Re: [Gcl-devel] 2.6.9

[Donald Winiecki]

While I'm not privvy to the historical specifics and context of GCL (that is, KCL), your conjecture about page and heap limits seems sensible if we consider that KCL was developed when PCs were still affected by MS-DOS and IBM-DOS limits.  (Okay, Win-Whatever is still affected by those things, but arguably much less so, and Win7 has some substantive changes that make for more smiles deep down inside...)

Camm, have you committed the changes mentioned earlier such that I could safely test things on a Windows machine?

Best,

_don

On Mon, Jun 10, 2013 at 2:28 PM, Camm Maguire <address@hidden> wrote:

Greetings!  OK, I've got wine back up, basically, and have a working
build which I will commit shortly.  I'll give you a shout when done.

Just wanted to say now that I've noticed the win build appears limited
to 65k pages or about 256M heap.  Does this seem right to you?  Is this
to preserve legacy memory compatibility?

Here is roughly what I see, test area at ~ #x4d0000, C stack at ~
#x990000, heap start hardcoded at #x10100000, VirtualAlloc will max out
at a heap end of #x20000000.

Insights appreciated!

Donald Winiecki <address@hidden> writes:

> Hi Camm,
>
> Still no joy.  Attached are the CONFIGURE.LOG and MAKE.LOG files.
>
> Not sure why an amd64 processor should affect Wine -- though Wine has pretty hard limits.  I have an Ubuntu setup with VirtualBox and Windows XP on
> it and it behaves well -- sometimes better than the machine running WinXP as a native OS!
>
> Best,
>
> _don
>
> On Fri, Jun 7, 2013 at 4:07 PM, Camm Maguire <address@hidden> wrote:
>
>     Greetings!
>
>     Donald Winiecki <address@hidden> writes:
>
>     > Camm, Nico and Richard,
>     >
>     > Since CLTL1 is the usual and WinXP is the OS version about which most Win32 questions appear to be focused on at present, I started there.
>     >
>     > Included immediately below is a link to a compressed file containing the successful build of gcl-269pre.  The CLTL1 Win32 executable is /
>     > gcl-268pre-Win32-cltl1-06062013/unixport/saved_gcl.exe
>     >
>     > http://dl.dropboxusercontent.com/u/18163692/gcl-268pre-Win32-cltl1-06062013.zip
>     >
>
>     Thanks so much again!
>
>     > ​gcl-269pre fails in MAKE.  I'm attaching logs for CONFIGURE, MAKE and INSTALL on this attempt to build it.  Camm, please signal when you've
>     got
>     > the brk emulation ready and I'll give it a go​.
>     >
>
>     OK, if you could try again now, that would be most helpful.
>
>     I've shaken out a few bugs, and built on the mac.  acl2/maxima
>     regression passes.
>
>     Alas, now that I have an amd64 laptop, my wine setup is not working.
>     VirtualAlloc is returning 0.  Will try to debug this, but perhaps your
>     useful configure and make logs would be more effective at this juncture.
>
>     I'm uploading 2.6.9 to the debian autobuilders shortly for a test.
>
>     > ​(I'm not currently fit to ​look at the logs except to note errors at the end of MAKE (and of course, INSTALL).  We got off the plane from a
>     stay
>     > in Seoul only a few hours ago and while my body thinks it's 15 hours in the future and I'm not able to sleep, I'm not ready for the level of
>     effort
>     > required to do a proper analysis either.)
>     >
>
>     Take it easy!  Only as time and interest permit.
>
>     Take care,
>
>
>

--
Camm Maguire                                        address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah