February 20th, 2011 — GNOME
The obvious lesson learned this time is that you either have a handyman or you need to do it yourself. Bear with me and see the results below. The Pan 0.134 release showcases work done by several volunteering people and awaited by a vivid community around the Pan USENET Newsreader.
By the way, the Pan Newsreader is one of the older FLOSS projects out there. As far as I can tell, it went public in July 1999 and the C++ rewrite beta series, which this release is part of, started back in 2006, again under the lead of Charles Kerr. 0.134 wouldn’t be possible without K. Haley who continued with the development after Charles ceased his active involvement in the project.
And last, but not least, C++ developers, translators, artists, testers and other contributors are most welcome!
0.134 “Wait for Me”
What is Pan?
Pan is a newsreader which attempts to be pleasing to both new and experienced users. In addition to the standard newsreader features, Pan also supports yEnc, offline newsreading, article filtering, multiple connections, and more.
It’s also the only Unix newsreader to get a perfect score on the Good Net-Keeping Seal of Approval evaluations.
This release brings a large number of feature improvements and bugfixes that have accumulated over the last 30 months of release hiatus. The code is based on the Pan bugfix & testing repository maintained by K. Haley and it is recommended for everyone using older versions of Pan to upgrade.
Changes since 0.133
- Fix compilation with GCC 4.4. (Charles Kerr, #573722)
- WARNING: setting an adjustment with non-zero page size is deprecated. (Charles Kerr, #579753)
- No window icon when running in NZB-mode. (Charles Kerr, #574419)
- Remove deprecated glib/gdk/gtk calls to prepare for GNOME 3. (Charles Kerr, #596648)
- Use GRegex instead of PCRE. (Charles Kerr, #596653)
- Tooltips missing from two ‘Post Article’ toolbar buttons. (Charles Kerr, #548860)
- Infinite loop with server that doesn’t support LIST NEWSGROUPS. (Charles Kerr, #545220)
- ‘Add port to server dialog’ — apply commits 862da67, af30418 from lostcoder. (K. Haley, Charles Kerr, #527313)
- Support 64 bit article numbers. (Charles Kerr, #549655)
- Don’t queue xover for 0 connections. (Charles Kerr, #596682)
- Make wrapping honor changes in compose-wrap pref. (Charles Kerr, #596680)
- Change nzb task saving delay & add pref. (Charles Kerr, #596683)
- Port to GMime 2.4. (K. Haley, #541676)
- Spawn editor asynchronously. (K. Haley, #465763)
- Change allocation buffer for article tree. (K. Haley)
- Save some more memory by re-ordering a few variables. (K. Haley)
- Don’t save files as executable. (K. Haley)
- Fix handling of multibyte spaces in text-massager. (K. Haley)
- Re-write multipart handling for viewing. (K. Haley)
- Change default mime-type for incoming attachments. (K. Haley, #135734)
- Add some additional mime types. (K. Haley)
- Add memchunk.h to Makefile.am. (K. Haley)
- GTK+ 2.16 required. (K. Haley)
- Fix bug in multipart article mids. (K. Haley)
- Add console support for windows. (K. Haley)
- Update filter-info and scorefile-test to support non-overview headers. (K. Haley)
- Skip non-overview test if not cached. (K. Haley)
- Make ArticleCache::get_message const. (K. Haley)
- Allow scoring article on all headers. (K. Haley)
- Promote rescore_articles to data interface. (K. Haley)
- Update score when article is cached. (K. Haley)
- Add regex support to search. (Jack Cuyler, K. Haley, #351196)
- Additional info for about & User-Agent. (K. Haley, #424083)
- Add support for compiling with gmime-2.5. (K. Haley)
- Allow subject line use in save path. (K. Haley, #403797)
- Improve regexes used for squashing subject line. (K. Haley)
- Reduce memory allocation for multiparts. (K. Haley)
- Allow newsrc paths relative to PAN_HOME. (K. Haley)
- Save tasks on exit. (K. Haley, #609355)
- Always show full revision info in UA hdr. (K. Haley)
- Remember size of post window. (K. Haley)
- Add tests for subject line. (K. Haley)
- Make separator user configurable. (K. Haley)
- Replace deprecated gdk_pixmap_create_from_data. (K. Haley)
