Debian (titchy) Vs. Debian Eabi ?

[ZDevil]

Presence of broken packages is a fact.

So file bugs instead of saying 'Presence of broken packages is a fact' in some random forum.

Again, I will do so. I have kept asking people here around which packages are broken. Many say so so I assume that's true. But I do need to find out exactly which.
And if the devs know that, some notes in the wiki would prove a great time-saver.
And this is a bit tricky to me. I do use some (content of extracted) Angstrom packages, but only when the arm or armel Debian packages fail.
Can I really ask questions about armel in the Angstrom mailing list?

[koen]

Presence of broken packages is a fact.

So file bugs instead of saying 'Presence of broken packages is a fact' in some random forum.

Again, I will do so. I have kept asking people here around which packages are broken. Many say so so I assume that's true. But I do need to find out exactly which.
And if the devs know that

The devs don't know until someone files a bug. Why the $*(@$*(@ heck do you think I keep hammering on people to file bugs?!?!

, some notes in the wiki would prove a great time-saver.

You're confusing wiki and bugtracker. At one point people started adding bugs into the wiki (that were already in the bugtracker), bugs got fixed and the person that added the bugs to the wiki didn't remove them, so lots of people thought the bugs were still present ("the wiki says so"). To avoid that:

* The bugtracker is the only place that has bugs and is canonical
* The wiki is unofficial and not the gospel truth (notice that is isn't on angstrom-distribtion.org).

Once a wiki page reaches a certain level of correctness it is removed from the wiki and put on angstrom-distribtion.org (e.g. like the c7x0 series install instructions).
The bugzilla also has meta bugs like http://bugs.openembedded.org/showdependenc...hide_resolved=0 or http://bugs.openembedded.org/showdependenc...hide_resolved=0 where you can see all the bugs for a given subject (am ipaq and altboot in this case).

[ZDevil]

Fair enough. Why don't you simply stick this in the Angstrom wiki and/or the very first page of the homepage?

The very first page of Angtrom seems to me is more for communication between developers than to the ordinary users.
It simply states the draconian rules without telling the ordinary users what they are.

By ordinary users I refer to those not very conversant with complicated shell scripts and commands, let alone knowing the ins and outs of the kernel.
(Does the users of Angstrom need to know this to be "qualified" to use it?)

Your view already assumes a good deal of *nix background of the end-users. Prove that I am wrong.
Angstrom is a great product. But can you push this to clients who are *nix noobs?  Ok, I am a newbie. I've got stuck. So, go file a bug or check out the bugs in the bugtracker...

Bugtracker is for the developers and advanced users to talk about things seriously techinically.
Wikis and forums are for all the users and developers to communicate in a friendly way.
Mailinglists too. But frankly the format and organization grow a bit old fashioned and it is not convenient to search for relevant information. Digging into the "threads" is time-consuming and somethings doesn't pay off.
Forums can get pretty messy once the threads grow so big. I agree with your opinion about the bad side of forums.
Wikis retrieve, summarize and update useful information. Not for discussion but can grow as a result of the discussions in forums and mailing lists.

[tux]

 Glad to see a constructive conversation is going on!  

By the way ZDevil, I seem to be making progress with OABI+OABI! Hope I'm not kidding myself.

[koen]

Bugtracker is for the developers and advanced users to talk about things seriously techinically.

Anyone can file a bug, developer can then have a technical discussion in that bug. As a user you don't have to follow or understand that discussion, just make some effort to provide the information the developers ask for. It really isn't rocket science, filing a bug is as easy as 'abiword crashes on startup', the developers can then do some digging and notice abiword only crashes when using the clearlooks theme and gtk 2.6, but not with other themes or clearlooks + gtk 2.8.

[ZDevil]

Anyone can file a bug, developer can then have a technical discussion in that bug. As a user you don't have to follow or understand that discussion, just make some effort to provide the information the developers ask for. It really isn't rocket science, filing a bug is as easy as 'abiword crashes on startup', the developers can then do some digging and notice abiword only crashes when using the clearlooks theme and gtk 2.6, but not with other themes or clearlooks + gtk 2.8.

If it's that simple and userfriendly, then :
-- why so few users file bugs there? (perhaps because Angstrom just works)
-- what is the essential difference between forums and bugtrackers?

Just ask a new noob user. What do you think of when you look for support and useful tips & tricks & easy fixes in the product's website?
A. Forum
B. Wiki
C. Bugtracker

[Chero]

I changed my view on things a bit and edited post 2 of this thread.

Chero.
[size=](only fools never change their mind)[/size]

[koen]

-- why so few users file bugs there? (perhaps because Angstrom just works)

Dunno

-- what is the essential difference between forums and bugtrackers?

A bugtracker is a *central* place where only *bugs* get discussed. This is a zaurus forum, angstrom is not only about zaurusses. This forums lose messages, attachements and is down every day that ends in -day. Almost all our developers dislike forums for their idle chatter, interface and find them hard to navigate. For other devices there are other forums, and it seems every user expects the developers to monitor and use 6 different forums instead of getting emailed by the bugtracker and spending the time gained on solving bugs. For most developers this is done in their spare time after work. Do you really expect them to spend an hour each day reading each and every post on a number of different fora because there might be a bug mentioned somewhere? Whatever you expect, they don't do that, they just read the 6 to 10 emails from the bugtracker to see if they can fix something.
There's a new & closed bug summary posted every week on the OE list for people that aren't on the buglog list.

It looks something like this:

Code: [Select]

Bugs fixed:
bug_id resolution     short_desc
3135 FIXED     'unexport' attribute of variable still doesn't work sometimes
2714 FIXED     [PATCH] lock mtd-utils to a specific git version so builds
will not break unexpectedly
3061 DUPLICATE     usb-gadget is broken (at least in card reader and
file-backed mode)
3155 FIXED     Patch for libgstmad to generate 16-bit output instead of 32-bit
3179 FIXED     dbus fails to start on x86/Angstrom
3183 INVALID     maxima-5.13.0-r0-do_configure
3184 FIXED     [PATCH] reanimate xorg-app/mkfontscale recipies
3185 FIXED     [PATCH] reanimate xorg-app/xkbutils recipie
3186 FIXED     [PATCH] reanimate xorg-app/xlsfonts recipie
3187 FIXED     [PATCH] reanimate xorg-app/xkbcomp recipie and add
xkbcomp-native
3191 FIXED     openmoko-dates2-0.1.0+svnr640-r0-do_populate_staging
3194 WORKSFORME     2007-10-20 'bitbake gpe-image' failed on several
missing DEPENDS
3199 FIXED     midori fails to build with syntax error in document.h
3202 FIXED     Add a machine definition for MPC8323E-RDB
3203 FIXED     updater.sh.tosa is a forbidden file
3206 FIXED     Add uClibc configuration for MPC8323E-RDB
3207 FIXED     Add linux configuration for MPC8323E-RDB


Bugs opened:
bug_id resolution     short_desc
3181      Building Angstrom  for x86 without IMAGE_FSTYPES defined in
local.conf results in error
3182      bb.data.setVar(PARALLEL_MAKE
3200      rm_work could do with the ability to specify specific packages to
keep
3174      sound problem over CE-RH2.
3175      Toshiba eseries kernel recipe
3176      empathy-0.14-r0-do_configure
3177      Sepukku.bbclass can't attach logs anymore
3178      binutils 2.18 / uclibc : linking error
3180      bitbake gtk+ fails on 20071020 (and back at least a week) due to
missing DEPENDS
3188      [PATCH] fix libx11-native package
3189      [PATCH] new fakeroot package to fix download failure
3190      [PATCH] fix mkfontdir-native download location
3192      Linksys WCF-11 Card works on OZ Opie
3193      source code of screeensaver unable to be downloaded
3195      can't mount cf card on collie
3196      gpephone-image fails on librecord due to problems in Makefile.am
and configure.ac
3197      Proposed keymap for Zaurus spitz and akita
3198      syslogd fails to start at boot with busybox 1.7.2
3201      Fix: missing dependency for xdm
3204      libxml2-2.6.29-r5: do_populate_staging: failed
3205      binutils-2.18-r1 fails to link with libiberty.a
3208      Tree 1.5.1 doesn't build
3209      Git fetcher failing on x86_64 installs (Fedora 7)
3210      minimo_cvs fails to build; cannot access nsBuildID.h

In total 24 bugs have been created and 17 bugs were closed.
So if people after reading this still refuse to file their bugs to the bugtracker: Learn to live with the problem, you don't want to invest time in a bugreport, so we wont to invest time in a fix, since we don't know the problem is there.

And I repeat, since it doesn't seem to sink in: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.
I repeat again: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.

[tux]

I changed my view on things a bit and edited post 2 of this thread.

Chero.
[size=](only fools never change their mind)[/size]

 I'm going to go through this thread to see if it will help me with the X problems I'm wading through in trying to do the OABI + OABI thing. But it might change my mind also!

[ZDevil]

No, there is nothing wrong in it.
I also remember things felt so difficult to set up as discussed in the "Uboot and Debian" thread months ago.
The major reason why I try the EABI is to find out how things may be different now. The kernels are newer. The installation steps seem less complicated. Many tricks and tips have surfaced. And of course the EABI performance is hard to resist ...
The Angstrom kernels are indeed great works thanks to the devs' ardent and clever improvement.  (Despite the missing information for new users to get started)
And since last time my system screwed up, I was thinking to myself: "See? I've already tried quite a number of kernels and installation methods. Why not have a go with EABI? If things go bad I can always fall back to OABI because I am already pretty familiar with the process."

Right now I am writing the guide, including the bits for setting up EABI. Will post it later tonight.  

[Chero]

No, there is nothing wrong in it.
I also remember things felt so difficult to set up as discussed in the "Uboot and Debian" thread months ago.
The major reason why I try the EABI is to find out how things may be different now. The kernels are newer. The installation steps seem less complicated. Many tricks and tips have surfaced. And of course the EABI performance is hard to resist ...
The Angstrom kernels are indeed great works thanks to the devs' ardent and clever improvement.  (Despite the missing information for new users to get started)
And since last time my system screwed up, I was thinking to myself: "See? I've already tried quite a number of kernels and installation methods. Why not have a go with EABI? If things go bad I can always fall back to OABI because I am already pretty familiar with the process."

Right now I am writing the guide, including the bits for setting up EABI. Will post it later tonight.  

There's still nothing wrong in it - I only added some things I read here and in another thread.

[ZDevil]

*SNIPPED*
There's a new & closed bug summary posted every week on the OE list for people that aren't on the buglog list.

It looks something like this:

Code: [Select]

Bugs fixed:
bug_id resolution     short_desc
3135 FIXED     'unexport' attribute of variable still doesn't work sometimes
2714 FIXED     [PATCH] lock mtd-utils to a specific git version so builds
will not break unexpectedly
*SNIPPED*
In total 24 bugs have been created and 17 bugs were closed.
So if people after reading this still refuse to file their bugs to the bugtracker: Learn to live with the problem, you don't want to invest time in a bugreport, so we wont to invest time in a fix, since we don't know the problem is there.

To be honest: it is a huge mistake to tell/force/indoctrinate/preach to your target users/clients to read this, especially if you are targeting "the world". This is no end-user support.

And I repeat, since it doesn't seem to sink in: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.
I repeat again: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.

Everybody knows that. Many many of us also go to other forums, mailing lists, wikis, web sites, what have you.  Like many others, I have never treated *this* forum as MY forum. But for the Zaurus if you can tell me there is another better forum than this, do drop me a line. And I don't care how Angstrom is used in the rest of the world because I am not using those other things.  
After all, a gadget, or several gadgets, only occupy a very tiny part in our life.

[Chero]

So if people after reading this still refuse to file their bugs to the bugtracker: Learn to live with the problem, you don't want to invest time in a bugreport, so we wont to invest time in a fix, since we don't know the problem is there.

And I repeat, since it doesn't seem to sink in: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.
I repeat again: angstrom (and OE) is not only about zaurusses, this forum in not the only forum in the world.

sorry, I can't resist to interfere.

I think that when we (at least I) have found a bug, that is not due to the user, we will file a bug report.
But, when we ask/discuss about a problem in a forum before doing so, a lot of problems happen to be no bugs but user problems. You see, that exactly is the benefit of a forum, without forums you would get lots and lots of bug-reports about non existing bugs. And I can assure you that is very discouraging cause you have to spend your time filtering the real bugs from the imagined ones. I really wish there was a community or a user-group using a forum about the CAD-apps I'm developing in real life. I can answer 90% of the questions they ask me with RTFM, but I point them to the correct page of the manual instead.
Let the forum exist and do what it's supposed to do : filter.

BTW : who ever claimed this was the only forum in the world ?
BTW : who ever told the devs to read through this forums to find bugs ?

But, when a dev does happen to read a user problem in a forum while occasionally passing by and he/she knows the solution, then I think that he/she can as well point the user in the right direction instead of making a laugh ...

Chero.

[Chero]

After all, a gadget, or several gadgets, only occupy a very tiny part in our life.

My wife says other things about my gadgets, especially my zaurusses (zauri) ...  

[ZDevil]

And ...

I think, koen, you may have missed the big point of the enthusiastic discussions in this new Debian forum over this couple of weeks.

We do like the Angstrom kernels. If, as you think, these jerks just keep blaming Angstrom for the imaginary bugs or find faults with your fruit of hard work or simply get anti-social and delusional, then i am sorry, you are mistaken. We like the Angstrom kernels. What else are we playing with here?
Discussions in a forum is such a good thing, to us, probably not to you as the architect of the Angstrom. We tell each others about different possibilities. We share our experiences. We find a lot of fun using and playing with the Z in a friendly community. We don't expect the mighty Angstrom devs to come down and salvage the poor and ignorant souls. Even less do we expect the devs mock at us by not knowing (how come!) the deep secrets about kernel (for many of us it is just a black box).

We are discussing a lot of things, some may be bugs, many more may be user errors. This is just like every other forum in the world.
Have we spread our wild imaginary bugs around the world? Do we plot together against the great Angstrom project?
No.
But we keep poking around. And look, many of us have succeeded in getting Debian running, and with the Angstrom kernels (or their modified/patched builds).
And we do appreciate both Angstrom and Debian a lot.

If you feel uneasy about these unjustified and biased and lousy discussions, prove yourself and your good work in your website and your wiki.