#5214: Setup koji tags and repo for Docs
Reporter: crantila | Owner: rel-eng@…
Type: task | Status: new
Milestone: Fedora 18 Alpha | Component: koji
Keywords: docs | Blocked By:
Posted here as per [https://fedorahosted.org/fedora-
To support simpler and more robust updates to docs.fp.o, need to setup
koji to handle a new tag and repo for Docs.
<Sparks> What would be required to stand up a separate instance of koji
for Docs? <dgilmore> Sparks: why would you want a seperate instance?
<dgilmore> Sparks: i dont see any valid reason to do so <Sparks> dgilmore:
For the Docs website. Publican has the ability publish documentation from
packages (separate repo from the Fedora repo) for the website. We want to
replace the git repo that operates it now. <dgilmore> Sparks: and why does
that need a seperate koji? <Sparks> The git repo has gotten HUGE and is
becoming an issue. <dgilmore> Sparks: so, why does that mean a seperate
koji instance <dgilmore> Sparks: we could use a seperate tag and targets
in koji <dgilmore> i dont see why it would need its own koji <Sparks>
dgilmore: It's either that or try to get everyone setup as packagers.
<dgilmore> Sparks: get everyone setup as packagers <Sparks> dgilmore:
Except that they really won't be packagers. <dgilmore> Sparks: though you
dont need to be a a packager to get a koji cert <dgilmore> you just need
fas <dgilmore> Sparks: what would be the workflow? <Sparks> dgilmore:
Okay, and with that we can send packages through koji and tag them
separately? <Sparks> dgilmore: Basically we just tell Publican to build
the package and submit it to koji. The Publican software does all the
work. <dgilmore> Sparks: I still really dont know what your trying to do.
pretend im an idiot(not really that hard) and explain what it is and how
it should work <Sparks> dgilmore: I'm not far off... <Sparks> dgilmore: So
Publican will make an SRPM package, submit it to koji destined for a repo.
Our Publican backend will install those packages and publish the data to
the website. <Sparks> ...as I understand it <dgilmore> Sparks: so we would
need to set up seperate tags and tagets for it, defining a koji policy
allowing srpms to be built. likely we would add a new group to koji and
add people allowed to build docs to it and limit access to the
tags/targets to people in that group <dgilmore> Sparks: so its all doable
<Sparks> dgilmore: Well that's a lot easier. :) <Sparks> dgilmore: We
already have a group in FAS for people that should have access (docs-
publishers). Should I open a ticket? <dgilmore> Sparks: koji doesnt know
anything about fas <dgilmore> Sparks: please file a ticket. We wont be
able to make changes until after f17 is done <Sparks> dgilmore: That's
fine. We're not completely ready for the transition so after F17 would be
good. <Sparks> dgilmore: Thanks!
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5214>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
The Defensive and coding guide sports an interesting and forgotten feature
in its file organization: directories.
Beyond directories, it also happens to show files having the same name
(such as Language.xml, found in Python/, Java/, CXX/, C/, etc. or
You tell me, that is what directories are for. Fair enough.
The real issue here is that pushing .po resulting from those XML files to
Transifex which only handles uniquely named resources in a flat hierarchy
leads to a mess, where there is no predictibility in having the Language
resource always the same on Transifex, depending how those are pushed.
Could it be possible to rename all those files to unique names, ideally
with the subdir as a prefix (such as CXX-Language.xml or C-schema.xml) so
that we have unique resource names in Transifex?
Furthermore, could we add a requirement to documentation authors in order
to ensure that all XML files in their document hierarchy show unique names?
Currently looking at the translation of the FreeIPA docs, the glossary
will pose an interesting challenge: strings in the DocBook XML file
are ordered, as those are supposed to be in the alphabetic order.
However, when the translation phase happens, words will change
(obviously), and for latin alphabet-based languages (not speaking
about other alphabets or languages not using an alphabet which will
have the same kind of challenge, maybe not based on word initials),
the initials of those words won't be kept. And thus the order of
appearance for the words.