- Add support for Face header. (K. Haley)
- Fix crash possibly due to change in cairo. (K. Haley)
- Change from GdkPixmap to GdkPixbuf. (K. Haley)
- Remove deprecated function through gtk 2.18. (K. Haley)
- Compatibility with -std=c++0x. (K. Haley)
- Fix crash in gio_func on OSX. (K. Haley)
- Updated translations: Arabic (Djihed Afifi), Brazilian Portuguese (Flamarion Jorge), Spanish (Jorge Gonzalez), Portuguese (Duarte Loreto), French (Claude Paroz, Bruno Brouard), Catalan (Joan Duran), German (Mario Blättermann, Andreas Kuhlen), Basque (Inaki Larranaga Murgoitio), Swedish (Daniel Nylander), Czech (Petr Kovar), Slovenian (Andrej Žnidaršič), Danish (Joe Hansen), Simplified Chinese (Aron Xu)
January 31st, 2011 — GNOME, Localization
Just a word of preface to fellow quarterly report writers: we still lack ca. 8 reports, so if you haven’t submit your update yet, now is the right time to do so. The deadline was postponed to February 4, 2011.
On October 16, Gil Forcada presented results of the GNOME I18N Survey which was referred to in the previous report. A brief analysis of the results was included.
Discussion on the possibility and feasibility of translating schema files within separated gettext domains or catalogs emerged from the survey analysis debate, as well as the point of localizing certain types of strings that are usually not user-visible. Especially the price of splitting limited resources within smaller translation teams was compared with the eventual need to make significant changes to the current GNOME i18n infrastructure and also to various module build systems.
With regard to the Release Team’s second proposal for moduleset reorganization from October 7, which would allow various software projects outside of the GNOME infrastructure to become officially endorsed GNOME software, members of the GNOME Translation Project expressed strong preference for working on l10n support within the GNOME official i18n and SCM infrastructure.
In the debate which spread over the gnome-18n and desktop-devel-list groups, GNOME translators were mainly concerned about translation quality, string freeze periods and release schedules, about expecting developers or maintainers to integrate translations manually to their respective repositories in a suitable, timely manner, and generally about changing the current module requirements by dropping them and/or making them optional for official GNOME software and GNOME developers.
Several proposals were made to (require to) allow the DL infrastructure on l10n.gnome.org auto-commit translations to code repositories not hosted on git.gnome.org, to migrate from the DL application altogether and replace it with Transifex, and generally to specify l10n requirements for official modules more narrowly and precisely. No final resolution was made in this regard.
Sysadmin work on DL auto-commit, providing translators a way to manage l10n support without interacting with Git system directly, was resumed during October and November. Furthermore, GTP members discussed options to integrate automatic QA checking with l10n.gnome.org.
There were also changes in coordination of the Persian and Romanian team in October and November, respectively.
October 28th, 2010 — GNOME, Localization
(As drafted on the gnome-i18n mailing list several days ago.)
On July 29, Andre Klapper represented the GNOME Translation Project at the AGM meeting at GUADEC with a Project update report. At GUADEC, he also gave a talk on “Identifying software projects and translation teams in need” where he provided an overview of interesting data combined & gathered from Damned Lies, GNOME Bugzilla and other relevant sources.
Gil Forcada, with the feedback from other community members, conducted the GNOME I18N Survey by sending a questionnaire on August 13 to every GTP language coordinator, and collecting answers for two weeks. Out of 120 coordinators, 36 answered. The rationale behind the survey was to know each other within the GNOME translation community better, and thus to find ways the GTP can improve the overall experience of translating GNOME.
