[Caml-list] configure failure under cygwin when compiling ocaml

[Caml-list] configure failure under cygwin when compiling ocaml

[Simon Kirby]

I've used Ocaml happily under cygwin before, but just bought a new 
machine and installed the latest cygwin (1.3.19-1) and latest ocaml 
sources (3.06). I can't even run ./configure without an error:

Configuring for a i686-pc-cygwin ...
gcc found
The C compiler is ANSI-compliant.
Checking the sizes of integers and pointers...
OK, this is a regular 32 bit architecture.
64-bit "long long" integer type found (printf with "%ll").
This is a big-endian architecture.
      3 [main] tst 1104 cygheap_fixup_in_child: Couldn't reserve space 
for cygwin's heap (0x615D0000 <0x0>) in child, Win32 error 487
C:\cygwin\home\simon\src\ocaml-3.06\config\auto-aux\tst.exe: *** 
m.AllocationBase 0x615D0000, m.BaseAddress 0x615D0000, m.RegionSize 
0x5000, m.State 0x1000

Doubles must be doubleword-aligned.
      3 [main] tst 2212 cygheap_fixup_in_child: Couldn't reserve space 
for cygwin's heap (0x615D0000 <0x0>) in child, Win32 error 487
C:\cygwin\home\simon\src\ocaml-3.06\config\auto-aux\tst.exe: *** 
m.AllocationBase 0x615D0000, m.BaseAddress 0x615D0000, m.RegionSize 
0x5000, m.State 0x1000

64-bit integers must be doubleword-aligned.
Signal 8
Something went wrong while checking native division and modulus, please 
report it.

...

If I try "make world" anyway, ignoring the above error, I get a core 
dump from "pervasives.cmi".

Help!

Thanks in advance,
Simon

-- 
Simon Kirby              Language Evolution and Computation Research Unit
simon@ling.ed.ac.uk                   Theoretical and Applied Linguistics
http://www.ling.ed.ac.uk/~simon/                  University of Edinburgh

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

Re: [Caml-list] configure failure under cygwin when compiling ocaml

[Karl Zilles]

Simon Kirby wrote:

 > I've used Ocaml happily under cygwin before, but just bought a new 
machine and installed the latest cygwin (1.3.19-1) and latest ocaml 
sources (3.06). I can't even run ./configure without an error:
 >
 > Configuring for a i686-pc-cygwin ...
 > gcc found
 > The C compiler is ANSI-compliant.
 > Checking the sizes of integers and pointers...
 > OK, this is a regular 32 bit architecture.
 > 64-bit "long long" integer type found (printf with "%ll").
 > This is a big-endian architecture.


Configure thinks your machine is big-endian?  What exactly do you have 
Cygwin installed on? :)  PCs are little endian.

(This is where my configure output begins to differ from yours.  Perhaps 
you could look into the big endian/little endian test that configure is 
performing to see why it is getting so confused.)

Karl


-- 
1969
Karl Zilles | Technical Architect, Principal
t 310.566.7565 | f 310.566.7569
13101 Washington Boulevard, Suite 248, Los Angeles, CA 90293
http://www.1969.ws | zilles@1969.ws

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

Re: [Caml-list] configure failure under cygwin when compiling ocaml

[Simon Kirby]

Hmmm... well spotted! There is definitely something going badly wrong 
with the configure script. Not sure quite what to do - I manage to get 
an older install of cygwin off another machine and it compiled OK - but 
got an error on first running about "pervasives.mli". 

Not sure what to do next - at the moment, I'm sticking with the native
windows port, but since I don't have VC++ I can't compile optimised
code, which is a pain.

Has anyone else had configure errors under the latest cygwin? Can anyone 
test this?

Thanks!
Simon

On Fri, 7 Feb 2003, Karl Zilles wrote:

> Simon Kirby wrote:
> 
>  > I've used Ocaml happily under cygwin before, but just bought a new 
> machine and installed the latest cygwin (1.3.19-1) and latest ocaml 
> sources (3.06). I can't even run ./configure without an error:
>  >
>  > Configuring for a i686-pc-cygwin ...
>  > gcc found
>  > The C compiler is ANSI-compliant.
>  > Checking the sizes of integers and pointers...
>  > OK, this is a regular 32 bit architecture.
>  > 64-bit "long long" integer type found (printf with "%ll").
>  > This is a big-endian architecture.
> 
> 
> Configure thinks your machine is big-endian?  What exactly do you have 
> Cygwin installed on? :)  PCs are little endian.
> 
> (This is where my configure output begins to differ from yours.  Perhaps 
> you could look into the big endian/little endian test that configure is 
> performing to see why it is getting so confused.)
> 
> Karl
> 
> 
> 

