Bug #61286 for ExtUtils-MakeMaker: generated PM_TO_BLIB can be too long for ActivePerl

Sun Sep 12 05:51:15 2010 MITHALDU [...] cpan.org - Ticket created

Subject:

generated PM_TO_BLIB can be too long for ActivePerl

Under windows, ActivePerl installs dmake by default, which apparently cannot handle long lines as well as unix makes. Specifically, when trying to handle DateTime-Timezone MM generates this file: http://gist.github.com/575964 And trying to process that, dmake stumbles because PM_TO_BLIB is over 32 kbytes. D:\DateTime-TimeZone-1.21>dmake dmake: makefile: line 1316: Error: -- Input line too long, increase MAXLINELENGTH Can this be fixed by special-casing the MAXLINELENGTH macro for dmake, splitting PM_TO_BLIB if it becomes too long or other methods?

Sun Sep 12 18:46:41 2010 schwern [...] pobox.com - Correspondence added

Subject:	Re: [rt.cpan.org #61286] generated PM_TO_BLIB can be too long for ActivePerl
Date:	Sun, 12 Sep 2010 15:46:27 -0700
To:	bug-ExtUtils-MakeMaker [...] rt.cpan.org
From:	Michael G Schwern <schwern [...] pobox.com>

On 2010.9.12 2:51 AM, Christian Walde via RT wrote: Show quoted text

> Under windows, ActivePerl installs dmake by default, which apparently > cannot handle long lines as well as unix makes. > > Specifically, when trying to handle DateTime-Timezone MM generates this > file: http://gist.github.com/575964 > > And trying to process that, dmake stumbles because PM_TO_BLIB is over 32 > kbytes. > > D:\DateTime-TimeZone-1.21>dmake > dmake: makefile: line 1316: Error: -- Input line too long, increase > MAXLINELENGTH

Gah. Thanks for reporting that. Show quoted text

> Can this be fixed by special-casing the MAXLINELENGTH macro for dmake, > splitting PM_TO_BLIB if it becomes too long or other methods?

If you put MAXLINELENGTH = 65536 at the top of the Makefile it works? Would you please try: http://github.com/Perl-Toolchain-Gang/Extutils-MakeMaker/zipball/dmake_maxlinelength -- Life is like a sewer - what you get out of it depends on what you put into it. - Tom Lehrer

Sun Sep 12 18:46:42 2010 The RT System itself - Status changed from 'new' to 'open'

Sun Sep 12 19:12:18 2010 MITHALDU [...] cpan.org - Correspondence added

On Sun Sep 12 18:46:41 2010, schwern@pobox.com wrote: Show quoted text

> Gah. Thanks for reporting that.

No need to thank me. I'm currently on a crusade to make ActivePerl a bit nicer to use. So, thanks to you for answering so quickly. :) Show quoted text

> If you put MAXLINELENGTH = 65536 at the top of the Makefile it works?

If i edit the file manually, DateTime-Timezone dmakes and tests fine. However: Show quoted text

> Would you please try: > http://github.com/Perl-Toolchain-Gang/Extutils- > MakeMaker/zipball/dmake_maxlinelength

This breaks. I tried installing this and got a linelength error on its Makefile. Turns out it inserted: # Get dmake to read long commands like PM_TO_BLIB MAXLINELENGTH = 1.#INF So, once you get it to print an actual number, it will work fine. :)

Sun Sep 12 19:36:11 2010 schwern [...] pobox.com - Correspondence added

Subject:	Re: [rt.cpan.org #61286] generated PM_TO_BLIB can be too long for ActivePerl
Date:	Sun, 12 Sep 2010 16:35:50 -0700
To:	bug-ExtUtils-MakeMaker [...] rt.cpan.org
From:	Michael G Schwern <schwern [...] pobox.com>

On 2010.9.12 4:12 PM, Christian Walde via RT wrote: Show quoted text

> No need to thank me. I'm currently on a crusade to make ActivePerl a > bit nicer to use. So, thanks to you for answering so quickly. :)

I've reported the problem upstream to dmake suggesting they eliminate the MAXLINELENGTH concept entirely. It doesn't look too bad to do in the code. Show quoted text

> However:

>> Would you please try: >> http://github.com/Perl-Toolchain-Gang/Extutils- >> MakeMaker/zipball/dmake_maxlinelength

> > This breaks. I tried installing this and got a linelength error on its > Makefile. Turns out it inserted: > > # Get dmake to read long commands like PM_TO_BLIB > MAXLINELENGTH = 1.#INF > > So, once you get it to print an actual number, it will work fine. :)

I was asking for a mere 64 to the 1024th power bytes of memory. Please try it again. -- I'm pale as Formica, social skills stunted small. But I'm accurate like a pica, I know the capital of Nepal. I'm the nemesis of error, dreadful diction fears my skills, more inquisitive than Jim Lehrer, snottier than Beverly Hills. -- I.L.O.P. Secret Rap http://goats.com/archive/020830.html

Sun Sep 12 20:27:53 2010 MITHALDU [...] cpan.org - Correspondence added

On Sun Sep 12 19:36:11 2010, schwern@pobox.com wrote: Show quoted text

> I've reported the problem upstream to dmake suggesting they eliminate

the Show quoted text

> MAXLINELENGTH concept entirely. It doesn't look too bad to do in the

code. That is very nice of you. Thank you. :) Show quoted text

> I was asking for a mere 64 to the 1024th power bytes of memory. > > Please try it again.