The sent questionnaire consisted of more than 20 questions on various areas of community l10n in GNOME, e.g. inquiring about general team information, coordinatorship & membership, team workflow, QA processes, use & evaluation of GNOME Damned Lies infrastructure, collaborating with downstream translators, other translation teams, and language institutions, community knowledge sharing, etc.
As for the GNOME development itself, GTP language teams have been busy working on providing l10n support for the new GNOME stable release 2.32, which was delivered on September 29. GTP has been also investigating approaches to help out language teams that seemed to be considerably short on manpower and/or proper coordinatorship, this included the Persian and Welsh teams.
We also communicated with GNOME developers to try to solve i18n issues with translating strings within submodules, strings with constructed sentences, and some other problems that (re)appeared during the Q3 period.
During Q4, apart from working on l10n support for the upcoming GNOME 3, GTP community aims for identifying issues with the current i18n & l10n infrastructure inside and outside the GNOME Project, like the Git commit functionality, and solving them, hopefully implementing the necessary GTP support for repositories hosted at git.gnome.org and elsewhere. This is to be done in conjunction with the Release Team’s proposal for moduleset reorganization.
October 6th, 2010 — Localization
You are a developer and you want to keep your project moving forward. You set up various communication channels and organize an open community around. Your vivid project starts to attract new people, amongst them are people who intend to contribute code to the project. Great, because that’s what you were waiting for when you started building your FLOSS community.
Naturally, you do not allow anyone on the net to directly contribute code without any, more or less formal, review process. That’s good, since you care about happy community of contributing members, but you also want quality that you can be proud of.
And then there’s the world of community localization. You are most likely not a polyglot, and you can hardly do a review process with tons of languages from around the world, apart from making sure that the localization work you’ve been provided is not missing some obvious bits of technicalities. So you simply open the submit process for l10n to anyone, or reach to some nice outsourcing tool, hoping that translators will eventually cope with it and the project’s l10n will be worthy, as is the code. But really?
But Really, It’s For Your Own Good
That is, not to stick to the openness at any cost. The fact is that, quite similarly to the code contribution, the quality work in l10n will not miraculously show up. It needs reviews, proper management, suitable workflow. It needs community.
One of the first things you need to do if you want to facilitate building a real l10n community is to set up, more or less formal, rules. So you turn the translation teams option on. You encourage work in translation teams & projects, so your translators can interact with each other and share knowledge. You keep an eye on l10n. You are responsive to the needs of your translation team members. Then you are a great developer with FLOSS project that deserves quality & efficiency in l10n very much comparable to that of the professional (as in commercial) translators with plenty of ISO & DIN certificates. And your project has it. For free. (Almost.)
August 26th, 2010 — GNOME, Localization
If you happen to be one of the GNOME Translation Project language team coordinators, and you haven’t done so yet, now is the right time to participate in the GNOME i18n survey conducted by Gil Forcada on behalf of the GNOME Translation Project!
The rationale behind this survey is that the GNOME i18n community (which the language team coordinators are naturally part of) wants to know better each other, so that the GNOME Translation Project can improve the overall experience of translating GNOME, as Gil outlined in his survey email. Some of you may remember that there was a similar survey conducted by an Ubuntu i18n community in the past which greatly inspired this effort.
On August 13, Gil sent out a plain text file with survey questions to all coordinators’ addresses we could gather. Nevertheless, only a fraction of coordinators have responded so far, so once again, in case you are one of the majority, please don’t hesitate to take a few moments to fill out the questionnaire! Or if you know any of those coordinators, please ping them! Yes, it’s quite important!
For those of you interested in knowing what the survey questions are, you can find them attached in the aforementioned Gil’s email, and a final draft is available on live.gnome.org. Also keep an eye there for results.
