Skip Menu |

This queue is for tickets about the Devel-Size CPAN distribution.

Report information
The Basics
Id: 33530
Status: resolved
Worked: 10 min
Priority: 0/
Queue: Devel-Size

People
Owner: TELS [...] cpan.org
Requestors: ANDK [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Normal
Broken in: 0.69
Fixed in: (no value)



Subject: Not compiling under bleadperl
In bleadperl regular expressions became first class objects and some structs have been redesigned. This currently leads to: /home/src/perl/repoperls/installed-perls/perl/p3KWVui/perl-5.10.0@33346/bin/perl "-Iinc" /home/src/perl/repoperls/installed-perls/perl/p3KWVui/perl-5.10.0@33346/lib/5.11.0/ExtUtils/xsubpp -typemap /home/src/perl/repoperls/installed-perls/perl/p3KWVui/perl-5.10.0@33346/lib/5.11.0/ExtUtils/typemap Size.xs > Size.xsc && mv Size.xsc Size.c cc -c -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -DVERSION=\"0.69\" -DXS_VERSION=\"0.69\" -fPIC "-I/home/src/perl/repoperls/installed-perls/perl/p3KWVui/perl-5.10.0@33346/lib/5.11.0/i686-linux-64int/CORE" Size.c Size.xs: In function 'regex_size': Size.xs:202: error: 'REGEXP' has no member named 'nparens' Size.xs:203: error: 'REGEXP' has no member named 'precomp' Size.xs: In function 'thing_size': Size.xs:387: error: duplicate case value Size.xs:369: error: previously used here
Attached are the simple fixes for the compilation errors. But the tests crash at arrayrefs and recurse. t/basic........ Failed 8/13 subtests t/pod..........ok t/pod_cov......ok t/recurse...... Failed 51/54 subtests -- Reini Urban
difforig Devel-Size-0.69 2008-03-16 13:41:59 <rurban@x-ray.at> diff -ub Devel-Size-0.69/Size.xs.orig --- Devel-Size-0.69/Size.xs.orig 2007-08-12 11:13:44.000000000 +0200 +++ Devel-Size-0.69/Size.xs 2008-03-16 13:39:29.984375000 +0100 @@ -198,12 +198,17 @@ UV total_size = 0; total_size += sizeof(REGEXP); - /* Note hte size of the paren offset thing */ +#if (PERL_VERSION < 11) + /* Note the size of the paren offset thing */ total_size += sizeof(I32) * baseregex->nparens * 2; total_size += strlen(baseregex->precomp); - +#else + total_size += sizeof(struct regexp); + total_size += sizeof(I32) * SvANY(baseregex)->nparens * 2; + /*total_size += strlen(SvANY(baseregex)->subbeg);*/ +#endif if (go_yell && !regex_whine) { - carp("Devel::Size: Calculated sizes for compiled regexes are incomple, and probably always will be"); + carp("Devel::Size: Calculated sizes for compiled regexes are incompatible, and probably always will be"); regex_whine = 1; } @@ -383,12 +388,14 @@ total_size += sizeof(NV); #endif break; +#if (PERL_VERSION < 11) /* Is it a reference? */ case SVt_RV: #ifndef NEW_HEAD_LAYOUT total_size += sizeof(XRV); #endif break; +#endif /* How about a plain string? In which case we need to add in how much has been allocated */ case SVt_PV:
From: RURBAN [...] cpan.org
0.71 2008-04-14 rurban 68 tests * Fixed failing RV under 5.10 and 5.11. AV and HV were pushed directly onto the pending_array, and not the RV, which caused #33530. So I rewrote the logic to deref the RV inside the array traversal. * Fixed 5.11 RV/IV logic. * Removed one duplicate total_size arrayref test. * Added dbg_printf's. -- Reini Urban

Message body is not shown because it is too large.

Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Mon, 14 Apr 2008 17:20:25 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: Tels <nospam-abuse [...] bloodgate.com>
On Monday 14 April 2008 15:44:32 Reini Urban via RT wrote: Show quoted text
> Queue: Devel-Size > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=33530 > > > 0.71 2008-04-14 rurban 68 tests > * Fixed failing RV under 5.10 and 5.11. AV and HV were pushed > directly onto the pending_array, and not the RV, which caused #33530. > So I rewrote the logic to deref the RV inside the array traversal. * > Fixed 5.11 RV/IV logic. > * Removed one duplicate total_size arrayref test. > * Added dbg_printf's.
Thanx for the patch, but I will not apply the change from "thing" to "sv", renaming this variable doesn't add enough value (from my POV) to justify touching a lot of code. ATM I am happy Devel::Size works. Release of v0.70 will be done shortly, thanx again. Best wishes, Tels -- Signed on Mon Apr 14 17:19:32 2008 with key 0x93B84C15. Get one of my photo posters: http://bloodgate.com/posters PGP key on http://bloodgate.com/tels.asc or per email. This email violates U.S. patent #4,197,590: for (x = 0; x < widht; x++) { for (y = 0; y < height; y++) { setPixel (x+xm, y+ym, getPixel(x+xm,y+ym) ^ getCursorPixel(x,y); } }
Download (untitled)
application/pgp-signature 481b

Message body not shown because it is not plain text.

Is there an estamation possible when the next release is going to happen? Thanks,
Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Sat, 23 Aug 2008 09:21:43 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: Tels <nospam-abuse [...] bloodgate.com>
On Saturday 23 August 2008 06:21:00 Andreas Koenig via RT wrote: Show quoted text
> Queue: Devel-Size > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=33530 > > > Is there an estamation possible when the next release is going to > happen?
Five minutes ago :) Just pushed the well-aged v0.70 from my HD to PAUSE. Briefly tested to compile and pass the tests on v5.8.8 and v5.10.0 - but I am not sure if it really fixes all issues. I believe there is still an unresolved segfault somewhere hidden. Tests and patches welcome :) Tels PS: Sorry for the delay... :( -- Signed on Sat Aug 23 09:20:33 2008 with key 0x93B84C15. Get one of my photo posters: http://bloodgate.com/posters PGP key on http://bloodgate.com/tels.asc or per email. "We're looking at a future where only the very largest companies will be able to implement software, and it will technically be illegal for other people to do so." -- Bruce Perens, 2004-01-23
Download signature.asc
application/pgp-signature 481b

Message body not shown because it is not plain text.

Thanks for the quick response. In the meantime bleadperl has again moved and broken Devel-Size. It was this patch: Change 33656 by nicholas@mouse-mill on 2008/04/07 11:29:51 Eliminate cop_label from struct cop by storing a label as the first entry in the hints hash. Most statements don't have labels, so this will save memory. Not sure how much. It leads to this gcc error message: Size.xs: In function ‘op_size’: Size.xs:333: error: ‘COP’ has no member named ‘cop_label’ Size.xs:334: error: ‘COP’ has no member named ‘cop_label’ Besides that the tests are failing for a wide range of bleadperls. I have not found a single bleadperl after 5.10 that passes the tests. Thanks again,
Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Sun, 24 Aug 2008 00:19:58 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: Tels <nospam-abuse [...] bloodgate.com>
On Saturday 23 August 2008 19:54:35 Andreas Koenig via RT wrote: Show quoted text
> Queue: Devel-Size > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=33530 > > > Thanks for the quick response. > > In the meantime bleadperl has again moved and broken Devel-Size. It > was this patch: > > Change 33656 by nicholas@mouse-mill on 2008/04/07 11:29:51 > > Eliminate cop_label from struct cop by storing a label as the first > entry in the hints hash. Most statements don't have labels, so this > will save memory. Not sure how much. > > > > It leads to this gcc error message: > > Size.xs: In function ‘op_size’: > Size.xs:333: error: ‘COP’ has no member named ‘cop_label’ > Size.xs:334: error: ‘COP’ has no member named ‘cop_label’ > > > Besides that the tests are failing for a wide range of bleadperls. I > have not found a single bleadperl after 5.10 that passes the tests.
Unfortunately I neither have the time nor expertise to look into that. But I'll gladly apply any patches and release new versions. All the best, Tels -- Signed on Sun Aug 24 00:19:04 2008 with key 0x93B84C15. Get one of my photo posters: http://bloodgate.com/posters PGP key on http://bloodgate.com/tels.asc or per email. "Un bon mot ne prouve rien." -- Voltaire
Download signature.asc
application/pgp-signature 481b

Message body not shown because it is not plain text.

