Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Perl: porters

mktables gets stuck on Win32

 

 

Perl porters RSS feed   Index | Next | Previous | View Threaded


steve.m.hay at googlemail

Nov 23, 2009, 3:38 PM

Post #1 of 5 (312 views)
Permalink
mktables gets stuck on Win32

Bleadperl currently isn't building on Win32: it gets stuck on the
mktables call with CPU running at 100%.

The win32/Makefile runs this (from win32/):

cd ..\lib\unicore && ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib
-I..\..\cpan\Cwd mktables -check ..\lib\unicore\Canonical.pl
..\lib\unicore\Canonical.pl

(and normally likewise for each other file in UNIDATAFILES (?)).

I tried manually running this instead (from lib/unicore/):

..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables

but that also 'hangs' with 100% CPU, as does this (from win32/):

..\miniperl -I..\lib -I..\cpan\Cwd\lib -I..\cpan\Cwd
..\lib\unicore\mktables -C ..\lib\unicore

What's happened?


public at khwilliamson

Nov 23, 2009, 4:09 PM

Post #2 of 5 (286 views)
Permalink
Re: mktables gets stuck on Win32 [In reply to]

Steve Hay wrote:
> Bleadperl currently isn't building on Win32: it gets stuck on the
> mktables call with CPU running at 100%.
>
> The win32/Makefile runs this (from win32/):
>
> cd ..\lib\unicore && ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib
> -I..\..\cpan\Cwd mktables -check ..\lib\unicore\Canonical.pl
> ..\lib\unicore\Canonical.pl
>
> (and normally likewise for each other file in UNIDATAFILES (?)).
>
> I tried manually running this instead (from lib/unicore/):
>
> ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables
>
> but that also 'hangs' with 100% CPU, as does this (from win32/):
>
> ..\miniperl -I..\lib -I..\cpan\Cwd\lib -I..\cpan\Cwd
> ..\lib\unicore\mktables -C ..\lib\unicore
>
> What's happened?
>
There's a new mktables, which takes more real-time. There are also
Makefile patches to call it. I just submitted a revised patch for
those; perhaps these will fix it. I don't know how such changes get
propagated to other platforms' Makefiles. I did test mktables on
Windows, but I don't have a development environment set up, I just
manually used an ActiveState perl on it. You could manually add the -v
option and email the output. And you could try running a regular perl
on it to see if that helps; it runs slower on miniperl.


SteveHay at planit

Nov 24, 2009, 12:29 AM

Post #3 of 5 (286 views)
Permalink
RE: mktables gets stuck on Win32 [In reply to]

karl williamson wrote on 2009-11-24:
> Steve Hay wrote:
>> Bleadperl currently isn't building on Win32: it gets stuck on the
>> mktables call with CPU running at 100%.
>>
>> The win32/Makefile runs this (from win32/):
>>
>> cd ..\lib\unicore && ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib
>> -I..\..\cpan\Cwd mktables -check ..\lib\unicore\Canonical.pl
>> ..\lib\unicore\Canonical.pl
>>
>> (and normally likewise for each other file in UNIDATAFILES (?)).
>>
>> I tried manually running this instead (from lib/unicore/):
>>
>> ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd
mktables
>>
>> but that also 'hangs' with 100% CPU, as does this (from win32/):
>>
>> ..\miniperl -I..\lib -I..\cpan\Cwd\lib -I..\cpan\Cwd
>> ..\lib\unicore\mktables -C ..\lib\unicore
>>
>> What's happened?
>>
> There's a new mktables, which takes more real-time. There are also
> Makefile patches to call it. I just submitted a revised patch for
> those; perhaps these will fix it. I don't know how such changes get
> propagated to other platforms' Makefiles. I did test mktables on
> Windows, but I don't have a development environment set up, I just
> manually used an ActiveState perl on it. You could manually add the
-v
> option and email the output. And you could try running a regular perl
> on it to see if that helps; it runs slower on miniperl.

Just tried again on my (much faster) machine at work, and it does run,
but takes over a minute even here. I don't know how long it would have
taken on my slower machine at home--I gave up waiting. Obviously I need
to leave it longer, but that's a real pain.

It's not happy, though. Here's the output from my work machine, where it
just ran:

cd ..\lib\unicore && ..\..\miniperl.exe -I..
-I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
..\lib\unicore\Canonical.pl ..\lib\unicore\Canonical.pl

mktables: Failed to open 'mktables.lst', turning on -globlist option
instead:
No such file or directory; at mktables line 13483
Use of uninitialized value $file in pattern match (m//) at
..\..\cpan\Cwd\lib/File/Spec/Win32.pm line 113.
Use of uninitialized value $file in pattern match (m//) at
..\..\cpan\Cwd\lib/File/Spec/Win32.pm line 119.
Use of uninitialized value $path in pattern match (m//) at
..\..\cpan\Cwd\lib/File/Spec/Win32.pm line 212.
cd ..\lib\unicore && ..\..\miniperl.exe -I..
-I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
..\lib\unicore\Exact.pl ..\lib\unicore\Canonical.pl
Skipping as check params are not the same.
cd ..\lib\unicore && ..\..\miniperl.exe -I..
-I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
..\lib\unicore\Properties ..\lib\unicore\Canonical.pl
Skipping as check params are not the same.
cd ..\lib\unicore && ..\..\miniperl.exe -I..
-I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
..\lib\unicore\PVA.pl ..\lib\unicore\Canonical.pl
Skipping as check params are not the same.