August 24th, 2010 — Uncategorized
In case your hardware still consists of very legacy NVIDIA chips like GeForce2 or GeForce4 and you are experiencing some problems running X server on Fedora 13 with the now and for some time default nouveau driver (like screen freezing after a period of time with an apparent need to hard restart), you can still switch to the old nv driver and refrain from using either an extremely basic VESA driver, or a legacy proprietary driver produced by NVIDIA. (Which, looking at rpmfusion.org, doesn’t seem to be packaged for Fedora 13 to provide a support for GeForce2 etc. I’ve read something about compatibility issues with version 1.8 of the X server, but I haven’t looked into that further.)
One of the drawbacks is that you will have to live without hardware acceleration, since only software acceleration is provided. (Okay, these chips weren’t GNOME Shell-ready anyway.)
A quick memo on how to accomplish that switch on a fresh Fedora 13 install:
First, you need to create your own
xorg.conf file, since this isn’t delivered anymore by latest distribution releases. In order to do that, install the
system-config-display package, and then run:
Simply setting the desired driver in the
xorg.conf file, however, is not enough, because the nv driver is interfering with nouveau, so you will likely get this error message when trying to run the X server:
The PCI device has a kernel module claiming it.
This driver cannot operate until it has been unloaded
(EE) No devices detected.
Fatal server error:
no screens found
That means you need to disable the nouveau driver e.g. by modifying your
/etc/grub.conf file and adding the following at the end of the line starting with
Then you can simply reboot your system and that’s it, you should be done.
August 9th, 2010 — Documentation
Sylpheed FAQ revision 2.2 released on 2010-08-09
New revision of Sylpheed FAQ has been officially released from the Sylpheed Documentation Project to reflect changes in the upcoming Sylpheed 3.1.
You can view the FAQ either as a multi-page or single-page HTML document at:
Or download it together with source DocBook XML files in a .tar.gz or .zip archive from:
The source DocBook XML files are also available in the Project CVS repository, see:
Changes over the past 4 months
* New Q&A "Can I run multiple instances of Sylpheed?"
* New Q&A "Execute command for my dynamic signature seems not to be working!"
* Updated Q&A on environment variables
* Updated Q&A on automatic name completion
* Updated Q&A on Sylpheed plug-ins
* Other minor edits throughout the document
* To better comply with the GFDL license, the source DocBook XML files together with the plain text copy of the GFDL license and appropriate legal notice are now distributed in tarballs with each documentation release, and the exact way of how to obtain the source files is mentioned explicitly in the document; thanks to Ricardo Mones for pointing these legal issues out
All other information on the Sylpheed Documentation Project including how to contribute to the documentation effort is available at:
Contributors to this and previous releases
Paul Kater, Jens Oberender, Francois Barriere, Olivier Delhomme, Petr Kovar
August 4th, 2010 — GNOME
Usenet may not be as popular as it used to be years ago, but this worldwide net definitely still has its users. According to Wikipedia, it was established in 1980, so Usenet users observe a nice anniversary this year. Given its popular stance throughout the Internet history, there exist, without much surprise, many FLOSS solutions to access Usenet, or to work with the NNTP. Using GNOME software, you can accomplish it e.g. with the official GNOME PIM Evolution.
And then there’s Pan, quite minimalist, HIG respecting, easy-to-use newsreader for GNOME, that has been developed for a decade. Unfortunately, owing to the limited time resources of the main Pan developer Charles Kerr (whom you might also know from the Transmission project), the Pan development has slowed down considerably during the last two or three years, and it’s now officially in hiatus.
The community around the newsreader will hopefully be able to organize itself enough to resume and continue with the active development in the Pan official repository hosted on git.gnome.org. Luckily for Pan users, there’s a competent developer K. Haley around who has been working on Pan during the last few years, although not in the official repository.
Anyway, what the project now needs is, preferably, a bunch of volunteers with interest in Usenet and NNTP who are willing to lend a hand and get involved in the project.