-- 
Simon Kirby              Language Evolution and Computation Research Unit
simon@ling.ed.ac.uk                   Theoretical and Applied Linguistics
http://www.ling.ed.ac.uk/~simon/                  University of Edinburgh

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

RE: [Caml-list] configure failure under cygwin when compiling ocaml

[Alex Romadinoff]

It seams that you have broken cygwin\usr\include\sys\param.h

This file should contain lines like this:

/* All known win32 systems are little endian.  */
#define BYTE_ORDER	LITTLE_ENDIAN

I've checked Cygnus CVS
http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/winsup/cygwin/include/s
ys/?cvsroot=src and found that there are no either sys\param.h nor
sys\types.h required to check endianess.  So configure willn't work on
this version.

MinGW doesn't seam to have such a problem.

Alex

-----Original Message-----
From: Simon Kirby [mailto:simon@ling.ed.ac.uk] 
Sent: Monday, February 10, 2003 2:10 PM
To: Karl Zilles
Cc: caml-list@inria.fr
Subject: Re: [Caml-list] configure failure under cygwin when compiling
ocaml

Hmmm... well spotted! There is definitely something going badly wrong 
with the configure script. Not sure quite what to do - I manage to get 
an older install of cygwin off another machine and it compiled OK - but 
got an error on first running about "pervasives.mli". 
Not sure what to do next - at the moment, I'm sticking with the native 
windows port, but since I don't have VC++ I can't compile optimised 
code, which is a pain. 
Has anyone else had configure errors under the latest cygwin? Can anyone

test this? 
Thanks! 
Simon 
On Fri, 7 Feb 2003, Karl Zilles wrote: 
> Simon Kirby wrote: 
> 
>  > I've used Ocaml happily under cygwin before, but just bought a new 
> machine and installed the latest cygwin (1.3.19-1) and latest ocaml 
> sources (3.06). I can't even run ./configure without an error: 
>  > 
>  > Configuring for a i686-pc-cygwin ... 
>  > gcc found 
>  > The C compiler is ANSI-compliant. 
>  > Checking the sizes of integers and pointers... 
>  > OK, this is a regular 32 bit architecture. 
>  > 64-bit "long long" integer type found (printf with "%ll"). 
>  > This is a big-endian architecture. 
> 
> 
> Configure thinks your machine is big-endian?  What exactly do you have

> Cygwin installed on? :)  PCs are little endian. 
> 
> (This is where my configure output begins to differ from yours. 
Perhaps 
> you could look into the big endian/little endian test that configure
is 
> performing to see why it is getting so confused.) 
> 
> Karl 
> 
> 
> 
-- 
Simon Kirby              Language Evolution and Computation Research
Unit 
simon@ling.ed.ac.uk                   Theoretical and Applied
Linguistics 
http://www.ling.ed.ac.uk/~simon/                  University of
Edinburgh 
------------------- 
To unsubscribe, mail caml-list-request@inria.fr Archives:
http://caml.inria.fr 
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ:
http://caml.inria.fr/FAQ/ 
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners 

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

Re: [Caml-list] configure failure under cygwin when compiling ocaml

[Karl Zilles]

Alex Romadinoff wrote:
> It seams that you have broken cygwin\usr\include\sys\param.h
> 
> This file should contain lines like this:
> 
> /* All known win32 systems are little endian.  */
> #define BYTE_ORDER	LITTLE_ENDIAN
> 
> I've checked Cygnus CVS
> http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/winsup/cygwin/include/s
> ys/?cvsroot=src and found that there are no either sys\param.h nor
> sys\types.h required to check endianess.  So configure willn't work on
> this version.
> 
> MinGW doesn't seam to have such a problem.
> 
> Alex

Cygwin distributes the version from

http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/newlib/libc/sys/cygwin/sys/?cvsroot=src

This file hasn't changed in 22 months, and is included in the 
cygwin-1.3.19-1 distribution.  I've actually installed 1.3.19-1 from 
scratch and compiled ocaml successfully, so I know it does work.

(just to muddy the waters, Cygwin released 1.3.20 on Saturday)

Karl

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners