Thanks, now I see something useful:

$ echo foo | ./munge.exe

The exported object file has version 5.71 but this library supports 5.70

(then it hangs)

I have Poly/ML 5.7 installed in /usr/local (by Homebrew on Mac OS X 10.11), and 
the new 5.7.1 manually installed into $HOME.  So I guess, when building 
`munge.exe` it must have been linked with libpolyml.a (or libpolymain.a) in 
/usr/local/lib because I didn’t put $HOME/lib into DYLD_LIBRARY_PATH 
(LD_LIBRARY_PATH on Linux).

After completely removing Poly/ML 5.7 and rebuilding and installing 5.7.1 into 
/usr/local, the problem is resolved. Now munge.exe works correctly on my side.

P. S. but multiple copies of Poly/ML actually worked well in all other cases, I 
normally use Poly/ML 5.6 (installed into a dedicated directory) to build HOL4 
K11 release but never needed to build and run munge.exe from that version.

Any way, sorry for disturbing.

Regards,

Chun

> Il giorno 21 nov 2017, alle ore 12:18, [email protected] ha 
> scritto:
> 
> Those that passed are not tests of the munger per se: they are building the 
> necessary background theories, and the munger executables. Those tests that 
> you are trying to run are actually trying to execute the mungers.
> 
> If you test a munger interactively with
> 
>    $ echo foo | ./munge.exe
> 
> what do you see?
> 
> Michael
> 
> On 21/11/17, 19:52, "Chun Tian" <[email protected]> wrote:
> 
>    Hi,
> 
>    I rebuilt again with -t option. By default Holmake have 4 parallel 
> threads, now all are blocked by munge tests:
> 
>      501 60326 60299   0  9:48am ttys000    0:00.00 /bin/sh -c ./tymunge.exe 
> < tyabb_input > tyabb_output
>      501 60327 60326   0  9:48am ttys000    0:00.00 ./tymunge.exe
>      501 60330 60299   0  9:48am ttys000    0:00.00 /bin/sh -c ./munge.exe < 
> input > output
>      501 60331 60330   0  9:48am ttys000    0:00.00 ./munge.exe
>      501 60332 60299   0  9:48am ttys000    0:00.00 /bin/sh -c ./utymunge.exe 
> < tyabb_input > utyabb_output
>      501 60333 60332   0  9:48am ttys000    0:00.00 ./utymunge.exe
>      501 60334 60299   0  9:48am ttys000    0:00.00 /bin/sh -c ./munge.exe < 
> gh242_input > gh242_output
>      501 60335 60334   0  9:48am ttys000    0:00.00 ./munge.exe
> 
>    However some munge tests indeed passed:
> 
>    Building directory src/TeX/theory_tests [21 Nov, 09:48:01]
>    Recursively calling Holmake in /Users/binghe/ML/HOL.stdknl/tools/cmp
>    Finished recursive invocation in /Users/binghe/ML/HOL.stdknl/tools/cmp
>    tyabbrevTheory                                                             
>   OK
>    foo248Theory                                                               
>   OK
>    bag248Theory                                                               
>   OK
>    mdtTheory                                                                  
>   OK
>    untyabbrevTheory                                                           
>   OK
>    pp248Theory                                                                
>   OK
>    tymunge.exe                                                                
>   OK
>    munge.exe                                                                  
>   OK
>    utymunge.exe                                                               
>   OK
>    munge248.exe                                                               
>   OK
> 
>    Regards,
> 
>    Chun Tian
> 
>> Il giorno 21 nov 2017, alle ore 03:57, [email protected] ha 
>> scritto:
>> 
>> Does HOL build with the -t option?
>> 
>> This does some testing of the munging technology.
>> 
>> Michael
>> 
>> On 21/11/17, 10:03, "polyml on behalf of Chun Tian" 
>> <[email protected] on behalf of [email protected]> wrote:
>> 
>>   Hi Ramana,
>> 
>>   I found that “munge.exe” (generated from theories) doesn’t work with this 
>> new Poly release: it simply hangs when converting *.htex to *.tex files … 
>> and when I switched back to poly 5.6, it worked again.
>> 
>>   Can you confirm this on your side?
>> 
>>   —Chun
>> 
>>> Il giorno 20 nov 2017, alle ore 22:58, Ramana Kumar 
>>> <[email protected]> ha scritto:
>>> 
>>> The 5.7.1 candidate appears to work for HOL4 and CakeML - I haven't done 
>>> extensive tests, but normal builds seem fine.
>>> 
>>> On 21 November 2017 at 08:41, Makarius <[email protected]> wrote:
>>> On 20/11/17 14:23, David Matthews wrote:
>>>> Thanks to everyone who sent in bug reports.  We're almost ready to
>>>> release 5.7.1.  I've updated Git master ( 44b7b88 ) with pre-built
>>>> compilers for 5.7.1.  This is the last chance check it before the
>>>> release which will probably happen at the end of the week.
>>> 
>>> I have successfully built this version on Linux, Windows, mac OS, and
>>> ran vorious manual tests of Isabelle + AFP. It all looks fine.
>>> 
>>> Note that I will be on travel on Wed--Fri this week, so if there are any
>>> last-minute issues, I cannot do any tests during these days.
>>> 
>>> 
>>>       Makarius
>>> 
>>> _______________________________________________
>>> polyml mailing list
>>> [email protected]
>>> http://lists.inf.ed.ac.uk/mailman/listinfo/polyml
>>> 
>>> _______________________________________________
>>> polyml mailing list
>>> [email protected]
>>> http://lists.inf.ed.ac.uk/mailman/listinfo/polyml
>> 
>> 
>> 
> 
> 
> 

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
polyml mailing list
[email protected]
http://lists.inf.ed.ac.uk/mailman/listinfo/polyml

Reply via email to