Bug#996907: xfce4-power-manager: Screen blank is active but not configurable when xfce4-power-manager is NOT installed

[Phil Endecott]

Package: xfce4-power-manager
Version: 4.16.0-1
Severity: normal

Dear Maintainer,

I did a miminal install of xfce, i.e. "apt-get install xfce4", which 
did NOT install xfce4-power-manager.

The resulting system had screen blank configured to activate after 
10 minutes, but there was no settings panel to adjust or disable 
this.

After a lot of googling and trial-and-error I discovered that the 
display power management settings are part of this xfce4-power-manager 
package. After installing this, I was able to adjust the timeouts.

It seems to me that (1) maybe screen blank settings should be installed 
as part of a default xfce4 install, and/or (2) if the screen blank 
settings panel is not installed, then screen blank should be disabled 
by default.


Regards, Phil.

Bug#996773: barrier: man pages lack description of crypto mode

[Phil Endecott]

Package: barrier
Version: 2.3.3+dfsg-1.1
Severity: normal

Dear Maintainer,

The man pages for barriers and barrierc don't say anything about 
how to use the --enable-crypto mode.

This time I was setting up a new client. By looking at the output 
when run with -d DEBUG -f, I determined that on the client I needed 
to create a file
~/.local/share/barrier/SSL/Fingerprints/TrustedServers.txt 
containing the fingerprint of the server, which I found in 
~/.local/share/barrier/SSL/Fingerprints/Local.txt on the server.
Is there supposed to be a simpler way to do that?

(I have a Linux barrier server, one Mac client (homebrew package) 
and one Linux client. I enabled crypto on the server because 
the barrier GUI on the Mac doesn't have an option to disable it. 
I recall having some trouble doing that, but I forget the details 
now.)


Thanks, Phil.

Bug#869535: lightdm: service fails to start due to incorrect data dir perms

[Phil Endecott]

I've also seem this warning message and the root.root permissions on 
/var/lib/lightdm/data/

This system previously had a very minimal (debootstrap'd) 
Debian Bullseye installation. I then did:

# apt-get install openbox lightdm

and rebooted. X starts and shows a background image but there are 
no login controls or anything else. On this first boot, 
/var/log/lightdm/lightdm.log says:

[+0.06s] WARNING: Could not enumerate user data directory 
/var/lib/lightdm/data: Error opening directory '/var/lib/lightdm/data': No such 
file or directory

Looking at the permissions:

# ls -ld /var
drwxr-xr-x 12 root root 4096 Oct 11 16:36 /var

# ls -ld /var/lib
drwxr-xr-x 23 root root 4096 Oct 17 16:34 /var/lib

# ls -ld /var/lib/lightdm/
drwxr-x--- 4 lightdm lightdm 4096 Oct 17 16:36 /var/lib/lightdm/

# ls -ld /var/lib/lightdm/data/
drwxr-xr-x 3 root root 4096 Oct 17 16:36 /var/lib/lightdm/data/

# ls -ld /var/lib/lightdm/data/lightdm/
drwxrwx--- 2 lightdm lightdm 4096 Oct 17 16:36 /var/lib/lightdm/data/lightdm/

# ls -l /var/lib/lightdm/data/lightdm/
total 0

Observe the ownership going down the hierarchy is root then lightdm 
then root then lightdm, which seems peculiar.

On the second boot, this warning was not shown - presumably the 
missing directory was created on the first boot albeit with the 
surprising ownership. lightdm still didn't work, but that was due 
to other issues.


Regards, Phil.

Bug#947319: lightdm: fails during startup (no invite)

[Phil Endecott]

I have also seen this message in the lightdm log:

[+0.06s] WARNING: Error getting user list from 
org.freedesktop.Accounts: 
GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name 
org.freedesktop.Accounts was not provided by any .service files


This went away when I installed accountsservice.

I believe that either lightdm itself or maybe the greeter package
should depend on accountsservice.


Regards, Phil.

Bug#996489: awscli: Cryptic error if awscli and boto versions mismatch

[Phil Endecott]

Package: awscli
Version: 1.20.53-1
Severity: normal

Dear Maintainer,

On a system running bullseye (11.1), I wanted to install a newer 
version of awscli. So I added sources.list entries for testing, set 
APT::Default-Release "bullseye"; and then

# apt-get install -t testing install awscli

This installed awscli 1.20.53-1, but it did not change 
python3-botocore which remained at version 1.20.0+repack-1 
(from bullseye) because awscli only requires python3-botocore 
>= 1.20.0.

"aws help" worked but this failed:

  $ aws lambda update-function-configuration help
  'StructureShape' object has no attribute 'is_document_type'

I then upgraded python3-botocore to the bookworm version, 
1.21.53+repack-1:

# apt-get install -t testing python3-botocore

and now awscli seems to function correctly.

It seems to me that awscli should have a tighter dependency on 
python3-botocore.

(I am a bit confused by the version numbers; it looks as if 
awscli 20.x corresponds to boto3 21.x ? What's going on there?)


Thanks, Phil.

Bug#990242: libapache2-mod-xsendfile: /usr/share/doc/libapache2-mod-xsendfile/Readme.html contains Google Analytics tracking script

[Phil Endecott]

Package: libapache2-mod-xsendfile
Version: 0.12-2
Severity: normal

Dear Maintainer,

/usr/share/doc/libapache2-mod-xsendfile/Readme.html has the following 
at the end:




  var _gaq = _gaq || [];
  _gaq.push(['_setAccount', 'UA-2632517-2']);
  _gaq.push(['_setDomainName', '.tn123.org']);
  _gaq.push(['_trackPageview']);

  (function() {
var ga = document.createElement('script'); ga.type = 
'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 
'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; 
s.parentNode.insertBefore(ga, s);
  })();




This seems undesirable. Please consider removing it.


Regards, Phil.



-- System Information:
Debian Release: 10.10
  APT prefers stable
  APT policy: (990, 'stable'), (500, 'oldstable')
Architecture: amd64 (x86_64)

-- no debconf information

Bug#988060: barrier: misleading error when too many open files

[Phil Endecott]

The reason for the too-many-open-files appears to be that 
sockets are leaked when client connections fail due to SSL 
errors.

The particular SSL error in my case seemed to be due to the 
server certificate having expired. Barrier seems to automagically 
create SSL certs that are valid for one year.

I have been starting the Barrier server from my .config/openbox/autostart :
barriers -c /home/phil/.synergy.conf --enable-crypto

It appears that that doesn't do anything about expired SSL 
certs. On the other hand, if I start the barrier GUI:
$ barrier --no-tray

then something does seem to create a certificate - at least, 
it did when I removed the expired cert; I'm not sure if it 
actually checks the expiry date. (Looking at the source the 
only attempt to generate certs is in the GUI code, but I'm 
not sure if that is run on expiry or only if no cert exists.)

In summary there are three issues here:

1. Misleading error reporting on server when too-many-open-files 
prevents the SSL certificate file from being read.

2. Server socket leak when SSL error (certificate expired?) causes 
client connection to fail.

3. Server SSL certificate expiry does not seem to be handled when 
barriers is invoked directly, rather than the barrier GUI; it does 
not attempt to generate new certificates, nor does it report any 
useful error about expired certs to syslog when client connections 
fail. (Nor do clients when they receive expired certs.)


Thanks, Phil.

Bug#988060: barrier: misleading error when too many open files

[Phil Endecott]

Package: barrier
Version: 2.1.2+dfsg-1~bpo9+1
Severity: normal

Dear Maintainer,

I have Barrier server version 2.1.2+dfsg-1~bpo9+1 on a Debian box 
and Barrier client version 2.3.3 on a Mac (from Homebrew).

After restarting the Mac, Barrier failed to start. Its log says 
repeatedly:

ERROR: ssl error occurred (system call failure)
ERROR: failed to connect to secure socket

On the Debian system, syslog shows:

ERROR: ssl certificate doesn't exist: 
/home/phil/.local/share/barrier/SSL/Barrier.pem

But that file does exist:

-rw--- 1 phil phil 2798 Feb 20  2020 
/home/phil/.local/share/barrier/SSL/Barrier.pem

Investigating with strace:

[pid 15729] openat(AT_FDCWD, "/home/phil/.local/share/barrier/SSL/Barrier.pem", 
O_RDONLY) = -1 EMFILE (Too many open files)

So there appears to be an issue with misleading error reporting.

Looking at the source:
https://github.com/debauchee/barrier/blob/master/src/lib/net/SecureSocket.cpp
Line 344
It just checks std::ifstream::good().

