You can't save spelling words.

Clear message

    PaulHowarth/Blog/2010-06

Paul's Blog Entries for June 2010

Tuesday 1st June 2010

Fedora Project

  • Updated perl-Mail-Mbox-MessageParser to fix warning messages when used with perl 5.12.0:

  • "Mail::Mbox::MessageParser::OLDSTDERR" used only once: possible typo at /usr/share/perl5/Mail/Mbox/MessageParser.pm line 297.
  • This warning was causing the test suite for grepmail to fail, so I raised CPAN RT#58053 about it

  • Built grepmail for perl 5.12.0 in Rawhide

Local Packages

  • Updated dovecot (1.2) to build with Solr full text search support where possible (Bug #595933), adding build requirements libcurl-devel and expat-devel

  • Updated dovecot (2.0) and pigeonhole to 20100529 snapshots, enabling Solr full text search where possible, adding buildreqs libcurl-devel and expat-devel for Solr support, and building with -fno-strict-aliasing for the moment

  • Updated perl-DBM-Deep to 1.0024, and updated the patch for old Test::More versions:

    • Stale references now compare equal to undef, "" and 0

    • Warn and assign undef if you assign a stale reference to a database location

    • Improve diagnostics when using a stale reference as a hash or array ref
    • Fix compatibility with perl 5.12.0

    • Singleton support has been re-enabled in the File back end

    • Fix t/43_transaction_maximum.t to detect maximum number of open files

  • Updated perl-Mail-Mbox-MessageParser as per the Fedora package

  • Updated perl-YAML-Syck to 1.09:

  • Updated sendmail to use the proper scriptlets for user and group creation

  • Rebuilt perl-DBD-SQLite, perl-DBI, perl-Socket6, and perl-String-Format for RHEL-6 Beta

Wednesday 2nd June 2010

Local Packages

  • Updated perl-Moose to 1.06 (added '0+' overloading in Moose::Meta::TypeConstraint so that we can more uniformly compare type constraints between 'classic' Moose type constraints and MooseX::Types based type constraints)

  • Rebuilt perl-PAR-Dist, perl-Params-Util, perl-Params-Validate, perl-Parse-CPAN-Meta, perl-Perl-Tidy, perl-Pod-Spell, perl-Probe-Perl, perl-Readonly-XS, and perl-Readonly for RHEL-6 Beta

Thursday 3rd June 2010

Fedora Project

  • Updated perl-Crypt-DSA to fix its META.yml that claimed it would work with perl 5.005 onwards, which isn't true due to the use of a 3-argument open in Crypt::DSA::KeyChain, as pointed out by current versions of Perl::MinimumVersion (Bug #600012) - raised upstream as CPAN RT#58094

Local Packages

  • Rebuilt perl-Array-Compare, perl-Net-DNS-Resolver-Programmable, perl-Net-LibIDN, perl-Net-SMTP-SSL, perl-Net-SSLeay, perl-Net-Server, perl-Number-Compare, perl-Package-Constants, and perl-Package-Generator for RHEL-6 Beta

IPv6 Connectivity

I'm having a go at setting up IPV6 connectivity for my home network, but I'm not going to go into any detail on it at the moment as I don't really know what I'm doing and I wouldn't want to publish bad information.

What I can say though is that I was quickly and easily able to get an IPv6 over IPv4 tunnel established via Hurricane Electric's excellent free Tunnel Broker service with no waiting around for ages, and I'm basically following Wolfgang S. Rupprecht's Fedora IPv6 Static Tunnel Guide, except using shorewall6 to handle routing and firewalling rather than tweaks to /etc/sysconfig/network and ip6tables.

Friday 4th June 2010

Local Packages

  • Updated curl to add a workaround for broken applications using curl multi (Bug #599340) and enable threaded DNS lookup instead of using c-ares for F-12 and F-13 builds in addition to the existing Rawhide build

  • Rebuilt perl-Hook-LexWrap, perl-IO-Tty, perl-IPC-Run, perl-IPC-Run3, perl-Jcode, perl-LMAP-CID2SPF, perl-Mail-SPF-Test, perl-Mail-SPF, perl-Mail-Sender, perl-MailTools, perl-Module-Info, and perl-Module-Pluggable for RHEL-6 Beta

Saturday 5th June 2010

Local Packages

  • Rebuilt perl-Compress-Raw-Bzip2, perl-Compress-Raw-Lzma, perl-Compress-Raw-Zlib, perl-IO-Compress-Lzma, perl-IO-Compress, perl-IO-Multiplex, perl-IO-Socket-SSL, and perl-IO-Zlib for RHEL-6 Beta

Sunday 6th June 2010

Fedora Project

  • Updated gtkwave in Rawhide to 3.3.7 (minor bugfixes)

Local Packages

  • Updated gtkwave to 3.3.7 (minor bugfixes)

  • Rebuilt perl-File-Remove and perl-File-Which for RHEL-6 Beta

Monday 7th June 2010

Local Packages

  • New package perl-Package-Stash (0.03)

  • Updated moin to 1.9.3 (lots of bug fixes)

  • Updated perl-Class-MOP to 1.03 (make CMOP::Package a thin wrapper around Package::Stash), adding buildreq perl(Package::Stash)

  • Updated perl-DBM-Deep to 1.0025 (fixed t/39_singletons.t to work on Windows)

