I ran tests on iPhone, all memory mapped file related tests (eg. gmbx)
failed but passed with the patch. arm in iOS works as a little endian
32-bit cpu which is the same as a 32-bit intel cpu, so I'm curious if
jmf works correctly on 32-bit Mac Intel.
Пн, 06 янв 2014, Eric Iverson писал(а):
> Bill,
>
> Regarding your questions on jmf vs mac:
>
> I haven't looked at anything now, but I can say with confidence that
> the test suites passed with all combinations of mac intel/ppc and
> 32/64 for all official Jsoftware releases. The code is right until
> someone shows a counter example.
>
> /Eric
>
> On Sat, Jan 4, 2014 at 1:44 AM, bill lam <bbill....@gmail.com> wrote:
> > While I running test suite, I noticed that the api for osx
> > needs a condition branch because offset is 64-bit even on
> > 32-bit platforms os that it has to push another 4 bytes of zero
> > to stack. But it does not look quite right to me.
> >
> > NB. ---------------------- beware of line wrap
> > if. UNAME -: 'Darwin' do.
> > c_mmap=: }:@:('mmap * * i i i i i i' api)@:(}: , (<0)"_ , {:)
> > if. ({.a.)={. 1&(3!:4) 1 do. NB. endian - powerpc vs intel
> > c_lseek=: (0 1 3 4&{)@:('lseek i i i i i' api)@:(0&{ , (<0)"_ , 1&{ ,
> > 2&{)
> > end.
> > end.
> > NB. ----------------------
> >
> > this branch is not protected by IF64 so the first question if it
> > will work for 64-bit
> >
> > c_mmap is not inside the endian test branch, so the second
> > question is will it work on intel.
> >
> > what I guess is that it should be
> >
> > NB. ----------------------
> > if. (UNAME -: 'Darwin') > IF64 do. NB. apply only to 32-bit
> > if. ({.a.)={. 1&(3!:4) 1 do. NB. endian - powerpc vs intel
> > c_mmap=: }:@:('mmap * * i i i i i i' api)@:(}: , (<0)"_ , {:)
> > c_lseek=: (0 1 3 4&{)@:('lseek i i i i i' api)@:(0&{ , (<0)"_ , 1&{ ,
> > 2&{)
> > else.
> > c_mmap=: }:@:('mmap * * i i i i i i' api)@:(}: , (<0)"_ ,~ {:)
> > c_lseek=: (0 1 3 4&{)@:('lseek i i i i i' api)@:(0&{ , (<0)"_ ,~ 1&{ ,
> > 2&{)
> > end.
> > end.
> > NB. ----------------------
> >
> > Did anyone test how jmf work on OSX (ppc vs intel, 32 vs 64)?
> >
> > for example using the test script gmbx.ijs
> >
> > NB. mapped boxed arrays ---------- > > > > load 'jmf'
> > 18!:4 <'base'
> >
> > 1 [ unmap_jmf_ 'q'
> > f=: <'q.jmf'
> > 1 [ createjmf_jmf_ f,<4e6 NB. 4e6 bytes for data
> > map_jmf_ (<'q'),f,'';0 NB. map q to jmf file
> > '' -: q
> >
> > 1 [ unmap_jmf_ 'r'
> > f1=: <'r.jmf'
> > 1 [ createjmf_jmf_ f1,<7e5
> > map_jmf_ (<'r'),f1,'';0
> > '' -: r
> >
> > g =: 6 7 4 5 2 3 0 1&{"1 @: (5&}.) @: (3!:3) @: ({."1)
> > mean=: +/ % #
> >
> > NB. ----------------------
> >
> > Thanks.
> >
> >
> > --
> > regards,
> > ====================================================
> > GPG key 1024D/4434BAB3 2008-08-24
> > gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3
> > gpg --keyserver subkeys.pgp.net --armor --export 4434BAB3
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
--
regards,
====================================================
GPG key 1024D/4434BAB3 2008-08-24
gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3
gpg --keyserver subkeys.pgp.net --armor --export 4434BAB3
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