At the very minimum, the error message "ssl certificate doesn't exist" 
should be replaced with "ssl certificate file could not be opened". 
Much better, it should use errno and e.g. strerror to give a useful 
message.

(Note I'm using an older version of Barrier but I'm looking at the 
current source code.)

I'm continuing to look at what the underlying issue might be.


-- System Information:
Debian Release: 9.9
  APT prefers stable
  APT policy: (990, 'stable'), (500, 'oldstable-updates'), (500, 'oldstable')
Architecture: arm64 (aarch64)
Foreign Architectures: armhf

Kernel: Linux 3.14.29+ (SMP w/2 CPU cores; PREEMPT)
Locale: LANG=en_GB.utf8, LC_CTYPE=en_GB.utf8 (charmap=UTF-8), 
LANGUAGE=en_GB.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages barrier depends on:
ii  libavahi-compat-libdnssd1  0.7-4+b1
ii  libc6  2.28-10
ii  libcurl3   7.52.1-5+deb9u9
ii  libgcc11:6.3.0-18+deb9u1
ii  libice62:1.0.9-2
ii  libqt5core5a   5.7.1+dfsg-3+deb9u1
ii  libqt5gui5 5.7.1+dfsg-3+deb9u1
ii  libqt5network5 5.7.1+dfsg-3+deb9u1
ii  libqt5widgets5 5.7.1+dfsg-3+deb9u1
ii  libsm6 2:1.2.2-1+b3
ii  libssl1.1  1.1.0k-1~deb9u1
ii  libstdc++6 6.3.0-18+deb9u1
ii  libx11-6   2:1.6.4-3+deb9u1
ii  libxext6   2:1.3.3-1+b2
ii  libxi6 2:1.7.9-1
ii  libxinerama1   2:1.1.3-1+b3
ii  libxrandr2 2:1.5.1-1
ii  libxtst6   2:1.2.3-1
ii  openssl1.1.0k-1~deb9u1

barrier recommends no packages.

barrier suggests no packages.

-- no debconf information

Bug#966575: Symbol `grub_calloc' not found: AWS instance

[Phil Endecott]

I've been affected by this issue on an AWS EC2 instance. 

The particular issue with AWS is that the device names 
may depend on the particular instance types; on newer 
hardware disks appear as NVMe devices, and on older 
hardware as /dev/xvd? or /dev/sd?.  The Debian cloud 
instances have unattended updates enabled and I guess 
that the grub update was installed while the instance 
was running on hardware with NVMe disks, while it had 
originally been installed when it was running on older 
hardware.  My fstab refers to the disks using UUIDs; I 
believe that some distributions may install symlinks in 
/dev to avoid problems like this but Debian doesn't 
seem to.

Rescue is not too difficult once you know how: detach 
the borked instance's root volume, attach it to another 
(temporary) instance, repair, and move it back.  To 
make it appear as the root volume when moved back you 
need to give exactly the same device name as is shown 
as "Root Device Name" in the image's AMI details; it 
took me a long time to work out that I needed to enter 
"xvda" rather than "/dev/xvda" here (YMMV).

To actually repair it I followed the advice in this bug 
to bind-mount /dev,proc,sys and chroot.  I then tried 
Colin's advice in message 184 but without success:

# dpkg-reconfigure grub-pc
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-4.19.0-10-amd64
Found initrd image: /boot/initrd.img-4.19.0-10-amd64
Found linux image: /boot/vmlinuz-4.19.0-9-amd64
Found initrd image: /boot/initrd.img-4.19.0-9-amd64
Found linux image: /boot/vmlinuz-4.9.0-5-amd64
Found initrd image: /boot/initrd.img-4.9.0-5-amd64
  WARNING: Device /dev/nvme0n1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p14 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p15 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme1n1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme1n1p1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p14 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme0n1p15 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme1n1 not initialized in udev database even after 
waiting 1000 microseconds.
  WARNING: Device /dev/nvme1n1p1 not initialized in udev database even after 
waiting 1000 microseconds.
Found Debian GNU/Linux 10 (buster) on /dev/nvme0n1p1
done

There were a couple of curses dialogs during that asking about 
kernel command lines, for which I accepted the defaults.  Note 
that /dev/nvme0n1p1 is the rescue system's root device, not the 
one that needs repairing.  This didn't work.

So I tried again with grub-install:

# grub-install /dev/nvme1n1
Installing for i386-pc platform.
Installation finished. No error reported.

(Note nvme1n1, not nvme1 or nvme1n1p1.)  This has worked, in as 
much as the system now works again.  I take it that I should now 
dpkg-reconfigure from within the restarted system (though that 
will not prevent future breakage if I move to hardware with 
different device names, right?).

I hope a fix is planned for this; cloud images can have quite long 
uptimes so there may still be a lot of undiscovered affected systems.


Regards, Phil.

Bug#915127: cloud.debian.org: Please add AWS image for new ARM instances

[Phil Endecott]

Noah Meyerhans wrote:

It's on its way.


Glad to hear it, many thanks!

Bug#915127: cloud.debian.org: Please add AWS image for new ARM instances

[Phil Endecott]

Package: cloud.debian.org
Severity: wishlist

Dear Maintainer,

AWS have recently announced new instance types that use the 64-bit ARM 
(aka aarch64) architecture.  Machine images are currently available for 
"Amazon Linux 2", RHEL, Ubuntu and Fedora.  It would be great to also 
have official Debian images.


Many thanks,  Phil.

Bug#839961: Upstream is now at v5.7.0

[Phil Endecott]

Hi,

libjs-d3 in Debian has not been updated for more than 2 years.
Version 3.5.17 isn't really very useful.
Please consider updating it, or perhaps officially orphan it if
you are no longer able to maintain it.


Thanks, Phil.

Bug#851099: "auto-apt update" tries to fetch wrong URL

[Phil Endecott]

I think this would have worked in Wheezy, but not since Jessie.

I.e. this is OK:

http://ftp.debian.org/debian/dists/wheezy/Contents-armhf.gz

But since Jessie it needs "main" or whatever between the Debian version
and the "Contents-ARCH.gz" bit:

http://ftp.debian.org/debian/dists/jessie/main/Contents-armhf.gz

Has auto-apt really not worked since Wheezy?

Ah!  Now I see bug #813976 (with the unhelpful title "repair auto-apt"
which proposes splicing "main" into the URL exactly like that.

I'm going to (try to) increase the severity, as it seems that this
package has been broken for years; is no-one using it?

Bug#851099: "auto-apt update" tries to fetch wrong URL

[Phil Endecott]

Package: auto-apt
Version: 0.3.24

I believe that the first thing I must do after installing auto-apt is
"auto-apt update", right?

# auto-apt update
Downloading http://ftp.debian.org/debian/dists/stretch 
Contents-arm64.gz ...

2017-01-11 23:47:13 ERROR 404: Not Found.


My /etc/apt/sources.list contains:
deb http://ftp.debian.org/debian stretch main contrib non-free


It seems that auto-apt invokes wget to fetch 
/debian/dists/stretch/Contents-arm64.gz
rather than /debian/dists/stretch/main/Contents-arm64.gz .

Bug#848152: Please add Multi-Arch: foreign to ntopng-data

[Phil Endecott]

Package: ntopng-data
Version: 2.2+dfsg1-2

Hi,

Currently ntopng doesn't install properly on a multi-arch
system (e.g. arm64/armhf).  I believe that the fix for this
is to add Multi-Arch: foreign to ntopng-data in debian/control.

(If you want to read my tale of woe about this, see:
https://lists.debian.org/debian-arm/2016/12/msg00033.html )

Thanks, Phil.

Bug#843896: Logging issues fixed in new upstream version

[Phil Endecott]

Hi,

I believe that these issues are fixed in the new upstream
version, 1.1.0.


Thanks, Phil.

Bug#840061: Enable I2C/IPMI and EDAC for X-Gene

[Phil Endecott]

Hi Ben,

Ben Hutchings wrote:

It went into 4.8.7.  I'll enable the drivers for the next upload, which
includes 4.8.7.


Thanks.

Have you also enabled the xgene EDAC and DMA configs that I mentioned
towards the end of the original report?

Also, I'm confused by the changelog; have you accidentally pasted the
upstream 4.8.6 changelog instead of 4.8.7?


Cheers,  Phil.

Bug#843896: Fix logging and merge config in mod_authnz_pam

[Phil Endecott]

Package: libapache2-mod-authnz-pam
Version: 1.0.1-1
Tags: patch

I have fixed a couple of issues with mod_authnz_pam's logging:

1. I've changed the log level for the "passed" message from 
notice to info.  "Notice" is special in Apache; it is included 
in the log irrespective of the LogLevel setting.  Successful 
authentication isn't such an important event that it must be 
logged unconditionally.

2. I've made the module name appear correctly in the log statements; 
previously they looked like:

[Tue Nov 08 19:17:24.340078 2016] [:notice] [pid 13338:tid 140735843676592] 
[client 192.168.1.42:51330] mod_authnz_pam: PAM authentication passed for user 
phil

Now they look like:

[Tue Nov 08 19:17:24.340078 2016] [authnz_pam:notice] [pid 13338:tid 
140735843676592] [client 192.168.1.42:51330] PAM authentication passed for user 
phil

I think that this was a result of not using the AP_DECLARE_MODULE 
macro.  Or possibly of using ap_log_rerror() rather than ap_log_error().

I think that this correction of the module name also means that you 
can successfully using the module name in a LogLevel statement, i.e. 
LogLevel authnz_pam:Debug.

I've forwarded these as suggestions to mod_authnz_pam's author, Jan 
Pazdziora, who replies:

> thank you for the suggestions and analysis. I will review and 
> process them later but due to other work it might take some time 
> for me to do another release incorporating these changes.

Not related to logging, nor submitted to Jan, I've also added a 
configuration merge function so that configuration is inherited from 
a parent directory to a subdirectory.

My patch follows.  As this might be useful for other users, and as 
upstream might be slow to catch up, please consider adding this to 
the Debian version of the module.

Thanks,  Phil.


--- mod_authnz_pam-1.0.1/mod_authnz_pam.c   2015-11-09 21:42:37.0 
+
+++ libapache2-mod-authnz-pam-1.0.1/mod_authnz_pam.c2016-11-10 
13:49:20.702396637 +
@@ -40,6 +40,18 @@ static void * create_dir_conf(apr_pool_t
return cfg;
 }
 
+static void *merge_dir_conf(apr_pool_t *pool, void *basev, void *overv)
+{
+   authnz_pam_config_rec * base = basev;
+   authnz_pam_config_rec * over = overv;
+   authnz_pam_config_rec * cfg = apr_pcalloc(pool, 
sizeof(authnz_pam_config_rec));
+   
+   cfg->pam_service = over->pam_service ? over->pam_service : 
base->pam_service;
+   cfg->expired_redirect_url = over->expired_redirect_url ? 
over->expired_redirect_url : base->expired_redirect_url;
+
+   return cfg;
+}
+
 static const command_rec authnz_pam_cmds[] = {
AP_INIT_TAKE1("AuthPAMService", ap_set_string_slot,
(void *)APR_OFFSETOF(authnz_pam_config_rec, pam_service),
@@ -170,8 +182,8 @@ static authn_status pam_authenticate_wit
if (ret == PAM_NEW_AUTHTOK_REQD) {
authnz_pam_config_rec * conf = 
ap_get_module_config(r->per_dir_config, &authnz_pam_module);
if (conf && conf->expired_redirect_url) {
-   ap_log_error(APLOG_MARK, APLOG_ERR, 0, 
r->server,
-   "mod_authnz_pam: 
PAM_NEW_AUTHTOK_REQD: redirect to [%s]",
+   ap_log_rerror(APLOG_MARK, APLOG_ERR, 0, 
r,
+   "PAM_NEW_AUTHTOK_REQD: redirect 
to [%s]",
conf->expired_redirect_url);
apr_table_addn(r->headers_out, 
"Location", format_location(r, conf->expired_redirect_url, login));
return HTTP_TEMPORARY_REDIRECT;
@@ -181,14 +193,14 @@ static authn_status pam_authenticate_wit
}
if (ret != PAM_SUCCESS) {
const char * strerr = pam_strerror(pamh, ret);
-   ap_log_error(APLOG_MARK, APLOG_WARNING, 0, r->server, 
"mod_authnz_pam: %s %s: %s", stage, param, strerr);
+   ap_log_rerror(APLOG_MARK, APLOG_WARNING, 0, r, "%s %s: %s", 
stage, param, strerr);
apr_table_setn(r->subprocess_env, 
_EXTERNAL_AUTH_ERROR_ENV_NAME, apr_pstrdup(r->pool, strerr));
pam_end(pamh, ret);
return AUTH_DENIED;
}
apr_table_setn(r->subprocess_env, _REMOTE_USER_ENV_NAME, login);
r->user = apr_pstrdup(r->pool, login);
-   ap_log_error(APLOG_MARK, APLOG_NOTICE, 0, r->server, "mod_authnz_pam: 
PAM authentication passed for user %s", login);
+   ap_log_rerror(APLOG_MARK, APLOG_INFO, 0, r, "PAM authentication passed 
for user %s", login);
pam_end(pamh, ret);
return AUTH_GRANTED;
 }
@@ -270,10 +282,10 @@ static void register_hooks(apr_pool_t *
APR_REGISTER_OPTIONAL_FN(pam_authenticate_with_login_password);
 }
 
-module AP_MODULE_DECLARE_DATA authnz_pam_module = {
+AP_DECLARE_MODULE(authnz_pam) = {
STANDARD20_MODULE_STUFF,
cre

Bug#840061: Enable I2C/IPMI and EDAC for X-Gene

[Phil Endecott]

The buffer overflow fix is now upstream here:

  
https://git.kernel.org/cgit/linux/kernel/git/wsa/linux.git/commit/drivers/i2c?id=603616017c35f4d0fbdbcace72adf9bf949c4a65

I expect it to be in 4.10, and also some stable updates.

Bug#840061: Enable I2C/IPMI and EDAC for X-Gene

[Phil Endecott]

I've now tested the same changes on 4.8~rc8-1~exp1 and they work there too.

Bug#840061: Enable I2C/IPMI and EDAC for X-Gene

[Phil Endecott]

Source: linux
Version: 4.7.5-1

Hi,

I have a Gigabyte MP30-AR1 board with an ARM64 X-Gene1 processor.

To access its temperature, fan and voltage sensors using IPMI the 
following kernel configs are needed:

CONFIG_MAILBOX=y
CONFIG_XGENE_SLIMPRO_MBOX=m
CONFIG_I2C_XGENE_SLIMPRO=m
CONFIG_IPMI_HANDLER=m
CONFIG_IPMI_SSIF=m
CONFIG_IPMI_DEVICE_INTERFACE=m

It should then possible to read the sensors using
# ipmi-sensors --driver-type=OPENIPMI --driver-device=/dev/ipmi0

..but in practice it seems to fail due to a buffer overrun:

--- drivers/i2c/busses/i2c-xgene-slimpro.c.orig 2016-10-07 18:25:09.357520311 
+
+++ drivers/i2c/busses/i2c-xgene-slimpro.c  2016-10-07 18:25:22.425573533 
+
@@ -105,7 +105,7 @@
struct mbox_chan *mbox_chan;
struct mbox_client mbox_client;
struct completion rd_complete;
-   u8 dma_buffer[I2C_SMBUS_BLOCK_MAX];
+   u8 dma_buffer[I2C_SMBUS_BLOCK_MAX+2];
u32 *resp_msg;
 };

I'm in the process of reporting this upstream.

The processor also has error detection and correction for its main 
memory and caches; this can be enabled with:

CONFIG_EDAC=y
CONFIG_EDAC_MM_EDAC=m
CONFIG_EDAC_XGENE=m

I've also turned this on, though I don't know if it does much:

CONFIG_XGENE_DMA=m

Note that I'm currently using 4.7.5, but I'm reasonably sure that 
none of this has changed significantly in 4.8.

Please consider enabling these configs, and possibly the i2c patch, 
in the Debian kernel.


Thanks,  Phil.

Bug#838297: man page for make-kpgs should refer to nproc --all, not grep /proc/cpuinfo

[Phil Endecott]

Package: kernel-package
Version: 13.018

man make-kpkg says:

   CONCURRENCY_LEVEL
  If  defined,  this  environment variable sets the concurrency
  level of make used to compile the kernel and the modules set
  using -j flags to the sub make in the build target of make-kpkg.
  Should be a (small) integer,  if used. You can get the current
  number of CPUs using the command:

   grep -c '^processor' /proc/cpuinfo

This grep command does not give a useful answer on machines that use the
"hotplug" cpufreq governor, which switches off unused CPUs; these switched
off CPUs do not appear in /proc/cpuinfo.  A better way to get the 
number of

CPUs is the nproc command.  In this case, "nproc --all" would be appropriate.
(nproc is in coreutils, so everyone should have it.)

Bug#838292: kernel-package/README.gz refers to module-init-tools, not kmod

[Phil Endecott]

Package: kernel-package
Version: 13.018

/usr/share/doc/kernel-package/README.gz says:

 The packages suggested are:
 devel:gcc, libc5-dev/libc6-dev, binutils, make, and, for intel
   x86 platforms, bin86 (non-Intel platforms don't need
   this), modutils (or module-init-tools for 2.5.x+ kernels).

IIUC, module-init-tools was replaced by a package called "kmod" some 
time ago.

Bug#838036: Postgis package has very many GUI dependencies

[Phil Endecott]

Package: postgresql-9.5-postgis-2.2
Version: 2.2.2+dfsg-5

Attempting to install postgis on a headless server, i.e. a
system that doesn't have X or any sort of desktop environment
installed, wants to bring in vast numbers of dependencies.

The issue seems to be that the postgis package depends on
liblwgeom, which depends on libsfcgal, which depends on
libopenscenegraph.  This depends on lots of GUI things
including mesa and qt.  (There may also be other dependency
chains that I've not noticed.)

It seems reasonable to me that I should be able to add geographic
primitives to PostgreSQL in order to, for example, execute the
query that appears at the top of the PostGIS website:

SELECT superhero.name
FROM city, superhero
WHERE ST_Contains(city.geom, superhero.geom)
AND city.name = 'Gotham';

- without also having to install vast numbers of unwanted
GUI packages.  Would it be possible to downgrade some of the
dependencies to "suggests", or perhaps offer a "lite" package
that avoids them entirely?

Thanks, Phil.

Bug#837715: Debain installer fails on Gigabyte MP30-AR1 (ARM64 X-Gene1)

[Phil Endecott]

Package: debian-installer

I've been asked to file this bug to track my attempts to install 
Debian on my MP30-AR1.  This is an ARM64 board with an X-Gene 1 
processor.  Please see this mailing list thread for more details:

  https://lists.debian.org/debian-arm/2016/09/msg00022.html

In summary, this board is virtually identical to the MP30-AR0 
except that it is supplied with UEFI, not U-Boot, in its boot 
flash.  Apparently there are various known issues that prevent 
the Debian installer from running, including bugs in the 
device tree and lack of ACPI support in the Debian kernel.  
On the other hand, both the CentOS and Ubuntu installers run 
and install working systems.

The other distributions' installers that do work are:
  http://mirror.centos.org/altarch/7.2.1603/isos/aarch64/
  
http://ports.ubuntu.com/ubuntu-ports/dists/yakkety/main/installer-arm64/current/images/netboot/
 (dated 2016-08-28 01:23)

Neither is running a particularly new kernel (4.2 and 4.4 respectively) 
so either they've backported required fixes or the problems are in the 
configuration (or conceivably GRUB, I suppose.)  The Ubuntu kernel 
seems to use device tree while the CentOS one uses ACPI.

The Debian installer fails as follows:

EFI stub: Booting Linux Kernel...
ConvertPages: Incompatible memory types
EFI stub: EFI_RNG_PROTOCOL unavailable, no randomness supplied
EFI stub: Using DTB from configuration table
EFI stub: Exiting boot services and installing virtual address map...
OSBootEvent = Success
L3c Cache: 8MB
[ 0.00] Booting Linux on physical CPU 0x0
[ 0.00] Linux version 4.6.0-1-arm64 (debian-ker...@lists.debian.org) (gcc 
version 5.4.0 20160609 (Debian 5.4.0-4) ) #1 SMP Debian 4.6.2-2 (2016-06-25)
[ 0.00] Boot CPU: AArch64 Processor [500f0001]
[ 0.00] earlycon: uart8250 at MMIO32 0x1c02 (options '')
[ 0.00] bootconsole [uart8250] enabled
[ 0.00] efi: Getting EFI parameters from FDT:
[ 0.00] EFI v2.40 by American Megatrends
[ 0.00] efi: ACPI 2.0=0x807ff43000 ESRT=0x807e161e18 SMBIOS 3.0=0x807e161c18
[ 0.00] Moving initrd from [4079b0f000-407ac457c3] to [7e0ac9000-7e1bff7c3]
[ 0.00] Unhandled fault: synchronous external abort (0x9610) at 
0xffbffe63e078
[ 0.00] Internal error: : 9610 [#1] SMP
[ 0.00] Modules linked in:
[ 0.00] CPU: 0 PID: 0 Comm: swapper Not tainted 4.6.0-1-arm64 #1 Debian 
4.6.2-2
[ 0.00] Hardware name: Gigabyte X-Gene MP30-AR0 board (DT)
[ 0.00] task: ff8008b88800 ti: ff8008b78000 task.ti: 
ff8008b78000
[ 0.00] PC is at __memcpy+0x100/0x180
[ 0.00] LR is at copy_from_early_mem+0x6c/0x94


Cheers,  Phil.