Haha, alright, this time it worked. With another however though: Due to some other issues ( http://gist.github.com/576646 ) I was not able to install your whole package. (Will investigate and report properly tomorrow.) Instead i copied over MM_Win32.pm manually and then dmaked and tested DateTime-Timezone.

Sun Sep 12 21:22:36 2010 schwern [...] pobox.com - Correspondence added

Subject:	Re: [rt.cpan.org #61286] generated PM_TO_BLIB can be too long for ActivePerl
Date:	Sun, 12 Sep 2010 18:22:20 -0700
To:	bug-ExtUtils-MakeMaker [...] rt.cpan.org
From:	Michael G Schwern <schwern [...] pobox.com>

On 2010.9.12 5:27 PM, Christian Walde via RT wrote: Show quoted text

> Queue: ExtUtils-MakeMaker > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=61286 > > > On Sun Sep 12 19:36:11 2010, schwern@pobox.com wrote:

>> I've reported the problem upstream to dmake suggesting they eliminate

> the

>> MAXLINELENGTH concept entirely. It doesn't look too bad to do in the

> code. > > That is very nice of you. Thank you. :) >

>> I was asking for a mere 64 to the 1024th power bytes of memory. >> >> Please try it again.

> > Haha, alright, this time it worked. With another however though: > > Due to some other issues ( http://gist.github.com/576646 ) I was not > able to install your whole package. (Will investigate and report > properly tomorrow.) Instead i copied over MM_Win32.pm manually and then > dmaked and tested DateTime-Timezone.

The miniperl.t failure may be due to the tricks that t/lib/MakeMaker/Test/NoXS.pm is using to disable XS loading for that test. The xs.t failures look like there's something genuinely wrong with DLL linking. Windows DLL linking is out of my domain of understanding. C:\MinGW\bin\g++.exe -out:blib\arch\auto\XS\Test\Test.dll -mdll -L"C:\Perl\lib\CORE" Test.o C:\Perl\lib\CORE\perl510.lib -lkernel32 -luser32 -lgdi32 -lwinspool -lcomdlg32 -ladvapi32 -lshell32 -lole32 -loleaut32 -lnetapi32 -luuid -lws2_32 -lmpr -lwinmm -lversion -lodbc32 -lodbccp32 -lcomctl32 -lmsvcrt -def:Test.def C:\MinGW\bin\..\lib\gcc\mingw32\3.4.5\..\..\..\..\mingw32\bin\ld.exe: cannot open output file ut:blib\arch\auto\XS\Test\Test.dll: Invalid argument collect2: ld returned 1 exit status dmake.exe: Error code 129, while making 'blib\arch\auto\XS\Test\Test.dll' -- Reality is that which, when you stop believing in it, doesn't go away. -- Phillip K. Dick

Sun Sep 12 22:11:39 2010 schwern [...] pobox.com - Correspondence added

Subject:	Re: [rt.cpan.org #61286] generated PM_TO_BLIB can be too long for ActivePerl
Date:	Sun, 12 Sep 2010 19:11:28 -0700
To:	bug-ExtUtils-MakeMaker [...] rt.cpan.org
From:	Michael G Schwern <schwern [...] pobox.com>

Oh yeah, here's the link to the issue I filed with dmake. http://www.openoffice.org/issues/show_bug.cgi?id=114470

Thu Sep 23 14:50:26 2010 MITHALDU [...] cpan.org - Correspondence added

Show quoted text

> The xs.t failures look like there's something genuinely wrong with DLL

linking. Windows DLL linking is out of my domain of understanding. Show quoted text

> > C:\MinGW\bin\g++.exe -out:blib\arch\auto\XS\Test\Test.dll -mdll

Actually, this is a known issue, for which you have a fix already. :D It is this one: https://rt.cpan.org/Public/Bug/Display.html?id=55752 ExtUtils::MM_Win32 checks for gcc presence by seeing if $Config{cc} ENDS with "gcc", which it never will on windows, as executables there end with ".exe". Thus it generates parameters for Visual Studio as default. I'm really hoping you can fix that in the next dev release. :)

Fri Oct 01 15:46:36 2010 mschwern [...] cpan.org - Correspondence added

The PM_TO_BLIB problem on dmake is resolved and will be in the next release. The gcc issue was taken care of in 55752.

Fri Oct 01 15:46:42 2010 mschwern [...] cpan.org - Status changed from 'open' to 'resolved'