Re: Check if fsck will be run on a partition

[Theo de Raadt]

> > How can I know if the partition needs to be checked by fsck, I'd like to 
> > test that.
> 
> Check the output of dumpfs.  clean=0 means that the filesystem is
> dirty and fsck should be run.

It is cheaper to just run fsck.  If it has no work to do, it finishes.

Re: Check if fsck will be run on a partition

[Christian Weisgerber]

On 2018-04-01, Mik J  wrote:

> How can I know if the partition needs to be checked by fsck, I'd like to test 
> that.

Check the output of dumpfs.  clean=0 means that the filesystem is
dirty and fsck should be run.

-- 
Christian "naddy" Weisgerber  na...@mips.inka.de

Check if fsck will be run on a partition

[Mik J]

Hello,
I have a script that mounts a partition and it works well except when the 
partition needs to be fsck checked.How can I know if the partition needs to be 
checked by fsck, I'd like to test that.If the partition needs to be checked by 
fsck, I run fsck firstElse I mount the partition
Happy easter

Re: bug tracking system for OpenBSD

[Christoph R. Murauer]

Thanks Bob, Theo, Chris for your detailed answers.

>From Ingos answer :

> ... Later on, when somebody would have time to look, such unprocessd
reports have low visibility because they are nothing but an old
posting on a mailing list, drowning in a lot of noise from invalid
and resolved reports.

Means for me, that a table of bugs / features are required. Instead of
reinventing the wheel and add a new interface, create a table (could
be a static site) including posting date, arch (release, stable,
current), subject, if it is not enough a description, bug / feature,
open / incomplete / WIP / closed. And, also a link back to bugs@.

As a new bug / feature is posted at bugs@ add them to the table as
open. If a dev answers, change it to incomplete / WIP and, if it is
solved then change it to closed.

A ideal world would provide complete bug reports so, the user has to
provide as much informations as possible. The dev could ask for more
informations if needed - if a dev didn´t ask for more details or the
user don´t provide them then nothing happened.

It is only a idea. I also have no team but maybe it is a option to
start for one person (from a given date and, let the past) where
others could join. Or, as Theo said, keep it as it is ... and expect,
that users ask again about - lets say - forgotten bugs / features.

Regards,

Christoph

Status of X i386 openbsd 6.2 on x200

[flipchan]

Hello all,

I have tried to installed 6.1 and 6.2 on a thinkpad x200 it works but X does 
work ...

Its works great with 6.0 but then i dont get the good 6.2 packages and features 
such as syspatch. 


It seems lika well known problem:
https://marc.info/?l=openbsd-bugs=150506076421862=2


Does anyone know the status of this/ if anyone is working on this ?


-- 
Take Care Sincerely flipchan layerprox dev

Re: bug tracking system for OpenBSD

[Consus]

On 11:01 Sun 01 Apr, Stuart Henderson wrote:
> Please don't.

I wasn't going to. This guy jsut asked for the feedback so I wrote him.

Chinese fonts on chrome

[Pau]

Hello:

I'm trying to get chrome or firefox display Chinese characters but I'm
failing. I have installed zh-wqy-bitmapfont and zh-fonts-kc and then I
run
fc-cache but they do not show up. In the font menu of chrome I cannot
find any option for those fonts to be used. Firefox is fine though.

I also followed this:

https://en.wikipedia.org/wiki/Help:Multilingual_support_%28East_Asian%29#Other_UNIX_Distributions

without success. Any idea of what I might be doing wrong? Thanks.

Re: bug tracking system for OpenBSD

[Chris Bennett]

I think the issue of triage deserves a more clear explanation of just
how huge a task this really is.

A team with members who are well familiar with each architecture and
having access to the actual models/devices showing the problem.
Even a "fix" might break other models within the same architecture!
So, access to a ton of hardware is a must have.

Knowing who is working in the area needed and if they actually want to
work on the bug at all. Look at the hackathon reports:
"I had planned on working on issue X, but I found myself working instead
on issue Y with so and so". Why not? Who doesn't want to enjoy themself?

They need to have members who know enough OpenBSD internals to at least
be able to make reasonable guesses about where the problem is probably
at, else how can they figure out who might be able to work on it.

OK, this is starting to sound like forming a team of developers.

Dedicated enough to really keep at it.
This is starting to sound like a team getting paid money to really keep
at it! Do you really want to spend 4 hours every night after going to
work and coming home tired to do rather difficult work?

So, at the very least, a mountain of hardware, expertise and probably
money too.

I think Theo's comment about hitting 'd' on many bug reports is
realistic reality.

It's a shame, but I occasionally (always?) see 'submit a diff' more
often than I see diff's submitted.

One new talented developer is worth more than trying to create a bug
tracker.

Chris Bennett

Re: bug tracking system for OpenBSD

[Theo de Raadt]

> No such team exists.  the tool used is irrelevant

Well, we have that tool today: it is a mailing list.

Also, we have a team which triages bugs on there: the developers.

Is it perfect?  No.

Do things slip through the cracks?  Sure.  Because not enough people
triage.  Not enough developers.

Would a system which tracks the things which slip through the cracks
help?  No.  A mailing list archive makes it obvious that things
slipped through the cracks, but it also makes it obvious there is one
reason only why that is the case:  THE VOLUME IS TOO HIGH.

Organization is only one piece of the puzzle.  The data won't organize
itself.  The missing piece is "Organizers", who triage.  There are no
new GROUPS of people being offered to do this work.

One person can't do it.  Yet organizing people together to help us is
hard, that's for sure.  Impossible I'd say.

Would it be awesome to have a gigantic pile of shitty things
collecting in some web gui that none of us look at?  No.  Yet that is
what other project have.

It has to be cleaned up by someone who does traige.  Some projects
have companies associated who pay people to do it.  Other projects
let it build up, and when the quality gets to messy, it gets ignored.
OpenBSD was there before: Since it was being ignored I took the approach
that helped our developers have a better experience: I deleted the bug
tracker, and our source tree was better off for it.

Creating a portal where developers can do that is a waste of time.
They won't do the triage you expect them to.

I triage bugs by hitting 'd' in my mail reader.  That includes ones
I assume someone else will look at.  It includes ones I just fixed.
It includes ones from previous releases.  It includes ones I can't
look at now because I have other things going on.

And tough shit, that's the way it is.  I have no additional cycles to
handle yet-another interface to the bug reports people submit.  None
of the developers have additional cycles.  Creating a new interface,
which will be full of crap because *noone is offering additional
triage services*, won't help.

triage/tracking is only a small piece of the process of making OpenBSD
a good operating system.

So I believe we might as well stay with what we have.

This is a volunteer open source project.  That does not mean just
anyone can walk on in and volunteer the addition of a "new process"
and expect such an addition will work within our processes.

OpenBSD as a development process works.  This thread is full of people
who assume that what they see working in other places will work here.

Re: bug tracking system for OpenBSD

[Bob Beck]

Christoph, your conversation is distracting.

Nobody gives a damn about the tool. Everyone gives a damn about the triage.

I hate to break it to you, but you are not the first person to broach
this discusson.

The only way this would work is with a dedicated team of people to
triage each area and clean it up constantly.

No such team exists.  the tool used is irrelevant


On Sun, Apr 1, 2018 at 9:44 AM, Christoph R. Murauer  wrote:
> My question was serious. I am not the enemy but I think this thing
> will only work if the people who use it accept / like to use it and so
> on.
>
>> bug tracking software is 1% of the solution.  At least 80% of the
> work is triage, and noone on this thread is serious about doing
> that.
>
>

Re: bug tracking system for OpenBSD

[Theo de Raadt]

Do we want the 1% solution?  No.

Will we accept something which comes with a full triage team?  Yes.

Is a triage team being offered? No.

> My question was serious. I am not the enemy but I think this thing
> will only work if the people who use it accept / like to use it and so
> on.
> 
> > bug tracking software is 1% of the solution.  At least 80% of the
> work is triage, and noone on this thread is serious about doing
> that.
> 
> 

Re: bug tracking system for OpenBSD

[Christoph R. Murauer]

My question was serious. I am not the enemy but I think this thing
will only work if the people who use it accept / like to use it and so
on.

> bug tracking software is 1% of the solution.  At least 80% of the
work is triage, and noone on this thread is serious about doing
that.