  • Updated perl-IO-Socket-INET6 to 2.63 (upstream patch for select timeout issue in t/io_multihomed6.t) and added a new patch for t/io_multihomed6.t on systems with broken getaddrinfo() (CPAN RT#58198)

  • Updated perl-Moose to 1.07 (fixed a minor metaclass compatibility fixing bug dealing with immutable classes and non-class metaclass traits)

  • Updated perl-YAML-Syck to 1.10 (bump Module::Install to 0.99), dropping my patch for Module::Install breakage, no longer needed

  • Rebuilt perl-ExtUtils-CBuilder, perl-ExtUtils-ParseXS, and perl-File-Find-Rule-Perl for RHEL-6 Beta

Tuesday 8th June 2010

Fedora Project

  • Updated mod_fcgid to drop the SELinux policy module for RHEL-6 onwards

  • New package perl-IO-Compress-Lzma (2.027) built for Rawhide in dist-f14-perltest branch; Fedora 13 build waiting on `perl(IO::Compress::Base)' 2.027 being built and released

  • Updated spamass-milter to make the postfix subpackage noarch on RHEL-6 onwards

Local Packages

  • Updated bluefish to 2.0.1-rc1, dropping upstreamed log10 linking patch and redundant buildreq pcre-devel, plus enabling a noarch subpackage for the shared data on RHEL-6 onwards

  • Updated libssh2 to make the docs subpackage noarch on RHEL-6 onwards

  • Updated mod_fcgid as per the Fedora package

  • Updated perl-DateTime to make the perl-DateTime-Locale and perl-DateTime-TimeZone packages noarch on RHEL-6 onwards

  • Updated perl-IO-Socket-INET6 to 2.64 (patch for CPAN RT#58198 upstreamed)

  • Updated spamass-milter as per the Fedora package

  • Rebuilt perl-Archive-Tar, perl-Archive-Zip, perl-B-Keywords, perl-Carp-Clan, perl-Class-Data-Inheritable, perl-Class-Inspector, perl-Class-Singleton, perl-Clone, perl-Config-Tiny, perl-Convert-TNEF, perl-Convert-UUlib, perl-Crypt-SmbHash, perl-Data-OptList, perl-Digest-MD4, perl-Digest-SHA, perl-Error, perl-Exception-Class, and perl-Expect for RHEL-6 Beta

Wednesday 9th June 2010

Local Packages

  • Updated perl-Software-License to 0.101600, enabling RELEASE_TESTING and adding build requirement perl(Test::Pod), adding a patch to build with ExtUtils::MakeMaker < 6.31 and another patch to build with Test::More < 0.88

  • Updated sendmail to move sasl2 config file from %{_libdir}/sasl2 to %{_sysconfdir}/sasl2 from RHEL-5 onwards, to add a patch for building with libdb version 5, and to build with libdb from Fedora 14 onwards

Thursday 10th June 2010

Fedora Project

Local Packages

  • Updated c-ares to 1.7.2:

    • Added ares_parse_mx_reply()

    • ares_init: Last, not first instance of domain or search should win

    • Improve alternative definition of bool

    • Fix VS2010 compiler warnings
  • Updated dovecot to 20100610 build:

    • lib-storage: fixed accessing uncommitted saved mails with dsync

    • example-config: moved ACL and quota settings to a separate .conf files

    • dbox, mdbox: fixed race conditions when creating mailboxes

  • Updated libssh2 to 1.2.6, dropping the upstreamed AES-CTR detection patch:

    • Added libssh2_sftp_statvfs() and libssh2_sftp_fstatvfs()

    • Added libssh2_knownhost_checkp()

    • Added libssh2_scp_send64()

    • wait_socket made c89 compliant and use two fd_sets for select()

    • OpenSSL AES-128-CTR detection fixed
    • Proper keyboard-interactive user dialog in the sftp.c example

    • Fixed build procedure for VMS
    • Fixed libssh2.dsw to use the generated libssh2.dsp

    • Several Windows-related build fixes
    • Fixed fail to init SFTP if session isn't already authenticated
    • Many tiny fixes that address clang-analyzer warnings
    • sftp_open: deal with short channel_write calls

    • libssh2_publickey_init: fixed to work better non-blocking

    • sftp_close_handle: add precaution to not access NULL pointer

    • sftp_readdir: simplified and bug-fixed

    • channel_write: if data has been sent, don't return EAGAIN

Friday 11th June 2010

Fedora Project

  • Wrote a patch to silence the warning sendmail emits if it can't find %{_libdir}/sasl/Sendmail.conf, which has just been moved to %{_sysconfdir}/sasl/Sendmail.conf in Rawhide; sendmail looks there next and of course finds what it's looking for

  • ChrisWeyl reviewed and approved my perl-Package-Stash submission

Local Packages

  • Updated c-ares to 1.7.3:

    • Builds on Android
    • Now includes all files necessary to build it (1.7.2 lacked a file)
  • Updated perl-Data-Section to 0.101620 (attempt to work with files with Win32 line endings)

  • Updated perl-IO-Socket-INET6 to 2.65 (further fixes for t/io_multihomed6.t)

  • Updated sendmail as per the Fedora version

Monday 14th June 2010

Fedora Project

  • Raised Bug #603717 about GTK tear-off menus not working in Fedora 13 (originally discovered as a bluefish issue)

  • Updated perl-Package-Stash in Rawhide (dist-f14-perltest tag) to 0.04 (get_package_symbol now doesn't autovivify stash entries; a new method get_or_add_package_symbol can now be used for that behaviour)