Would you have a recommendation about the technique to use to handle
such a sorted list of words, where the order would be kept, using
--- Comment #17 from Eric Christensen <sparks(a)redhat.com> ---
I only know of one other project that had its base (source) language that
wasn't EN. The "problem" was solved by translating it to English and then to
Unfortunately that is how our model is setup. Our expectation is that the
translators will translate text from English to another language. This should
limit someone's ability to write in their native language, though. There's
just another hop involved.
You are receiving this mail because:
You are the assignee for the bug.
The Fedora Documentation group will be holding the regular Monday
meeting at 1400UTC in #fedora-meeting on freenode.
-- Pete Travis
- Fedora Docs Project Leader
- 'randomuser' on freenode
As the F20 Alpha is tomorrow, we need final text ready today to hand it over
to people who do announcement (this time it's nirik). I'm not sure I'll make
today's marketing meeting, so quick recap:
I've done a quick review of current version - changed final release to early
December, removed Vagrant change as it's not yet ready. The other Changes
should be ok - marked as MODIFIED in Bugzilla. IRC chat posted down on the
page is commented out now.
One question - how do we want to reference Changes? When it's clear we talk
about Changes (reference to ChangeSet), we should use it but we also talk a lot
about features in the text...
Otherwise I think it looks good, thanks a lot. Formatting for email is still
needed, Robyn usually does it.
Fedora_20_Alpha_release_notes already points to F20_Alpha_release_announcement.
en-US/Server_Configuration_Tools.xml | 50 +++++++++++++-----------------------
1 file changed, 18 insertions(+), 32 deletions(-)
diff --git a/en-US/Server_Configuration_Tools.xml b/en-US/Server_Configuration_Tools.xml
index fb49b3f..978b5ca 100755
@@ -6,44 +6,30 @@
<title>Server Configuration Tools</title>
- <section id="section1">
- <title>App Installet</title>
- Fedora 20 includes a new application installer called gnome-software. gpk-application is still available.
- <section id="section2">
- <title>Acpica tools</title>
- The new acpica-tools package has been designed to replace the existing iasl and pmtools package. The tool brings the Fedora distribution up-to-date for ACPI development and exploration, improves the quality of the existing packaging by additionally running test suites.
- <section id="section3">
- <title>F20 Boost 154</title>
+ <section id="sect-F_&PRODVER;_Boost_154">
+ <title>F&PRODVER; Boost 154</title>
- Boost has been upgraded to version 1.54.0. Apart from a number of bugfixes, this brings in three new libraries: Boost.Log for logging, Boost.TTI for Type Traits Introspection, and Boost.TypeErasure for runtime polymorphism based on concepts.
+ Boost has been upgraded to version 1.54.0. Apart from a number of bugfixes, this brings in three new libraries:
+ <member>Boost.Log for logging</member>
+ <member>Boost.TTI for Type Traits Introspection</member>
+ <member>Boost.TypeErasure for runtime polymorphism based on concepts</member>
- <section id="section4">
- <title>No Default Syslog</title>
- F20 does not install a Syslog service by default anymore. For browsing log messages please type "journalctl" rather than "less /var/log/messages". Please type "journalctl -f" instead of "tail -f /var/log/messages". Please use "journalctl | grep foobar" instead of "grep foobar /var/log/messages". If the administrator needs /var/log/messages or support for the BSD syslog network protocol we recommend installing a syslog daemon such as rsyslog or syslog-ng with a command like like the following:
-$ yum install rsyslog
- <section id="section5">
+ <section id="sect-Snapshot_and_Tool">
<title>Snapshot and Rollback Tool</title>
- Fedora now provides tools for saving and restoring the system state of systems using a yum plugin and the roller-derby CLI tool.
+ Fedora &PRODVER; now provides tools for saving and restoring the system state of systems using a yum plugin and the Roller Derby CLI tool.In order to prevent loss of data when rebooting into an snapshot, the Roller Derby CLI will allow saving a snapshot of the current state before rolling back and will provide tools to allow mounting of that current state to recover changes that have occurred since the rollback point.
I've just published the draft Technical Notes for Fedora 20. The updated
content should be available within an hour or two, once the content
is picked up by the proxies. Technical notes are a great place for beat
writers to start, please take a look.
-- Pete Travis
- Fedora Docs Project Leader
- 'randomuser' on freenode