this is to let you know that boost has been upgraded to 1.50, sans
Python 3 support, which will arrive later.
After complaints in past years, we decided to request a dedicated
side-tag ("f18-boost") not to break a hundred or so packages that depend
on it in Rawhide. As always, all client packages, especially those that
depend on boost runtime libraries, need to be rebuilt, but this time
around you need to request the right tag explicitly:
fedpkg build --target=f18-boost
This tag will live for a couple weeks, but I think we should merge back
before alpha freeze, which is 2012-08-14, in about two weeks.
There's at least one disruptive change that I know about:
Boost.Filesystem v2 has been discontinued and anyone who hasn't yet
ported their code to v3 must do so now.
As usual, feel free to open bugs, ping me on IRC (_petr) or send e-mails
if one of your packages doesn't like the new boost. We'll figure out
how to work around the notorious API changes. (Or fix boost.)
Any chance Fedora moves from Koji to OBS?
Why should they?
OBS has a web interface so one can easily fix packages even from an
internet cafe/work/windows pc.
Projects can be developed separately and packages can be easily
branched and submitted.
openSUSE is entirely build by OBS.
Third party are unstable packages maintainers, can develop them at one
place for different distributions.
e.g as we do with unity for Fedora
What would stop Fedora from doing this switch?
I'm leaving for a vacation tonight and will be on the road through most of
Monday. Similarly next week when I return. In between I'll should have
some network access, but my schedule will be weird. If something important
cmoes up, I'll probably be able to deal with it.
I just wanted to confirm that for the first time ever, I could connect
to the IETF-v6 only network out of the box. Great job! I have complained
about this a lot in the last two years, so I figure I will also praise
the fix :)
ps. I'd mention the bugzilla item, but bugzilla.redhat.com is lacking
an IPv6 address :)
Always the same process (ocamlopt.opt) and always on 32 bit only.
The thing is, it *didn't* happen just 3 days ago. Nothing has changed
in the package, and ocamlopt.opt is the same as 3 days ago.
glibc has had a few memory-related fixes in the past 3 days:
- Revert patch for BZ696143, it made it impossible to use IPV6
addresses explicitly in getaddrinfo, which in turn broke
ssh, apache and other code. (#808147)
- Avoid another unbound alloca in vfprintf (#841318)
- Remove /etc/localtime.tzupdate in lua scriptlets
- Revert back to using posix.symlink as posix.link with a 3rd
argument isn't supported in the lua version embedded in rpm.
- Revert recent changes to res_send (804630, 835090).
- Fix memcpy args in res_send (#841787).
Has something changed in Koji or mock such as inherited signal masks?
I even went as far as building a 32 bit Rawhide VM to test this, but I
can't reproduce it there, and that's pretty odd considering it happens
reliably in Koji.
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages. http://libguestfs.org
I'm going to orphan tritonus package, it's an old java audio api that
does not build anymore on Rawhide and getting it working would require
too much time on OpenJDK 1.7.
Ping me if you are interested in any way.
Lukas "lzap" Zapletal
I'm seeking request for approval to upgrade the asciidoc package . The
latest version of AsciiDoc is 8.6.7, released this month. The version
currently packaged is 8.4.5, released May 2009 (not exactly what we would
An issue request has already been filed for this upgrade  and a link to
an updated spec file is available in the comments.
The request for approval has to do with backward compatibility of output.
(In case you haven't heard of it, AsciiDoc is a script that transforms
lightweight markup--ala markdown--into html, docbook, epub, etc). Any
document written for AsciiDoc 8.4.5 is expected to compile successfully
with the latest version, as AsciiDoc maintains backwards compatibility of
the syntax. However, the output markup is constantly being tweaked and
refined. Therefore, the output will be different. Better, but different.
There are a several compelling reasons why the package should be upgraded.
First and foremost, adoption of AsciiDoc is taking off, partly because
O'Reilly authors can write book drafts entirely in AsciiDoc. Several high
profile projects such as Neo4j and JBoss have also adopted AsciiDoc. As a
result, a growing number of developers are now looking for it for the first
time. When they see that the version in Fedora is so old, they either swap
to a different distro (Ubuntu Precise, Debian Sid and even Debian Stable
have newer versions), or they download it from the project website. The
fact that the other distros have newer versions works in the favor of this
request, since build systems are likely requiring the newer versions. It
also means those doc builds break when they run on Fedora.