Local Packages

  • Updated moin (1.8) to 1.8.8:

    • Fixed XSS issues (see MoinMoinBugs/1.9.2UnescapedInputForThemeAddMsg)

    • Fixed XSS in Despam action (CVE-2010-0828)

    • wikiutil.clean_input: avoid crash if it gets str type

    • Add RenderAsDocbook to actions_excluded if we have no python-xml

    • AttachFile._build_filelist: verifies readonly flag for unzip file link

    • attachUrl: fix wrongly generated tickets (e.g. for AttachList macro)

    • Disable usage of dircache, deprecate dc* functions

    • Remove all calls to filesys.dc* (dclistdir, dcdisable)

    • Fixed crash, see MoinMoinPatch/IncludeMacroWithDocBookFormatter

    • Avoid hardly recoverable crashes if #format specification is invalid

    • Fixed info action: AttachFile sub-actions that needed a ticket did not work

    • auth.ldap_login: add report_invalid_credentials parameter

  • Updated perl-Package-Stash as per the Fedora package

  • Updated perl-Test-LeakTrace to 0.11 (fix false-positive related to XS code - CPAN RT #58133)

Tuesday 15th June 2010

Fedora Project

Local Packages

  • Updated c-ares to fix the multilib file conflict in c-ares-devel (Bug #602880)

  • Updated perl-Test-Synopsis to add build and runtime dependencies on perl(Test::Builder::Module) and drop the perl ≥ 5.8.1 requirement, which is always met if we have Test::Builder::Module

  • Updated pptp to include patches from upstream CVS aimed at fixing segfaults due to re-use of freed data

  • Rebuilt sendmail to incorporate the trivial changes in the Rawhide 8.14.4-8 build

Wednesday 16th June 2010

Fedora Project

Local Packages

  • New package perl-Test-EOL (0.8), including patch to silence warnings on old perl versions (CPAN RT#58442)

  • Updated perl-Moose to 1.08 (refactored a small amount of Moose::Meta::Method::Constructor to allow it to be overridden more easily)

  • Updated perl-Package-Stash to 0.05 (packaging changes), adding buildreqs perl(Test::EOL) and perl(Test::NoTabs)

  • Updated perl-Test-EOL to 0.9 (patch for CPAN RT#58442 upstreamed)

  • Updated ppp to include the EAP-TLS patch

Thursday 17th June 2010

Kudos to Kustom PCs

Recently I have been having trouble with my i7 desktop/server box overheating. When I first built it, the sensors output looked like this:

$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0:      +41.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0001
Adapter: ISA adapter
Core 1:      +36.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0002
Adapter: ISA adapter
Core 2:      +37.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0003
Adapter: ISA adapter
Core 3:      +36.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0004
Adapter: ISA adapter
Core 4:      +40.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0005
Adapter: ISA adapter
Core 5:      +37.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0006
Adapter: ISA adapter
Core 6:      +37.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0007
Adapter: ISA adapter
Core 7:      +36.0°C  (high = +80.0°C, crit = +100.0°C)  

it8720-isa-0290
Adapter: ISA adapter
in0:         +1.04 V  (min =  +0.00 V, max =  +4.08 V)   
in1:         +1.52 V  (min =  +0.00 V, max =  +4.08 V)   
in2:         +0.14 V  (min =  +0.00 V, max =  +4.08 V)   
in3:         +3.02 V  (min =  +0.00 V, max =  +4.08 V)   
in4:         +0.14 V  (min =  +0.00 V, max =  +4.08 V)   
in5:         +0.10 V  (min =  +0.00 V, max =  +4.08 V)   
in6:         +0.10 V  (min =  +0.00 V, max =  +4.08 V)   
in7:         +2.16 V  (min =  +0.00 V, max =  +4.08 V)   
Vbat:        +3.07 V
fan1:       2033 RPM  (min =   10 RPM)
fan2:          0 RPM  (min =    0 RPM)
fan3:          0 RPM  (min =    0 RPM)
fan4:          0 RPM  (min =    0 RPM)
temp1:       +39.0°C  (low  = +127.0°C, high = +127.0°C)
temp2:       +26.0°C  (low  = +127.0°C, high = +60.0°C)
temp3:       +43.0°C  (low  = +127.0°C, high = +127.0°C)
cpu0_vid:   +2.050 V

In the last week or so, temperatures went up significantly, so much so in fact that the on-board temperature alarm had been going off when the machine is busy:

$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0001
Adapter: ISA adapter
Core 1:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0002
Adapter: ISA adapter
Core 2:      +60.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0003
Adapter: ISA adapter
Core 3:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0004
Adapter: ISA adapter
Core 4:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0005
Adapter: ISA adapter
Core 5:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0006
Adapter: ISA adapter
Core 6:      +60.0°C  (high = +80.0°C, crit = +100.0°C)  

coretemp-isa-0007
Adapter: ISA adapter
Core 7:      +61.0°C  (high = +80.0°C, crit = +100.0°C)  

it8720-isa-0290
Adapter: ISA adapter
in0:         +1.04 V  (min =  +0.00 V, max =  +4.08 V)   
in1:         +1.65 V  (min =  +0.00 V, max =  +4.08 V)   
in2:         +0.18 V  (min =  +0.00 V, max =  +4.08 V)   
in3:         +3.01 V  (min =  +0.00 V, max =  +4.08 V)   
in4:         +0.30 V  (min =  +0.00 V, max =  +4.08 V)   
in5:         +0.11 V  (min =  +0.00 V, max =  +4.08 V)   
in6:         +0.16 V  (min =  +0.00 V, max =  +4.08 V)   
in7:         +2.16 V  (min =  +0.00 V, max =  +4.08 V)   
Vbat:        +3.28 V
fan1:        915 RPM  (min =   10 RPM)
fan2:          0 RPM  (min =    0 RPM)
fan3:          0 RPM  (min =    0 RPM)
fan4:          0 RPM  (min =    0 RPM)
temp1:       +45.0°C  (low  = +127.0°C, high = +127.0°C) 
temp2:       +48.0°C  (low  = +127.0°C, high = +60.0°C)
temp3:       +59.0°C  (low  = +127.0°C, high = +127.0°C)
cpu0_vid:   +2.050 V

