[m-dev.] errors in Cygwin build 10.04.2

Robert Shiplett grshiplett at gmail.com
Wed Oct 13 01:28:37 AEDT 2010


I am watching the configure run - it complains that this new Cygwin
1.7+ does not have a recent enough jdk ( a complaint that halted my
Eralng build after getting WxWidgets -X11 to build

It complains of no /cgi-bin DIR in usr/lib  for a profiler install

but
  sh configure
itself appears to run clean and ends with no wanrnings.

It this make fails to install, I will try setting to $INSTALL to unset
( just now it has a value == path of install.exe)

...

Sad to say, but build fails again even after following your rm
instructions - is the problem in a makefile.in ??

I will start from scratch but unset that env INSTALL var

Here is the tail of make - note that -m 644 is being sent to ../../cp
which I think would fail even if cp accepted -m ( cp does not reside
there - but why `cp instead of install  ? )

 /bin/sh /cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/install-sh
-c -m 644 'sunc/x86.h
gdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-install/include/atomic_ops/sysdeps/sunc/x86.
 /bin/sh /cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/install-sh
-c -m 644 'sunc/x86_6
/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-install/include/atomic_ops/sysdeps/sunc/x
h'
make[8]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_op
eps'
make[7]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_op
eps'
make[7]: Entering directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_o
make[8]: Entering directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_o
make[8]: Nothing to be done for `install-exec-am'.
test -z "/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-install/include/atomic_ops/"
||
in/mkdir -p "/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-install/include/atomic_ops/"
 ../../cp -m 644 'generalize.h'
'/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-install/
e/atomic_ops//generalize.h'
/bin/sh: line 4: ../../cp: No such file or directory
 ../../cp -m 644 'generalize-small.h'
'/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops-in
include/atomic_ops//generalize-small.h'
/bin/sh: line 4: ../../cp: No such file or directory
make[8]: *** [install-privateHEADERS] Error 127
make[8]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_op
make[7]: *** [install-am] Error 2
make[7]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_op
make[6]: *** [install-recursive] Error 1
make[6]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src/atomic_op
make[5]: *** [install-recursive] Error 1
make[5]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops/src'
make[4]: *** [install-recursive] Error 1
make[4]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc/libatomic_ops'
make[3]: *** [libatomic_ops-install] Error 2
make[3]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc'
make[2]: *** [submake] Error 2
make[2]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2/boehm_gc'
make[1]: *** [boehm_gc] Error 2
make[1]: Leaving directory
`/cygdrive/i/cygwin/mercury-10.04.2/mercury-compiler-10.04.2'
make: *** [all] Error 2


On 12 October 2010 10:27, Robert Shiplett <grshiplett at gmail.com> wrote:
> I will try what you advise, because while correcting the odd value for
> $INSTALL was enought to prevent breaking the WxWidgets for Erlang
> install, it was not enough to get boehm_gc  to build (once again -m
> 644 being sent to cp  instead of to install )
>
> I should be able to get back within the hour ( I just endured a 9am
> job interview so I do need some quality computer-interaction time just
> now  ;-)
>
> R
>
> On 12 October 2010 04:59, Julien Fischer <juliensf at csse.unimelb.edu.au> wrote:
>>
>> Hi,
>>
>> On Tue, 12 Oct 2010, Julien Fischer wrote:
>>
>>> On Tue, 12 Oct 2010, Peter Ross wrote:
>>>
>>>> On Tue, Oct 12, 2010 at 4:21 AM, Robert Shiplett <grshiplett at gmail.com>
>>>> wrote:
>>>>>
>>>>> after
>>>>>  sh configure
>>>>>  make
>>>>>
>>>> I'm testing now to see if I can reproduce the same error.
>>>
>>> I've tried on Windows XP SP3 with Cygwin 1.7.7** and installation proceeds
>>> normally.  (I've just posted a fix for a problem with the boehm_gc
>>> directory in the source distribution that may fix the problem.)
>>>
>>> ** I'm also trying the same version of Cygwin on Windows 7.
>>
>> Couldn't reproduce the problem on Windows 7 either.
>>
>> Robert, could you please try again with a freshly unpacked 10.04.2
>> tarball and before doing configure; make etc do the following:
>>
>>   cd boehm_gc/libatomic_ops && make clean
>>
>> And then delete the following files:
>>
>>    boehm_gc/libatomic_ops/Makefile
>>    boehm_gc/libatomic_ops/tests/Makefile
>>    boehm_gc/libatomic_ops/src/Makefile
>>    boehm_gc/libatomic_ops/src/atomic_ops/Makefile
>>    boehm_gc/libatomic_ops/src/config.h
>>    boehm_gc/libatomic_ops/src/atomic_ops/sysdeps/Makefile
>>
>> The configure and make as-per normal from the level.
>>
>> Does that work?
>>
>> Julien.
>

--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at csse.unimelb.edu.au
Administrative Queries: owner-mercury-developers at csse.unimelb.edu.au
Subscriptions:          mercury-developers-request at csse.unimelb.edu.au
--------------------------------------------------------------------------



More information about the developers mailing list