SteveHay at planit

Nov 24, 2009, 1:26 AM

Post #4 of 5 (279 views)
Permalink
RE: mktables gets stuck on Win32 [In reply to]

Steve Hay wrote on 2009-11-24:
> karl williamson wrote on 2009-11-24:
>> Steve Hay wrote:
>>> Bleadperl currently isn't building on Win32: it gets stuck on the
>>> mktables call with CPU running at 100%.
>>>
>>> The win32/Makefile runs this (from win32/):
>>>
>>> cd ..\lib\unicore && ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib
>>> -I..\..\cpan\Cwd mktables -check ..\lib\unicore\Canonical.pl
>>> ..\lib\unicore\Canonical.pl
>>>
>>> (and normally likewise for each other file in UNIDATAFILES (?)).
>>>
>>> I tried manually running this instead (from lib/unicore/):
>>>
>>> ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd
mktables
>>>
>>> but that also 'hangs' with 100% CPU, as does this (from win32/):
>>>
>>> ..\miniperl -I..\lib -I..\cpan\Cwd\lib -I..\cpan\Cwd
>>> ..\lib\unicore\mktables -C ..\lib\unicore
>>>
>>> What's happened?
>>>
>> There's a new mktables, which takes more real-time. There are also
>> Makefile patches to call it. I just submitted a revised patch for
>> those; perhaps these will fix it. I don't know how such changes get
>> propagated to other platforms' Makefiles. I did test mktables on
>> Windows, but I don't have a development environment set up, I just
>> manually used an ActiveState perl on it. You could manually add the
-v
>> option and email the output. And you could try running a regular
perl
>> on it to see if that helps; it runs slower on miniperl.
>
> Just tried again on my (much faster) machine at work, and it does run,
> but takes over a minute even here. I don't know how long it would have
> taken on my slower machine at home--I gave up waiting. Obviously I
need
> to leave it longer, but that's a real pain.
>
> It's not happy, though. Here's the output from my work machine, where
it
> just ran:
>

Following 36ff7f95732aeec6ca1f7152b75dbedf5ce669d1 the output is now
improved, but I still see this:


cd ..\lib\unicore && ..\..\miniperl.exe -I..
-I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
..\lib\unicore\Decomposition.pl ..\lib\unicore\Decomposition.pl

mktables: Failed to open 'mktables.lst', turning on -globlist option
instead:
No such file or directory; at mktables line 13478


Is that expected?


rgs at consttype

Nov 24, 2009, 1:36 AM

Post #5 of 5 (277 views)
Permalink
Re: mktables gets stuck on Win32 [In reply to]

2009/11/24 Steve Hay <SteveHay [at] planit>:
> Steve Hay wrote on 2009-11-24:
>> karl williamson wrote on 2009-11-24:
>>> Steve Hay wrote:
>>>> Bleadperl currently isn't building on Win32: it gets stuck on the
>>>> mktables call with CPU running at 100%.
>>>>
>>>> The win32/Makefile runs this (from win32/):
>>>>
>>>> cd ..\lib\unicore &&  ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib
>>>> -I..\..\cpan\Cwd mktables -check ..\lib\unicore\Canonical.pl
>>>> ..\lib\unicore\Canonical.pl
>>>>
>>>> (and normally likewise for each other file in UNIDATAFILES (?)).
>>>>
>>>> I tried manually running this instead (from lib/unicore/):
>>>>
>>>> ..\..\miniperl.exe -I.. -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd
> mktables
>>>>
>>>> but that also 'hangs' with 100% CPU, as does this (from win32/):
>>>>
>>>> ..\miniperl -I..\lib -I..\cpan\Cwd\lib -I..\cpan\Cwd
>>>> ..\lib\unicore\mktables -C ..\lib\unicore
>>>>
>>>> What's happened?
>>>>
>>> There's a new mktables, which takes more real-time.  There are also
>>> Makefile patches to call it.  I just submitted a revised patch for
>>> those; perhaps these will fix it.  I don't know how such changes get
>>> propagated to other platforms' Makefiles.  I did test mktables on
>>> Windows, but I don't have a development environment set up, I just
>>> manually used an ActiveState perl on it.  You could manually add the
> -v
>>> option and email the output.  And you could try running a regular
> perl
>>> on it to see if that helps; it runs slower on miniperl.
>>
>> Just tried again on my (much faster) machine at work, and it does run,
>> but takes over a minute even here. I don't know how long it would have
>> taken on my slower machine at home--I gave up waiting. Obviously I
> need
>> to leave it longer, but that's a real pain.
>>
>> It's not happy, though. Here's the output from my work machine, where
> it
>> just ran:
>>
>
> Following 36ff7f95732aeec6ca1f7152b75dbedf5ce669d1 the output is now
> improved, but I still see this:
>
>
>        cd ..\lib\unicore &&  ..\..\miniperl.exe -I..
> -I..\..\cpan\Cwd\lib -I..\..\cpan\Cwd mktables -check
> ..\lib\unicore\Decomposition.pl ..\lib\unicore\Decomposition.pl
>
> mktables: Failed to open 'mktables.lst', turning on -globlist option
> instead:
>    No such file or directory; at mktables line 13478

That's expected, since mktables.lst is generated, the first time.

Perl porters RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.