I have uploaded a patch that fixes this to the CPAN as ANDK/Devel-Size-0.70-RURBAN-03.patch.gz It's just the patch from Reini, adjusted to current 0.70. Thanks,
Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Sun, 24 Aug 2008 11:28:45 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: Tels <nospam-abuse [...] bloodgate.com>
On Sunday 24 August 2008 06:03:44 Andreas Koenig via RT wrote: Show quoted text
> Queue: Devel-Size > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=33530 > > > I have uploaded a patch that fixes this to the CPAN as > > ANDK/Devel-Size-0.70-RURBAN-03.patch.gz
It would really have helped if you had given me a host to that file, too :) Also, compressing the file with gzip twice was unnecessary :) Show quoted text
> It's just the patch from Reini, adjusted to current 0.70.
Oh, yeah, I remember now, that infamous patch that only should change about 4 lines, but changes the code all over by renaming a variable. Reini never responded to my request for a more localized change. :( So I rewrote the patch, hopefully fixing the error and not changing unrelated things and not dropping anything. There is one thing that I did not take in: - RETVAL = sv_size(sv, tracking_hash); ... + if (SvOK(sv)) + RETVAL = sv_size(sv, tracking_hash); That change causes a test in basic.t to fail under v5.8.8 and v5.10.0, because RETVAL was not set, resulting in garbage being returned. If that change is nec. for 5.11, someone please drop me a note (preferable with a small patch :) Here is a copy to a pre-release package: http://www.bloodgate.com/perl/packages/devel/Devel-Size-0.71.tar.gz Would be good if someone could test that with blead as I don't have it here. All the best, Tels -- Signed on Sun Aug 24 10:15:09 2008 with key 0x93B84C15. View my photo gallery: http://bloodgate.com/photos PGP key on http://bloodgate.com/tels.asc or per email. Like my code? Want to hire me to write some code for you? Send email!
Download signature.asc
application/pgp-signature 481b

Message body not shown because it is not plain text.

