>> Hi,
>>>
>>> On Thu, 10 Sep 2009 16:32:55 +0900
>>> address@hidden wrote:
>>>>On Mon, 7 Sep 2009 13:02:28 +0900
>>>>address@hidden wrote:
>>>>>There are 3 groups of the problem.
>>>>>
>>>>>A) bdflib.c
>>>>
>>>>>B) t1decode.c & cffgload.c
>>>>
>>>>>C) ftcbasic.c & ftccmap.c
>>>>
>>>>Just I've fixed B) in git. Soon A) will be fixed.
>>>
>>> Just I've fixed A) in git, but by completely different way.
>>
>> but I've been off the radar since mid september due to personal issues
>> offline.
>>