The much needed roles include:
- a developer with experience in C++, to help out with the Pan main development, to review and accept patches that got accumulated during the years in the “Pan” product at bugzilla.gnome.org,
- possibly a developer who may be willing to take over the maintainership in the future, once it is needed, or in case K. Haley will resolve to participate in the project not as the maintainer,
- bug triagers, patch reviewers, testers,
- people who are willing to work on user documentation and Pan website that might be migrated from its current location, and which is in serious need of getting up-to-date; I, for myself, am willing to work on these and will appreciate any help,
- translators who are thankfully willing to work on Pan without break, as the last commits to the official Pan repository are those from the GNOME Translation Project members,
- users, users, users.
If interested, please contact the Pan community that gather together on the following mailing lists:
Naturally, as with other FLOSS projects, every help and every contributor is welcome. TIA.
July 31st, 2010 — GNOME
Being syndicated on Planet GNOME from yesterday, I’d like to thank Lucas Rocha for adding me to the list!
If interested, you can read more about me as a GNOME contributor e.g. on live.gnome.org, and about intentions behind this blog in my first post.
Okay, enough of self-promotion for now, back to the work.
July 29th, 2010 — Documentation
Last summer I found myself delving into documentation writing & maintaining. My email client of choice, Sylpheed, which many FLOSS users might remember from early 2000’s when it was one of the first full-featured GTK+ clients (at least as far as I know), was distributed with a set of long-time unmaintained documentation that quickly became more or less completely out-of-date.
In fact, it wasn’t touched for over three years and was left in a state of partially completed LinuxDoc SGML to DocBook 4 XML migration. The first thing that needed to be done, even before finishing the migration and previewing the actual documentation, was obviously evaluating whether it is actually worthwhile for a new maintainer without deep knowledge of the given maintenance rules & processes to try to continue with updating the original documents, or whether it would be better to start from the scratch, i.e. with a complete rewrite.
Since I wasn’t feeling very adventurous and my previous experiences were quite limited, to say the least, I eventually sticked with the former option.
Now I can say that the biggest obstacle is not a need to understand DocBook 4, neither it is adapting work flow procedures and writing style of documents with numerous authors in their history, it is actually dealing with the GNU Free Documentation License, the license the said documentation is distributed under. I believe that many rants have been written before already about this piece of legal text, so I won’t repeat others here. Still, I’m quite grateful for Creative Commons licenses as something fresh and needful coming into scene and replacing GFDL within many documentation teams & efforts. I wish it was more feasible for maintainers to relicense a documentation work done by many (inactive and probably unreachable) authors from GFDL to CC. That being said, the clause that appeared in the GFDL version 1.3 is/was apparently not very helpful, unless you wanted to relicense content on Wikipedia or something very similar.
Anyhow, what I have found in my experience (IANAL) to be least amusing about GFDL is:
- The idea behind invariant sections etc. (Yes, I can understand those good intentions that went pretty wrong, I would say.)
- The definition of a transparent format.
- The need for distributing the full text of the license along with a licensed document.
- The thing that a generated HTML file with included GFDL copy (i.e. a generated file converted from DocBook XML source files) is or at least might be seen as an opaque copy, that is one needs to distribute a (plain text) file with full text of the license along with HTML file distribution (be it via tarball or any other standard channel). This issue was brought to my attention by Ricardo Mones and discussed on the Sylpheed mailing list.
So after crying my eyes out, I realize that this is my first blog post that might land on Planet Fedora, so I would like to say hi to all the hopefully interested readers out there! You can find most of my Czech localization & documentation writing contributions upstream, however, with me joining the Fedora Czech translation team ca. two years ago, and enjoying my encounters with the Fedora l10n infrastructure, notably the Fedora Transifex instance, ever since. Though I do remember the times when translate.fedoraproject.org was running a Damned Lies fork, also. OK, not too far away history, anyway.