The reason would appear to be that the CPU fan was going at less than half the speed it had been originally, and never sped up even when the CPU was very hot. So the first thing I did was to clean out the inside of the machine, get rid of all of the dust, unplug and reconnect the CPU fan etc. It made not a jot of difference. I figured one of two things were wrong: either the CPU fan was faulty or the motherboard (a Gigabyte EX58-UD5) was making it run too slowly. Given that the former has moving parts and the latter doesn't, I thought the CPU fan was the most likely culprit. The CPU was bought retail from Kustom PCs in Ayr - they had the best price at the time I bought it and the CPU cooler came with the CPU. I had purchased the motherboard elsewhere. I emailed their Technical Support and quickly got a response back from Graeme, who was going to check what speed the stock cooler ran on a machine there. He found that it should indeed be running faster and mailed me a replacement cooler the next working day. Today I changed the cooler (a fiddly job since I had to extract the motherboard from the case in order to be able to safely apply enough pressure on the cooler retaining clips to get them to go in properly) and the new fan ran at the expected speed:

$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0:      +44.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1:      +46.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0002
Adapter: ISA adapter
Core 2:      +43.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0003
Adapter: ISA adapter
Core 3:      +39.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0004
Adapter: ISA adapter
Core 4:      +44.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0005
Adapter: ISA adapter
Core 5:      +46.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0006
Adapter: ISA adapter
Core 6:      +44.0°C  (high = +80.0°C, crit = +100.0°C)

coretemp-isa-0007
Adapter: ISA adapter
Core 7:      +39.0°C  (high = +80.0°C, crit = +100.0°C)

it8720-isa-0290
Adapter: ISA adapter
in0:         +1.04 V  (min =  +0.00 V, max =  +4.08 V)
in1:         +1.65 V  (min =  +0.00 V, max =  +4.08 V)
in2:         +0.16 V  (min =  +0.00 V, max =  +4.08 V)
in3:         +3.01 V  (min =  +0.00 V, max =  +4.08 V)
in4:         +0.16 V  (min =  +0.00 V, max =  +4.08 V)
in5:         +0.11 V  (min =  +0.00 V, max =  +4.08 V)
in6:         +0.10 V  (min =  +0.00 V, max =  +4.08 V)
in7:         +2.16 V  (min =  +0.00 V, max =  +4.08 V)
Vbat:        +3.26 V
fan1:       2027 RPM  (min =   10 RPM)
fan2:          0 RPM  (min =    0 RPM)
fan3:          0 RPM  (min =    0 RPM)
fan4:          0 RPM  (min =    0 RPM)
temp1:       +40.0°C  (low  = +127.0°C, high = +127.0°C)  sensor = thermistor
temp2:       +30.0°C  (low  = +127.0°C, high = +64.0°C)  sensor = thermal diode
temp3:       +48.0°C  (low  = +127.0°C, high = +127.0°C)  sensor = thermistor

So, kudos to Kustom PCs for their excellent customer service. I will have no hesitation in buying from them again. They didn't want the faulty cooler back either by the way.

Friday 18th June 2010

Fedora Project

  • Updated bluefish to 2.0.1 in Rawhide, F-13, and EPEL-6

Local Packages

  • Updated bluefish to 2.0.1

  • Updated curl to 7.21.0 with lots of new features:

    • added the --proto and --proto-redir options

    • new configure option --enable-threaded-resolver

    • improve TELNET ability with libcurl

    • added support for PolarSSL
    • added support for FTP wildcard matching and downloads
    • added support for RTMP
    • introducing new LDAP code for new enough OpenLDAP
    • OpenLDAP support enabled for cygwin builds
    • added CURLINFO_PRIMARY_PORT, CURLINFO_LOCAL_IP and CURLINFO_LOCAL_PORT

  • and bugfixes:
    • prevent needless reverse name lookups
    • detect GSS on ancient Linux distros
    • GnuTLS: EOF caused error when it wasn't
    • GnuTLS: SSL handshake phase is non-blocking
    • -J/--remote-header-name strips CRLF

    • MSVC makefiles now use ws2_32.lib instead of wsock32.lib

    • -O crash on windows

    • SSL handshake timeout underflow in libcurl-NSS

    • multi interface missed storing connection time

    • broken CRL support in libcurl-NSS

    • ignore response-body on redirect even if compressed
    • OpenSSL handshake state-machine for multi interface

    • TFTP timeout option sent correctly
    • TFTP block id wrap
    • curl_multi_socket_action() timeout handles inaccuracy in timers better

    • SCP/SFTP failure to respect the timeout
    • spurious SSL connection aborts with OpenSSL
  • Updated python-zope-filesystem to add zope.app, five, zc, and z3c namespaces as per the Fedora package

