devel March 2004

devel@lists.fedoraproject.org

238 participants
275 discussions

Pango development
by Leonard den Ottolander 10 Mar '04

10 Mar '04

Hello, I am seeing crashes in gnome-terminal when I use the midnight commander. This issue seems to be pango related. I reported this issue 7 weeks ago. See http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=113284 . I think it might have to do something with the fact that I use ISO-8859-15 as my charset, since the crash seems to appear in a call to g_utf8_next_char(). Should g_utf8_next_char() handle other charsets graciously? If not, then how can I use my system with another charset than UTF-8? Shouldn't pango handle other charsets as well? Has this changed in 1.3? When studying the pango rpm I noticed a reference to http://bugzilla.gnome.org/show_bug.cgi?id=121488 . In that bug report Owen states: "Is it really worth rewriting stuff in the 1.2.x branch, which will shortly be dead." What does this comment mean with respect to the support of pango for Fedora 1 and RHEL 3? Leonard. -- mount -t life -o ro /dev/dna /genetic/research

5 10

software suspend?
by Rui Miguel Silva Seabra 10 Mar '04

10 Mar '04

Hi, Is software suspend enabled in Linux 2.6.3-2.1.242 ? I have tried to echo S4 > /proc/acpi/sleep but nothing happens. What could I be doing wrong? Thanks in advance, Rui -- + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? Please AVOID sending me WORD, EXCEL or POWERPOINT attachments. See http://www.fsf.org/philosophy/no-word-attachments.html

3 3

Re: In the service of Aunt Tillie -- Zero-configuration networking
by Stuart Cheshire 10 Mar '04

10 Mar '04

Oops. Sorry. Forgot to correct he subject line last time, so it said "Re: fedora-devel-list digest, Vol 1 #475 - 13 msgs" instead of "Re: In the service of Aunt Tillie -- Zero-configuration networking" Stephen Smoogen wrote: >A $50 ethernet printer would be cool since most of the ones I see are >always 200 more than the regular printer. A 802.11 would be ok until >people started paying kids to drive by and 'fax' you the latest sales >items. Security is IMPORTANT! Sadly, too few vendors care at all ("We recommend our products are only used behind a firewall" -- whatever that means). I'll say no more on the subject. >Well point me to the source, I will try to make a fedora.us RPM and put >it in the cooker by Friday or so. <http://developer.apple.com/darwin/projects/rendezvous/> Check out top-of-tree mDNSResponder module from CVS. (Follow the instructions labeled "For APSL projects.") cd mDNSPosix make os=linux sudo make install os=linux Then cd ../Clients and make to build a simple client showing how to use the APIs. Stuart Cheshire <cheshire(a)apple.com> * Wizard Without Portfolio, Apple Computer, Inc. * www.stuartcheshire.org

1 0

Fedora Bug Day Tomorrow: March 10th 2004: Bug Bugs Day
by Leonard den Ottolander 09 Mar '04

09 Mar '04

Fedora Bug Day: Help in tackling some bugs and take some weight of the shoulders of the maintainers. Pick your favourite package, poke it a little and see what comes out. March 10th, starting at 14:00 UTC (09:00 EST) at #fedora-bugs irc channel on freenode Join the discussion and help out with squashing bugs. If you think you are not up to fixing bugs by yourself you can always help out in investigating the validity of outstanding issues. Some info on Fedora Triage can be found in the fedora-triage-list archives https://lists.dulug.duke.edu/pipermail/fedora-triage-list/ or at the following URLs: http://tinyurl.com/ywma3 - Jef "the man" Spaleta's vision on Fedora Triage http://tinyurl.com/23alw - Jef's short term goals and long term plan Hope to see you around tomorrow, Leonard. -- mount -t life -o ro /dev/dna /genetic/research

1 0

Accidental string changes in anaconda?
by Christian Rose 09 Mar '04

09 Mar '04

Ok, so why someone seems to have reverted all string changes that have happened the last months in anaconda beats me. Anaconda has had many string changes of the type: mention of tool "redhat-config-xxx" => mention of tool "system-config-xxx" mention of "Red Hat Linux" => mention of %s (which might be replaced by "Red Hat Linux" or "Fedora Core" by a compile time switch) Now all those string changes are reverted, at least in the anaconda po files, so we're back to square zero: anaconda happily tells users to use redhat-config-xxx tools, and calls the product "Red Hat Linux" in a hardcoded way. Accident, or some strange strategy that beats me? Christian