Re: bug tracking system for OpenBSD

[Theo de Raadt]

> Would the devs accept / use a bug tracker ? I ask because I find start
> something without the devs is burning time (see the .ru domain, the UI
> posts ... ).

We'd be happy to accept a bug-tracker which is slavishly quality-managed
and continually purified and kept current by a team of dedicated people

Oh, but what is being talked about it is?

We're being told we can use some sort of fly-by-night web portal which
has had a full-of-junk mailing list imported into it

bug tracking software is 1% of the solution.  At least 80% of the work
is triage, and noone on this thread is serious about doing that.

Using stmp auth for local account with PHP scripts

[Markus Rosjat]

Hi there,  

There are simple ways of relaying local mails(connection on lo0 on port 25) to 
a other mailserver. This is oky for logs and stuff but what's about mails 
created by a php on the local webserver? His do I get smtpd to still do a auth 
with username and pwd on lo0? Is it possible or do I need to configure the 
"external" addr too for this purpose?

Regards

Markus

Re: bug tracking system for OpenBSD

[Christoph R. Murauer]

Thanks for your fast answer.

I thought before, that the manual triage was the point.

Would the devs accept / use a bug tracker ? I ask because I find start
something without the devs is burning time (see the .ru domain, the UI
posts ... ).

If someone of the developers is interested a off list discussion could
maybe help.

Regards,

Christoph


> Hi Christoph,
>
> Christoph R. Murauer wrote on Sun, Apr 01, 2018 at 01:56:47PM +0200:
>
>> not a problem from the OpenBSD developers ?
>
> There *is* an actual problem:  Some bug reports are not processed
> immediately because at the time they are posted, nobody happens to
> find the time to investigate.  Later on, when somebody would have
> time to look, such unprocessd reports have low visibility because
> they are nothing but an old posting on a mailing list, drowning in
> a lot of noise from invalid and resolved reports.
>
> The task of a bugtracker would be
>
>  1. to collect as FEW as possible tickets -
> ideally only valid, unprocessed ones with complete information
>
>  2. to close them when resolved such that the number of open tickets
> stays low
>
> That's why manual triage and curation is the crucial point.
>
> Yours,
>   Ingo
>

Re: texmacs on 6.2

[Jeremie Courreges-Anglas]

On Sun, Apr 01 2018, Rudolf Sykora  wrote:
> Hello,
>
> I have probably a trivial (user) question.
>
> I've used texmacs on 6.1, amd64.
> Now I wanted to install it on 6.2, i386, but I can
> neither find a package nor i386 is mentioned
> among Archs in ports (the same for amd64, btw.).

>From the Makefile of the port:

  # XXX base clang: lots of undefined references to Qt functions
  COMPILER= base-gcc

> What can I do?

Try to fix TeXmacs so that it builds with clang; note that there are new
releases upstream.

> Why isn't i386 and amd64 available?

Because i386 and amd64 use clang as the base compiler.

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

Re: bug tracking system for OpenBSD

[Ingo Schwarze]

Hi Christoph,

Christoph R. Murauer wrote on Sun, Apr 01, 2018 at 01:56:47PM +0200:

> not a problem from the OpenBSD developers ?

There *is* an actual problem:  Some bug reports are not processed
immediately because at the time they are posted, nobody happens to
find the time to investigate.  Later on, when somebody would have
time to look, such unprocessd reports have low visibility because
they are nothing but an old posting on a mailing list, drowning in
a lot of noise from invalid and resolved reports.

The task of a bugtracker would be

 1. to collect as FEW as possible tickets -
ideally only valid, unprocessed ones with complete information

 2. to close them when resolved such that the number of open tickets
stays low

That's why manual triage and curation is the crucial point.

Yours,
  Ingo

Re: howto configure a virtual switch for vmd network

[Carlos Cardenas]

On Sun, Apr 01, 2018 at 10:04:08AM +0100, niya wrote:
> 
> > hi
> > 
> > how do i configure a virtual switch to connect the interfaces of several
> > vm's together.
> > the interfaces will have fixed addresses and i will also be bridging the
> > interfaces to the host.
> > i know how to set up the bridge and then configure it with vm.conf
> > but i not sure how to setup the switch and add the vio interfaces to the
> > switch.
> > 
> > shadrock
> 
> 
> i have read the following webpage and have the answers to the questions i
> asked i have?? a better understanding of how vmd works.
> http://blog.hermes-technology.de/openbsd/server/virtualmachine/network/2017/06/12/vmd-for-a-virtual-server-network.html
> thanks
> 

If you are using current or plan to use 6.3 when it's released, you need
to update your "switch" configuration in vm.conf as per:
https://www.openbsd.org/faq/upgrade63.html


+--+
Carlos

Re: PPPoE connection closing right after authentication?

[Mihai Popescu]

>The good news is that just PAP authentication should work, and does on 
>Windows.  The bad >news is that on OpenBSD I'm still not getting an IP address 
>from Tek/Telus after the Authenticate-Ack.

I remember some time ago, that someone requested some help with pppoe
implementation since his ISP was asking for a specific value of VLAN
field for a german ISP. I think someone coded right away that feature
in OpenBSD.
All the discussion is in the misc@.
In the meantime, you can email your ISP and explain that you are using
something else and you need some details about connection. First line
of support may think you are crazy, but ask for a more specialised
person and maybe they will tell you the full process.

texmacs on 6.2

[Rudolf Sykora]

Hello,

I have probably a trivial (user) question.

I've used texmacs on 6.1, amd64.
Now I wanted to install it on 6.2, i386, but I can
neither find a package nor i386 is mentioned
among Archs in ports (the same for amd64, btw.).

What can I do?
Why isn't i386 and amd64 available?

Thanks
Ruda

Re: bug tracking system for OpenBSD

[Christoph R. Murauer]

Just curious, could it be, that the problem is only a problem of the
OPs from the threads (bug tracker, github mirror and so on) and, not a
problem from the OpenBSD developers ?

If I read the last section from https://www.openbsd.org/report.html it
looks like that for me. I did not search the list but I can't
remember, that someone asks which things in sendbug(1) or bugs (or
their workflow) does not work for the developers / maintainers (and if
there are things like that - IMHO they will fix them). Maybe people
also don't want to accept that OpenBSD is not Linux.

IIRC Stuart (or others) mentioned not the first time, that a human had
to manage a bug tracker - if one is needed.

Maybe I am total wrong but thats are my 2 cents about it.


> Way to go guys, you've completely misunderstood the problem
> and therefore have no solution.
>
>

Re: PPPoE connection closing right after authentication?

[Jon Martin]

On Mon, Mar 26, 2018 at 09:47:18PM +0200, Sebastian Benoit wrote:
> Jon Martin(jmg...@gmail.com) on 2018.03.22 13:19:51 -0600:
>> 
>> To me this further indicates a "double authentication": a CHAP challenge
>> followed by PAP authentication.  I have no idea how to set up a config
>> to answer that though.
> 
> Yes, this is possible, and OpenBSD does not support this mode.

The good news is that just PAP authentication should work, and does on
Windows.  The bad news is that on OpenBSD I'm still not getting an IP
address from Tek/Telus after the Authenticate-Ack.

Re: bug tracking system for OpenBSD

[Stuart Henderson]

On 2018-03-31, Sergey Bronnikov  wrote:
> Setup another bugtracker which you like, connect it to bugs@ and send a
> link to this thread.

Please don't.

My suggestion of looking at bugs@ as a seed is for a *person* to read
new reports, collect information into one place, figure out if it
really is a bug, write up tickets. Sounds like hard work? Yes, it is.
A bug tracker needs curation/triage, that is not something that can be
automated.

If you look at a successful bug tracker, you will not be able to see
most of this work without grovelling through dead tickets because
the already-fixed issues and veiled support requests will no longer
be visible.

Re: howto configure a virtual switch for vmd network

[niya]

hi

how do i configure a virtual switch to connect the interfaces of 
several vm's together.
the interfaces will have fixed addresses and i will also be bridging 
the interfaces to the host.

i know how to set up the bridge and then configure it with vm.conf
but i not sure how to setup the switch and add the vio interfaces to 
the switch.


shadrock



i have read the following webpage and have the answers to the questions 
i asked i have  a better understanding of how vmd works.

http://blog.hermes-technology.de/openbsd/server/virtualmachine/network/2017/06/12/vmd-for-a-virtual-server-network.html
thanks