  • Updated python-zope-interface to 3.6.1

Saturday 19th June 2010

Local Packages

  • Updated dovecot to 1.2.12:

    • deliver: fix crash when message with Auto-submitted: header gets rejected

    • lib-storage: fixed header searches with multiple headers with same name

    • dict client: disconnect from dict server after 1 second of idling

    • dict: if process crashed, it wasn't automatically restarted

    • dict file: don't change dict file's GID unnecessarily
    • maildir: fixed a memory leak when copying with hardlinks

    • maildir: fix crash when expunging last messages with changed filenames

  • Updated sieve to 0.1.17:

    • ensure source code positions for compiler messages recorded at token starts
    • fixed a few potential memory leaks
    • made command line tools return proper exit status upon failure
  • Updated perl-Class-C3 to 0.23 (fix various documentation problems)

Monday 21st June 2010

Local Packages

  • Updated nmap to make nmap-frontend a noarch subpackage where possible, enable parallel build and drop patches that are no longer needed

Tuesday 22nd June 2010

Fedora Project

  • Updated perl-Math-Pari to fix build failures on ppc64 (Bug #551988) and s390x architectures; the failing test was 55_intnum.t, which ran out of stack space on one of the subtests so I doubled the stack space with the patch below and that resolved the problem

  • --- Math-Pari-2.01080604/pari-2.3.4/src/test/in/intnum  2008-01-16 15:45:16.000000000 +0000
    +++ Math-Pari-2.01080604/pari-2.3.4/src/test/in/intnum  2010-06-22 13:53:30.711784407 +0100
    @@ -1,5 +1,5 @@
     default(echo,1);
    -allocatemem(20 * 10^6);
    +allocatemem(40 * 10^6);
     gettime; oo = [1];
     \p96
     intcirc(s=1, 0.5, zeta(s)) - 1
    --- Math-Pari-2.01080604/pari-2.3.4/src/test/32/intnum  2008-01-16 15:45:01.000000000 +0000
    +++ Math-Pari-2.01080604/pari-2.3.4/src/test/32/intnum  2010-06-22 14:11:00.678792014 +0100
    @@ -1,4 +1,4 @@
    -? allocatemem(20*10^6);
    +? allocatemem(40*10^6);
     ? gettime;oo=[1];
     ? \p96
        realprecision = 96 significant digits
    --- Math-Pari-2.01080604/pari-2.3.4/src/test/64/intnum  2008-01-16 15:45:28.000000000 +0000
    +++ Math-Pari-2.01080604/pari-2.3.4/src/test/64/intnum  2010-06-22 14:11:10.119846014 +0100
    @@ -1,4 +1,4 @@
    -? allocatemem(20*10^6);
    +? allocatemem(40*10^6);
     ? gettime;oo=[1];
     ? \p96
        realprecision = 96 significant digits

Local Packages

  • Updated python-zope-interface to include the test modules, and to run the test suite during the package build

Wednesday 23rd June 2010

Fedora Project

  • Built new packages perl-Test-Synopsis and perl-Test-NoTabs, which passed review today

Local Packages

  • New package perl-File-Slurp (9999.13), needed for perl-PPIx-Utilities

  • New package perl-PPIx-Utilities (1.000001), needed for perl-Perl-Critic 1.08

  • Updated perl-Module-Build to exclude inc::* from the package for perl 5.12.x builds due to a file conflict with the perl package in Rawhide (raised as Bug #607505)

  • Updated perl-Module-CoreList to 2.35 (updated for 5.13.2)

  • Updated perl-Perl-Critic to 1.108, adding buildreqs perl(Task::Weaken) and perl(PPIx::Utilities::Statement); I skipped t/20_policies.t for distributions with Test::Builder < 0.82 as that test requires note(), todo_start() and todo_end()

  • Started the mass rebuild of perl module packages for perl 5.12.1 in Rawhide with perl-Archive-Tar, perl-Archive-Zip, perl-B-Keywords, perl-Class-C3, perl-Class-Data-Inheritable, perl-Clone, perl-Compress-Raw-Bzip2, perl-Compress-Raw-Lzma, perl-Compress-Raw-Zlib, perl-Config-Tiny, perl-Data-OptList, perl-Data-Section, perl-Devel-Leak, perl-Exception-Class, perl-ExtUtils-ParseXS, perl-File-Which, perl-IO-Compress-Lzma, perl-IO-Compress, perl-IO-String, perl-List-MoreUtils, perl-MRO-Compat, perl-Mixin-Linewise, perl-Module-Signature, perl-PAR-Dist, perl-PPI, perl-Params-Util, perl-Perl-MinimumVersion, perl-Perl-Tidy, perl-Pod-Eventual, perl-Pod-Readme, perl-Pod-Spell, perl-Probe-Perl, perl-Software-License, perl-String-Format, perl-Sub-Name, perl-Sub-Uplevel, perl-Taint-Runtime, perl-Task-Weaken, perl-Test-CPAN-Meta, perl-Test-Deep, perl-Test-Distribution, perl-Test-Exception, perl-Test-MinimumVersion, perl-Test-NoWarnings, perl-Test-Perl-Critic, perl-Test-Pod, perl-Test-Portability-Files, perl-Test-SubCalls, perl-Text-Template, perl-YAML-Syck, perl-YAML-Tiny, and perl-YAML

Thursday 24th June 2010

Local Packages

  • Updated perl-BerkeleyDB to build against libdb (Berkeley DB version 5) in Rawhide rather than db4

  • Updated perl-Sendmail-AccessDB to support build with libdb (db5) in Fedora 14 onwards, and to add a patch to remove deprecation warnings with perl 5.12 onwards:

  • The locked attribute for subroutines results in deprecation warnings in 5.12
    and has had no effect since 5.10.
    