CC: ANDK [...] cpan.org
Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Sun, 24 Aug 2008 14:08:20 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: andreas.koenig.7os6VVqR [...] franz.ak.mind.de (Andreas J. Koenig)
Show quoted text
>>>>> On Sun, 24 Aug 2008 05:29:56 -0400, "nospam-abuse@bloodgate.com via RT" <bug-Devel-Size@rt.cpan.org> said:
Show quoted text
>> I have uploaded a patch that fixes this to the CPAN as >> >> ANDK/Devel-Size-0.70-RURBAN-03.patch.gz
Show quoted text
> It would really have helped if you had given me a host to that file, > too :)
I may need a bit handholding for this one. Is it possible that you did not read the line above? "Uploaded ... to the CPAN" is this not sufficient for you? What would you suggest instead? That I wait until the patch has propagated and then send you a URL? Nothing easier than that, but please confirm that it is what you expect from me. So here is one of URL 300 possible URLs: ftp://ftp.gwdg.de/pub/languages/perl/CPAN/authors/id/A/AN/ANDK/Devel-Size-0.70-RURBAN-03.patch.gz Show quoted text
> Also, compressing the file with gzip twice was unnecessary :)
Where did you find it doubly compressed? Please be more specific, I do not know what you are talking about. Show quoted text
>> It's just the patch from Reini, adjusted to current 0.70.
Show quoted text
> [...]
Show quoted text
> That change causes a test in basic.t to fail under v5.8.8 and v5.10.0, > because RETVAL was not set, resulting in garbage being returned.
I have tested the patch with 5.8.8, 5.10.0, and current bleadperl and I did not see any test failure. Sounds suspicious. Show quoted text
> If > that change is nec. for 5.11, someone please drop me a note (preferable > with a small patch :)
Show quoted text
> Here is a copy to a pre-release package:
Show quoted text
Tested with bleadperl@34202, 5.10, and 5.8.8, passes the tests. Thanks, -- andreas
Subject: Re: [rt.cpan.org #33530] Not compiling under bleadperl
Date: Sun, 24 Aug 2008 14:46:29 +0200
To: bug-Devel-Size [...] rt.cpan.org
From: Tels <nospam-abuse [...] bloodgate.com>
On Sunday 24 August 2008 14:06:57 (Andreas J. Koenig) via RT wrote: Show quoted text
> Queue: Devel-Size > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=33530 > >
> >>>>> On Sun, 24 Aug 2008 05:29:56 -0400, "nospam-abuse@bloodgate.com > >>>>> via RT" <bug-Devel-Size@rt.cpan.org> said:
> >> > >> I have uploaded a patch that fixes this to the CPAN as > >> > >> ANDK/Devel-Size-0.70-RURBAN-03.patch.gz > >>
> > It would really have helped if you had given me a host to that > > file, too :)
> > I may need a bit handholding for this one. Is it possible that you > did not read the line above? "Uploaded ... to the CPAN" is this not > sufficient for you? What would you suggest instead? That I wait until > the patch has propagated and then send you a URL? Nothing easier than > that, but please confirm that it is what you expect from me. > > So here is one of URL 300 possible URLs: > > > ftp://ftp.gwdg.de/pub/languages/perl/CPAN/authors/id/A/AN/ANDK/Devel- >Size-0.70-RURBAN-03.patch.gz
The problem is that I can't click the original line as it is not a valid URL. So I did got to several hosts like "search.cpan.org", "pause.org" etc etc., appended that line and it and was never right or valid. In the end I browsed to search.cpan.org, clicked on AUTHORS and browsed until I found your page, then clicked on your CPAN directory, then used CTRL-F in my browser to finally find the file. That is an awful lot of work just to download one file. You could also just have attached it on your email :) (I am not that familiar with CPAN as you are, so excuse my ignorance.) Show quoted text
> > Also, compressing the file with gzip twice was unnecessary :)
> > Where did you find it doubly compressed? Please be more specific, I > do not know what you are talking about.
I tried it with wget and the FTP URL above and it was only once compressed. However, the version I downloaded with Firefox I uncompressed with gzip and afterwards the file was still gzipped (which confused me, because mc transparently ungzips the file before displaying, while vi does not, ending up in garbage). Can no longer tell if that was a browser issue, or my mistake in uncompressing the file, so it was probably "tels had not enough coffee yet" :) Show quoted text
> >> It's just the patch from Reini, adjusted to current 0.70. > >>
> > [...] > > > > That change causes a test in basic.t to fail under v5.8.8 and > > v5.10.0, because RETVAL was not set, resulting in garbage being > > returned.
> > I have tested the patch with 5.8.8, 5.10.0, and current bleadperl and > I did not see any test failure. Sounds suspicious.
Inserting that line does cause indeed test failures here, so I am not sure how that came into the patch. Show quoted text
> > If > > that change is nec. for 5.11, someone please drop me a note > > (preferable with a small patch :) > > > > Here is a copy to a pre-release package: > > > > http://www.bloodgate.com/perl/packages/devel/Devel-Size-0.71.tar > >.gz
> > Tested with bleadperl@34202, 5.10, and 5.8.8, passes the tests.
Cool, I upload v0.71 to CPAN then. One more thing, I believe it will report incorrect numbers on on bleedperl less than the COP patch (33656), but it might be that the testsuite won't catch this. (It usually only tests sizes relativ to each other, as the exact size on a specific architecture/perl version isn't known in advance). Not sure if this issue is worth doing anything about, because there will be only little people who use bleedperl between 5.10.0 and @33656 :) All the best, Tels -- Signed on Sun Aug 24 14:39:17 2008 with key 0x93B84C15. View my photo gallery: http://bloodgate.com/photos PGP key on http://bloodgate.com/tels.asc or per email. Marketing lesson #1: The synergy of the result driven leverage can *never* incentivize a paradigm shift. -- Walterk (124748) on 2004-01-16 at /.
Download signature.asc
application/pgp-signature 481b

Message body not shown because it is not plain text.

Closing this issue as v0.70 was already released :)