Other key enhancements include:
* HTML5, ePub and slidy output
* source highlighting using pygments
* new themes and a theme engine
* lots of smaller additions to the markup language
I've attempted to summarize the changes to AsciiDoc since the version in
Fedora 17, including gotchas for people upgrading, at the bottom of this
Can we proceed with the upgrade for Fedora 17? Fedora 18?
The Asciidoc numbered and toc attributes cause DocBook outputs to
include asciidoc-numbered and asciidoc-toc processing instructions,
these are used by DocBook XSL to include section numbering and
table of contents (like Asciidoc HTML backends). For backward
compatibility both numbered and toc attributes are defined by
default when the docbook backend is used.
most obvious change is that the toc.js script has been superceded by
asciidoc-xhtml11.js so if you’re linking you’ll need get a copy of the
xhtml11: Verse block and verse paragraph content enveloped in a pre
tag (instead of a div) so it renders better in text-only browsers.
CSS quirks mode is no longer the default xhtml11 output
The global configuration files directory is ignored by both
asciidoc and a2x if AsciiDoc configuration files are installed in
the same directory as the asciidoc executable. This change allows
both a system wide copy and multiple local copies of AsciiDoc to
coexist on the same host PC.
Break compatibility with quoted element positional color
and size attributes (HTML backends). To revert to the deprecated quote
behavior define the deprecated-quotes attribute in the global
asciidoc.conf file or on the command-line. For a more detailed
explanation of the rationale behind this change see
xhtml11: list bullets colored gray.
Renamed paragraph default style to normal.
attributes, titles and specialcharacters sections are now read from
the local asciidoc.conf file before the header is parsed. This
fixes a regression problem. See
Added text foreground and background color along with text size CSS
styles for XHTML outputs, see
xhtml11, html5: Made verse and quote block text darker to print
legibly in Google Chrome browser.
* Added html5 backend.
* The new a2x can output EPUB formatted documents (script now written
in python instead of bash).
* Pygments (http://pygments.org/) option added to source code
highlight filter. Based on Pygments source code filter written by
* A new flask theme for xhtml11 and html5 backends. A shameless
knock-off of the Flask website styling.
* xhtml11: Added a new theme (volnitsky). Written and contributed by
Leonid V. Volnitsky.
* xhtml11,html5: Added toc2 attribute which generates a scrollable
table of contents in the left hand margin. Based on customized CSS
written by Suraj Kurapati, see
* Implemented counter system attribute.
* Added style and role CSS classes to xhtml11 section templates.
* Added dblatex table-width processing instruction — tables generated
by dblatex now observe the AsciiDoc table width as a percentage
(thanks to Gustav Broberg for suggesting this enhancement).
* Added a start attribute to numbered lists to set the start number.
* Added slidy backend (based on Phillip Lord’s slidy backend
* a2x: Added --xsl-file option to allow custom XSL stylesheets to be
* html5,xhtml11: If the data-uri attribute is defined then icons from
the theme icons directory (if they exist) will be embedded in the
* Added args block attribute to source highlight blocks to allow
arbitrary parameters to be passed to the source highlighters.
* Implemented toc:: block macro and toc-placement attribute for
HTML backends to allow the Table of Contents placement to be set
manually by the author.
* fixes to Jython compatibility
Principal Software Engineer, Red Hat | Author of Seam in Action
Registered Linux User #231597
It looks like the latest f15 dracut to ever make it to -stable was
009-12. I ran across this when I found a laptop that hadn't been on in
a while and needed an upgrade to f17. The trick, of course, is that
convertfs went into dracut-009-13. After -13, -14 and -15 were built.
The Wiki links to the former bodhi page for -15 :
but apparently it never got enough karma to move to stable. I'm
presuming that when f15 went EOL all the bodhi pages went away too, so
the link went dead.
That leaves us with no yum upgrade path from 15 to 17 (due to
infrastructure details, not functionality). I think that means
preupgrade won't work either (though I use RAID-1 for /boot, so I don't
know too much about preupgrade).
Suggestions for the best way forward? I was thinking some constructive
1) out-of-cycle push to stable
2) place in -testing (easy enough to document, doesn't help trivial
3) host elsewhere on Fedora infrastructure (ditto)
4) host -15 off Fedora infrastructure (not awesome in my book)
We could always tell folks that they have to step through f16 first or
do a DVD upgrade, but those are somewhat less elegant and much more
Bill McGonigle, Owner
BFC Computing, LLC
Email, IM, VOIP: bill(a)bfccomputing.com
Social networks: bill_mcgonigle/bill.mcgonigle