    Use of :locked is deprecated at /builddir/build/BUILD/Sendmail-AccessDB-1.01/blib/lib/Sendmail/AccessDB.pm line 292.
    Use of :locked is deprecated at /builddir/build/BUILD/Sendmail-AccessDB-1.01/blib/lib/Sendmail/AccessDB.pm line 311.
    Use of :locked is deprecated at /builddir/build/BUILD/Sendmail-AccessDB-1.01/blib/lib/Sendmail/AccessDB.pm line 323.
    
    --- Sendmail-AccessDB-1.01/lib/Sendmail/AccessDB.pm     2010-06-24 17:20:53.596783652 +0100
    +++ Sendmail-AccessDB-1.01/lib/Sendmail/AccessDB.pm     2010-06-24 17:21:09.371899356 +0100
    @@ -289,7 +289,7 @@
         
     
     
    -sub _expand_ip : locked
    +sub _expand_ip
     {
         my $address = shift;
         my @expanded = ();
    @@ -308,7 +308,7 @@
         return @expanded;
     }
     
    -sub _expand_hostname : locked
    +sub _expand_hostname
     {
         my $hostname = shift;
         my @expanded = ($hostname);
    @@ -320,7 +320,7 @@
         return @expanded;
     } 
     
    -sub _expand_email : locked
    +sub _expand_email
     {
         my $address = shift;
         my @to_check = ($address);
  • The module needs tweaking properly to use a modern threading approach really (raised as CPAN RT#58832)

  • Continued the mass rebuild for perl 5.12.1 in Rawhide, rebuilding perl-Algorithm-C3, perl-Algorithm-Diff, perl-Crypt-SmbHash, perl-Digest-MD4, perl-Digest-SHA, perl-FileHandle-Unget, perl-FreezeThaw, perl-HTML-Tidy, perl-IO-Zlib, perl-IPC-Run3, perl-MLDBM, perl-Mail-Mbox-MessageParser, perl-Module-Info, perl-Package-Generator, perl-Pod-Coverage-TrustPod, perl-Pod-Coverage, perl-String-Escape, perl-Sub-Exporter, perl-Sub-Identify, perl-Sub-Install, perl-Sys-Hostname-Long, perl-Term-ReadKey, perl-Term-ReadLine-Gnu, perl-Test-ClassAPI, perl-Test-EOL, perl-Test-File, perl-Test-HTML-Tidy, perl-Test-LeakTrace, perl-Test-LongString, perl-Test-Manifest, perl-Test-NoTabs, perl-Test-Object, perl-Test-Output, perl-Test-Pod-Coverage, perl-Test-Prereq, perl-Test-Script, perl-Test-Signature, perl-Test-Spelling, perl-Test-Synopsis, perl-Test-Taint, perl-Test-Tester, perl-Test-Warn, perl-Text-Diff, perl-Text-Glob, perl-Tie-IxHash, perl-Tree-DAG_Node, perl-Try-Tiny, perl-UNIVERSAL-require, perl-Unicode-Map, perl-Unicode-String, perl-Unix-Syslog, perl-WeakRef, perl-XML-LibXML, perl-XML-LibXSLT, perl-XML-NamespaceSupport, perl-XML-SAX, perl-XML-Simple, perl-aliased, smbldap-tools, torrentsniff, and weblint

Friday 25th June 2010

Fedora Project

  • Finally got round to debugging a segfault in gtorrentviewer (Bug #542502 and Bug #572806), which turned out to be straightforward as one of the bug reporters took the trouble to provide a reliable reproducer of the bug; in the end it was the simple matter of a GValue being declared as a local variable in a function and not initialized to zero before use, resulting in this stack trace:

  • [New Thread 23163]
    Core was generated by `gtorrentviewer /storage/data/torrents-new/xxxx'.
    Program terminated with signal 11, Segmentation fault.
    #0  0x000000340cc28f9c in type_check_is_value_type_U (value=0x7fffdbacac80, 
        type=64) at gtype.c:3836
    3836    gtype.c: No such file or directory.
            in gtype.c
    
    ...
    