3 4

RE: package shepherding
by Erik LaBianca 09 Mar '04

09 Mar '04

Alexander Wrote: <snip> > > You have vast amounts of possible volunteers on these lists, which if > > managed well could take some of the burden of your shoulders. I can see > > that individual developers are taking part in managing these resources, > > but I don't see an overall strategy. But then maybe I am expecting too > > much of a BSDish approach... > > All the possible volunteers on this lists are great, but the fact is > that if I have to personally hand-hold them I won't have time for > anything else, which might be a net loss to the project. Its very much a > fact that volunteers in free software projects have to be very driven > self-starters. It might be an unfortunate fact, but its been true for as > long as I've done free software, and I don't see it changing. > I think this is the real issue. It took me many years of using open source to even attempt to make a contribution. I think what people are hoping for here is some guidance as to how to make it easier to get people started with the project. I think that's what Jef's been trying to do with triage days, and for my own part, I'm trying to make the Extra's QA process more straightforward with my quickstart guide. If fedora can come up with a set of simple tasks, that nonetheless provide real help, the barrier for entry will be drastically lowered. Expecting everybody to actually "fix a bug" means you eliminate many potential contributors due to the sheer background knowledge required. The trick IMHO is figuring out ways to give input that are easy AND useful. --erik

2 1

Re: human sheparding ( was package shepherding )
by Jef Spaleta 09 Mar '04

09 Mar '04

Alexander Larsson wrote: > Also, there is all this talk about "Management" with I just don't > understand. Redhat is a company with managers, but Fedora is about free > software development. People are supposed to work on it because they > like to, to scratch itches, to learn stuff, to have fun. If all you want > is to have a Redhat fix bugs for you the best way do to that is to spend > money on Redhat Enterprise Linux, that means we'll get more paid > developers and will be able to fix more bugs (in fedora too, fedora is > what will become the next RHEL version) You've read the leadership draft document right? Now talking about micro-management of developer time maybe isn't something productive to talk about (unless i knew who your manager was and had a way to bribe them into telling you to do exactly what i wanted partitioned into 15 minute blocks of time). But certainly there is room to talk about Red Hat's need to manage in broad strokes, the large human capital resources represented not only in red hat employees, but also the enthusiastic but maybe somewhat out-of-touch volunteers. This is not going to be a very successful community experiment, if Red Hat contains full control of the high level policy and decision making that directs the distro...but doesn't do a very good job of communicating and directing volunteers. Everyone involved in a process gets paid in different ways. Volunteers, get paid either through control to shape the direction of the project or through access to information, and both forms of metaphysical volunteer payment seems to be lost in the mail when it comes to the Fedora project. Personally I'd say Red Hat misjudged the cost of sending out those checks, and didn't spend enough money on stamps. Or dropping the cutesy allegory... I don't think Red Hat has allocated enough internal resources aimed directly at the problem of building and managing the volunteer community aspects of the project. A few very motivated and extremely technically knowledgable people have broken through the barren wastelands of communication separating the nation-states of Fedora users and Fedora developers, but I wouldn't call that sort of vision quest something a majority of volunteers are skilled enough to do. There is a need here to manage aspects of communication with volunteers, to take the day-to-day burden of dealing with common volunteer/developer communication off the shoulders of the developers. And I'm not talking about individual developer time, I'm talking about someone specifically whose job it is to act as a volunteer coordinator. If the Fedora project, as a community project, is really important to Red Hat corporate, they will allocate manhours specifically to address the horrible task of figuring out how to recruit and retain community volunteer interest in a project whose overall goals are set not by community..but by Red Hat. If Red Hat does tap someone to be a volunteer coordinator, and they make the mistake of not hiring someone with training and experience dealing with volunteer husbandry in the brick and mortar world, might I suggest they look at the first 18 or so pages of this volunteer management handbook pdf: http://tinyurl.com/33ban The section outlining on how to start off on the wrong foot, in terms of volunteer organization..reads like a history of the Fedora project. Read the first 18 pages or so and grimace. Actually...everyone involved in Fedora development should read the first 18 or so pages of that volunteer workbook. The majority of the pdf, concerns itself with template tools as example tools to address many of the issues raised in the first 18 or so summary pages. In the darker moments of my day (when I'm calibrating optical beam paths in the dead of night) i sometimes wonder what exactly is the point of opening this process up to the 'community.' Is it all about the binary bits? Or is the real potential value not in the bits at all but in the building of a vibrant community process, where an average fedora user (who lets be honest is going to be less technically inclined than users of some of the more prominent community development model distros) can learn how to contribute. People just wandering through, scratching their own itch, in my mind, isn't a really useful definition of community...its sort of like calling the mass of people standing in line at the DMV a community. Or too but it another way, idle technical proficient manpower at this point is probably in short supply, since anyone who has an itch to scratch already has a sourceforge project listing for some sort of mp3 playing something or other. Fedora either has to tap into existing volunteer manpower that is being used for other things by offering those people something of value inherent in the fedora contribution process, or by developing and nurturing less technically inclined user so that they can become active contributors. -jef"Communities are built through proactive leadership reaching out and building a process by which people feel empowered to be responsible for the project. I'm still very hopeful such a community can be built from the people standing in line at the Fedora's DMV"spaleta

1 0

Help Needed: thunderbird-0.5 and other stuff
by Warren Togami 09 Mar '04

09 Mar '04

https://bugzilla.fedora.us/show_bug.cgi?id=1113 Firefox and Thunderbird are the KILLER combination. fedora.us Extras firefox-0.8 was released last week, but we still need your help in order to get the thunderbird-0.5 into acceptable release shape. My personal time is too short to devote to this package now. The last remaining blocking is probably a complete rewrite of the /usr/bin/thunderbird script so the following requirements are met: 1) It must launch first time without problems while mozilla or firefox is running. 2) It must not pop-up that ugly error message when you attempt to launch it while it is already running. Ideally it should pop-up the already running thunderbird session if it is launched again. Read the recent bugzilla entries about the broken xremote behavior to see why this is a challenge. 3) It must not have that ugly lock-file that can get "stuck" after a crash or unexpected reboot. 4) It must be able to launch a composer window from both a state where thunderbird is not yet running, and thunderbird is already running. Again this is a challenge because xremote is broken. 5) https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=109738#c6 It must integrate with the defaults set by the Preferred Applications chooser as set by control-center after this patch above. I aim for this patch to eventually go into both upstream control-center and FC2, which should make the "Preferred Applications does not work" problem largely go away in many applications like htmlview, evolution, and a few others. A few other packages need simple patches for this to behave properly though. I highly suspect that after the above requirements are met, firefox-0.8 and thunderbird-0.5 will Just Work(TM) together and be nicely integrated into the desktop. After this is ready, I hope to have a nicely illustrated guide with graphics explaining the simple installation procedure for FC1 and FC2 (just install two RPMS from Extras), set the Preferred Applications, and suddenly Linux desktop nirvana. https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=114524 Also note that the above firefox & thunderbird pair will work together, while mozilla will be painful with thunderbird without this patch. I really hope for this patch to be applied before FC2, but I promised someone that I would never bug him directly about this anymore. I urge others to politely ask him instead. <evil grin> Below is a list of stuff that I really want to do, but there is no way I will have time to do so anytime soon. I ask for your help as a community to resolve. Delegate to Community --------------------- * https://bugzilla.fedora.us/show_bug.cgi?id=1167 fedora-rpmdevtools release I am not really sure why they included that "custom" stuff... thomasvs mentioned that he had a lot of ideas for kmodhelper We really need to get this update released, so please help. (Ville is on vacation, so another fedora.us release manager should probably roll another candidate package. <nudge Enrico>) * mdadm initscript proposal I never had the chance to write everything that I wanted to say about this. I have tested many different patches that would be good for a generic initscript solution, to be default instead of raidtools. I am sure however that many of you have similar ideas. Please discuss. * mailman https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=105638 1) rpm -V failure due to bytecode compiler A few other python packages suffer from this problem too, IIRC 2) issues during uninstall if you have postfix Just ugly error messages that could be cleaned up 3) leftover directories after uninstall (FC1) Ownership problems, should be fixed to prevent possible breakage * metadata DNS txt investigation Alan Cox had the interesting idea of using the DNS TXT field in order to distribute metadata for apt/up2date/yum when we switch to the new metadata format. Using it would be interesting and efficient because of DNS automatic propagation and caching. It could be used to distribute an official mirror list, or repository hashes, and since you can have up to 65K bytes within DNS TXT, it could easily be GPG signed. This idea while sounding extremely cool could easily be infeasible though, as I personally know very little about DNS. * https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=115218 Isolate radeon LCD blanking patch If anyone cares and wants to give this a try. No guarantee that mharris would accept it into the tree though if the patch turns out to be too invasive. * More KDE keybinding default fixes & unification with Mozilla If you care about KDE and Konqueror please continue pushing & discussing what I had posted on this list a while ago. These fixes are EASY but someone clueful only has to take the time to test it, get opinions and provide exact patches. * links lockup, kill -9... why? Seems to happen only on certain machines but not others... extremely annoying and I could never figure out why this happens. * cpuspeed module detection Originally the cpufreq scaling drivers were in rawhide kernels as modules, and you had to manually configure /etc/cpuspeed.conf in order to make it actually work. For now Arjan has included those modules into recent rawhide kernels, however it would be nice in the long run to write proper autodetection for this. Arjan and Dave Jones want such detection to be added to x86info, which already knows about the processor, while it would also need to make module decisions based upon chipset in some corner cases. Pretty low priority now that it is built-in to the kernel and not very large. Warren Togami wtogami(a)redhat.com