    Thread 1 (Thread 23160):
    #0  0x000000340cc28f9c in type_check_is_value_type_U (value=0x7fffdbacac80, 
        type=64) at gtype.c:3836
            tflags = G_TYPE_FLAG_VALUE_ABSTRACT
            node = 0x2224010
    #1  IA__g_type_check_value_holds (value=0x7fffdbacac80, type=64)
        at gtype.c:3878
    No locals.
    #2  0x000000340cc31219 in IA__g_value_set_string (value=0x7fffdbacac80, 
        v_string=0x22c4560 "http://tracker.openbittorrent.com/announce")
        at gvaluetypes.c:922
            __val = 0x7fffdbacac80
            __r = <value optimized out>
            _g_boolean_var_ = <value optimized out>
            new_val = 0x22c4560 "http://tracker.openbittorrent.com/announce"
            __PRETTY_FUNCTION__ = "IA__g_value_set_string"
    #3  0x0000000000407aa6 in on_RefreshTrackerButton_clicked (
        mwin=<value optimized out>, user_data=<value optimized out>)
        at mainwindow.c:753
            iter = {stamp = -1347712751, user_data = 0x22e5830, user_data2 = 0x0, 
              user_data3 = 0x0}
            itemvalue = {g_type = 35799056, data = {{v_int = 214096389, v_uint = 
        214096389, v_long = 223552395781, v_ulong = 223552395781, v_int64 = 
        223552395781, v_uint64 = 223552395781, v_float = 3.00216485e-31, 
                  v_double = 1.1044955880089772e-312, v_pointer = 0x340cc2da05}, {
                  v_int = 36582800, v_uint = 36582800, v_long = 36582800, 
                  v_ulong = 36582800, v_int64 = 36582800, v_uint64 = 36582800, 
                  v_float = 1.27988727e-37, v_double = 1.8074304708681154e-316, 
                  v_pointer = 0x22e3590 }
            }}
            err = 0x340cc1894c
  • The fix was as simple as this:
  • --- GTorrentViewer-0.2b/src/mainwindow.c.orig   2004-10-26 05:37:58.000000000 +0100
    +++ GTorrentViewer-0.2b/src/mainwindow.c        2010-06-25 14:43:22.391786427 +0100
    @@ -744,7 +744,7 @@
     {
       GtkListStore *list;
       GtkTreeIter iter;
    -  GValue itemvalue;
    +  GValue itemvalue = {0};
       gchar *tracker;
       GError *err;

Local Packages

  • Updated gtorrentviewer as per the Fedora package, fixing a crash when the "Refresh" button on the "Tracker Details" tab was pushed

  • Rebuilt libspf2 for perl 5.12.1 in Rawhide and added -fno-strict-aliasing compiler option

  • Updated perl-Number-Compare with patch from CPAN RT#58466 to fix "uninitialized value in comparison" warnings (raised as Bug #607982 in Fedora)

  • Updated perl-YAML-Tiny to 1.42:

    • No functional changes
    • Clarified documation regarding treatment of "bool" keywords

    • YAML::Perl comparison tests were behaving incorrectly

    • Author tests updated and moved into xt/

    • Updated the versions of the other YAML parsers for testing
  • Completed the rebuild of Rawhide packages for perl 5.12.1: bw-whois, grepmail, perl-Array-Compare, perl-Carp-Clan, perl-Class-C3-XS, perl-Class-Factory-Util, perl-Class-Inspector, perl-Class-MOP, perl-Class-Singleton, perl-ConfigReader-Simple, perl-Convert-BinHex, perl-Convert-TNEF, perl-Convert-UUlib, perl-DBD-SQLite, perl-DBI, perl-DBM-Deep, perl-DateTime-Calendar-Mayan, perl-DateTime-Format-Builder, perl-DateTime-Format-MySQL, perl-DateTime-Format-Strptime, perl-DateTime, perl-Declare-Constraints-Simple, perl-Devel-GlobalDestruction, perl-Devel-StackTrace, perl-Devel-Symdump, perl-Digest-BubbleBabble, perl-Email-Address, perl-Error, perl-Expect, perl-ExtUtils-CBuilder, perl-File-Find-Rule-Perl, perl-File-Find-Rule, perl-File-Remove, perl-FileHandle-Fmode, perl-HTML-Lint, perl-HTML-SimpleLinkExtor, perl-HTTP-SimpleLinkChecker, perl-HTTP-Size, perl-Hook-LexWrap, perl-IO-Multiplex, perl-IO-Socket-INET6, perl-IO-Socket-SSL, perl-IO-Tty, perl-IO-stringy, perl-IPC-Run, perl-Jcode, perl-LMAP-CID2SPF, perl-Locale-US, perl-MIME-tools, perl-Mail-IMAPClient, perl-Mail-SPF-Query, perl-Mail-SPF-Test, perl-Mail-SPF, perl-Mail-SRS, perl-Mail-Sender, perl-Mail-Sendmail, perl-MailTools, perl-Module-Pluggable, perl-Module-Refresh, perl-Moose, perl-Net-CIDR-Lite, perl-Net-CIDR, perl-Net-DNS-Resolver-Programmable, perl-Net-DNS, perl-Net-IP, perl-Net-LibIDN, perl-Net-SMTP-SSL, perl-Net-SSLeay, perl-Net-Server, perl-NetAddr-IP, perl-Package-Constants, perl-Package-Stash, perl-Params-Coerce, perl-Params-Validate, perl-Parse-CPAN-Meta, perl-Pod-Escapes, perl-Pod-Simple, perl-Readonly-XS, perl-Readonly, perl-Regexp-Common, perl-SUPER, perl-Scope-Guard, perl-Sendmail-PMilter, perl-Socket6, perl-TimeDate, perl-version, and weblint++

Saturday 26th June 2010

Fedora Project

  • Updated gtkwave to 3.3.8 in Rawhide:

    • add failure check on tempfile creation in fstReaderInit()

    • addstrace_repeat_count and an appropriate menu option

    • add drag-and-drop of signals from gtkwave into rtlbrowse

    • remove the "/File/Quit/Don't Quit" menu item if "fast exit" is enabled

    • more warnings clean-ups

Local Packages

  • Updated gtkwave to 3.3.8 as per the Fedora package

  • Split perl-Test-HTML-Lint out as a sub-package of perl-HTML-Lint as its dependency chain pulls in things like perl-devel

Monday 28th June 2010

Local Packages

  • New package perl-Data-Dumper-Names (0.03)

  • New package perl-PadWalker (1.9)

  • New package perl-Test-Differences (0.4801)

  • New package perl-Test-Most (0.21 for Fedora 9 onwards, 0.11 for Fedora 5 to Fedora 8)

  • Updated curl to add support for NTLM authentication with NSS, as per the Rawhide package (Bug #603783)

  • Updated dovecot (and pigeonhole) to 20100626 build:

    • moved disable_plaintext_auth to 10-auth.conf

    • mdbox: fixed assert-crash on storage rebuild if file got lost

    • lib-charset: don't assert-crash when iconv() skips lots of invalid input

    • master: fixed crash on deinit (maybe also on reload)

  • Updated getmail to 4.18.0

    • update broken link in FAQ
    • strip a few extra problematic characters when generating oldmail filenames

    • improve diagnostics for external programs writing to stderr

  • Updated perl-DateTime-Format-Strptime to 1.3000 (changed license to Artistic 2.0, fixed bug where specifiers such as %e that allowed for leading space before a number would previously cause DateTime.pm to throw an error if the date being parsed actually contained leading space - CPAN RT#58459); I also added buildreqs Test::EOL, Test::NoTabs, Test::Pod, Test::Pod::Coverage, and Test::Spelling and enabled AUTHOR_TESTING and RELEASE_TESTING for higher test coverage and added patches to support building with ExtUtils::MakeMaker < 6.31 and to get the spell check test working reliably

  • Updated perl-Exception-Class to 1.31 (changed license to Artistic 2.0, made it possible to use Exception::Class::Base without loading Exception::Class - CPAN RT#57269); I also added buildreqs Test::EOL and Test::NoTabs for higher test coverage and added a patch to support building with ExtUtils::MakeMaker < 6.31

  • Updated perl-Net-CIDR to 0.14:

    • corrected some documentation errors
    • cidrvalidate() will validate either an IP address or a CIDR

    • re-implemented addr2cidr without Math::BigInt, which is much faster for IPv6 addresses

Tuesday 29th June 2010

Local Packages

  • Updated getmail to fix this traceback due to a missing import of the regex library:

  • getmail version 4.18.0
    Copyright (C) 1998-2009 Charles Cazabon.  Licensed under the GNU GPL version 2.
    SimplePOP3Retriever:earlesfield@mail.plus.net:110:
    
    Exception: please read docs/BUGS and include the following information in any bug report:
    
      getmail version 4.18.0
      Python version 2.6.2 (r262:71600, Jan 25 2010, 18:46:47)
    [GCC 4.4.2 20091222 (Red Hat 4.4.2-20)]
    
    Unhandled exception follows:
        File "/usr/bin/getmail", line 685, in main
        go(configs)
        File "/usr/bin/getmail", line 102, in go
        retriever.initialize()
        File "/usr/lib/python2.6/site-packages/getmailcore/_retrieverbases.py", line 559, in initialize
        RetrieverSkeleton.initialize(self)
        File "/usr/lib/python2.6/site-packages/getmailcore/_retrieverbases.py", line 414, in initialize
        oldmail_filename = re.sub(
      NameError: global name 're' is not defined
    
    Please also include configuration information from running getmail
    with your normal options plus "--dump".
  • I did of course report the issue and send the patch upstream
  • Updated perl-Archive-Tar to 1.62 (partly address problem with exit status of Archive::Tar->extract_archive() - Bug #606813, CPAN RT#58636)

  • Updated perl-DateTime to 0.58, adding buildreqs Test::NoTabs, Test::EOL, Test::Spelling and aspell-en and enabling RELEASE_TESTING for higher test coverage, and adding my own perl-provides script as the default one prior to Fedora 7 was unable to extract them properly, causing broken dependencies for packages using DateTime:

    • Make DateTime::LeapSecond have the same $VERSION as every other .pm file

    • The set_formatter() method did not return the DateTime object, and did not actually validate the value provided for the formatter (CPAN RT#58506)

    • Improved docs on floating time zone (CPAN RT#56389)

    • Mention of end-of-month algorithms in docs on DateTime math (CPAN RT#58533)

    • License is now Artistic 2.0
  • Updated perl-DateTime-Format-Strptime to 1.4000 (actually update $VERSION in module file)

  • Updated perl-Exception-Class to 1.32 (removed Test::Most dependency, which was causing a circular dependency chain - CPAN RT#58889)

  • Updated perl-Test-Differences to 0.500, adding buildreqs perl(Test::Pod) and perl(Test::Pod::Coverage) for extra test cover:

    • Add support for all diff styles supplied by Text::Diff (CPAN RT#23579)

    • Add Build.PL

    • Convert to universally use Test::More instead of Test

    • Convert to modern Perl distribution.
    • Applied doc suggestion from CPAN RT#24297

    • Fix the { a => 1 } versus { a => '1' } bug (CPAN RT#3029)

Wednesday 30th June 2010

Local Packages

  • Updated getmail to 4.20.0 (fix crap code from bad svn merge that slipped into 4.18.0, triggering exceptions in MDA_external configs)

  • Updated perl-DateTime to 0.59:

    • Make sure that POD only shows up in the right files; in 0.56 through 0.58, some POD in the wrong place confused the search.cpan.org POD display code, and the main module's documentation wasn't viewable)

Previous Month: May 2010
Next Month: July 2010

Recent