3 2

Re: package shepherding
by Jef Spaleta 09 Mar '04

09 Mar '04

Mark McLoughlin wrote: > Hmm, I'm not creating a separate list of bugs outside of bugzilla > is such a great idea I concur. I would actually like to incorporate some of the structure in the table template back into bugzilla if possible... Example, having a field to link upstream bug reports/urls instead of burying that linkage between fedora and upstream in comments. And perhaps work towards a summary view of this sort as a possible way to view triaged bugzilla data. Taking a good look at the whats really in the table, other than the upstream bugreport linkage field..all you really need to do is use a few more keywords to be able to regenerate this sort of summary report only from bugzilla data. And to be really clear, what i especially liked about the summary template, is its potential to highlight issues that could use community patch building/verification/review for low priority but nagging issues. I already suckered notting into creating the PATCH keyword which i've been trying to use for community patches that have had some confirmation from the community that it works..sort of a 2 pair of eyes rule for community supplied patches to bugreports. > + Give package shepherds bugzilla permissions so they can > actually modify bugs. That process is in place... finally, its just woefully inefficient. Since having the ability to get community editting rights, my goal has been to get one or two people rights every week. I have not met that goal for a variety of reasons, some of which are out of my control being internal red hat communication issues. > + Come up with a triage guide that explains to shepherds how they > should: Why re-invent the wheel... gnome's guide to triage works and I use as a reference repeatedly in announcements. I'm tempted to just steal the document and replace Gnome with Fedora. That being said, its on my todo list, right after i meet critical mass of community triage volunteers who have bugzilla editting rights, so that I don't have to concentrate what time i do have to spend just following up on other people's attempts to identify bugs that need triage. I want to get to the point where there are just enough people with editting rights so i can sit back and watch them argue about the details...and take notes. > - Set the bug's priority/severity Just want to comment on this, I've been in conversations with fedora developers about these fields...and as far as i can tell these fields are not used consistently by fedora developers...which makes its difficult for triage to use effectively..so i'm inclined to ignore these fields completely. The rest of those bullet items, all have their own particular wrinkles that the few of us who do have triage rights are trying to wrap our heads around, and don't forget about the complication of legacy thrown into the mix. I jokingly call legacy...downstream. Though i think maybe PASSTHEBUCK or HOTPOTATO might make better keywords for bugs that roll off in fedora legacy's general direction. -jef"fedora developers are like information pinatas...you blindly poke them hard a few times you get rewarded with information spewed out all over the floor"spaleta

7 15

I want my Galeon back!!!
by Jean Francois Martinez 09 Mar '04

09 Mar '04

The need for Galeon is not as a a browser but because it is the only quarter decent tool for viewing Gnome help files. Neither Mozilla or Epiphany understand the ghelp "protocol" and for yelp it is an incredible piece of crap. Start gnumeric and try browsing its help. Yelp needed 4 (FOUR) CPU minutes to display something on a PII/400 and it was not a RAM problem but the stupid thing trying to parse the entire help hierarchy instead of the root file. A so long delay makes Gnumeric as good as if it were undocumented because the user cannot afford a so long delay. And you cannot print from Yelp (PLEASE, don't tell me there is a utility for printing ghelp files, the user does not see it on the Gnumeric help men). I have some users who wanted a spreadsheet running on a Unix platform and given that StarOffice is too heavy for the machine I thought Gnumeric was the answer but without a decent help browser I have had to give up. Fortunately I had an old Redhat 7.3 (with Galeon) otherwise I would have had to advise my users to use Excel. First time Galeon is started it prompts the user for he wants to use it for displaying gnome-help files (ie the user does not need to navigate in any config utility). In addition Galeon displays them nearly instantaneously AND it can print. I want my Galeon back!!!!! -- Jean Francois Martinez <jfm512(a)free.fr>

8 10

← Newer
1
...
15
16
17
18
19
20
21
...
28
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

devel March 2004