Bug#827079: triggerhappy: Generated event tables are empty, so nothing works.

[Phil Endecott]

Package: triggerhappy
Version: 0.3.4-2
Severity: grave
Justification: renders package unusable

Dear Maintainer,

Triggerhappy is entirely non-functional, because it doesn't recognise
any key names:

# thd --dump /dev/input/event4(a keyboard)
Unknown (null) event id on /dev/input/event4: 28 (value 0)
Unknown (null) event id on /dev/input/event4: 30 (value 1)
Unknown (null) event id on /dev/input/event4: 30 (value 0)

Having fetched the source using "apt source" and tried to build it myself,
I get an entirely empty evtable_KEY.h.

My guess is that this is the result of a change of behaviour of gcc, and
the awk 'NR==1 {print $2}' in the Makefile is broken by it.

Looking at your git repo I see that you changed the Makefile last year and
the way that it gets the event codes from the linux include file has changed.
This version does build correctly.  But the Debian package does not appear
to have picked up the change (and the version number is still 0.3.4).

(I'm doing this on a slightly exotic platform, but I don't see anything
architecture-related here.)

Regards,  Phil.


-- System Information:
Debian Release: stretch/sid
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: arm64 (aarch64)
Foreign Architectures: armhf

Kernel: Linux 3.14.29+ (SMP w/2 CPU cores; PREEMPT)
Locale: LANG=en_GB.utf8, LC_CTYPE=en_GB.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages triggerhappy depends on:
ii  libc6  2.22-11

triggerhappy recommends no packages.

triggerhappy suggests no packages.

-- Configuration Files:
/etc/default/triggerhappy changed [not included]

-- no debconf information

Bug#602424: elmerfem: More armel qreal trouble

[Phil Endecott]

Hi Adam,

I saw your message to debian-arm, a list which I sometimes read but am 
not subscribed to, hence this off-list reply.  I don't know anything 
about qt but I have been learning OpenGL recently.  Here's my suggestion:


inline void glGetQrealv(GLenum e, GLfloat* data) { glGetFloatv(e,data); }
inline void glGetQrealv(GLenum e, GLdouble* data) { 
glGetDoublev(e,data); }


Then replace the calls to glGetDoublev() that want qreals with calls to 
glGetQrealv().  This should work on platforms where qreal is of either type.


Good luck!


Phil.







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#399911: Still need to use lynx to enable remote administration

[Phil Endecott]

Four years later, and I'm installing CUPS for only the second time 
ever.  And I still need to use lynx to turn on remote administration.  
Am I the only one who runs CUPS on something that's not a full-featured 
desktop machine?







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527973: [Pkg-sysvinit-devel] Bug#527973: initscripts needs a newish ifupdown else mountnfs fails

[Phil Endecott]

Petter Reinholdtsen wrote:

[Phil Endecott 2009-05-19]

Hmmm, maybe that's not the (only) problem; NFS mounts now seem to
come up most but not all of the time.  I need to investigate some
more.


What did you find out.  Without more information, I believe this
report should be closed as not a bug.


Hi Petter,

I'm still seeing the problem and have not investigated thoroughly.  I 
understand that Debian is going to do (or has already done?) some major 
surgery on the init system, so presumably all issues like this one will 
need to be reviewed.  I agree that it should be closed.



Phil.







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#535952: xprint: Xprt can't find symbol PrinterFontRegisterFpeFunctions and fails to start

[Phil Endecott]

Hi Julien,

Julien Cristau wrote:

xprint support was removed from libXfont.


I see.  I guess I should work out why; there have obviously been some 
developments that I have not been following carefully enough.  (This 
machine has been in "not broken don't fix it" mode for a long time, and 
now whenever I do need to change something small it seems to ripple 
out, breaking other stuff like this.)



We should add a Breaks:
xprint to the libxfont1 package, as it doesn't look like xprint will be
coming back.


Right.  For the recond, "apt-get install libXfont1/stable" seems to 
have cured the problem without any immediately-obvious side effects.



Thanks,  Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#535952: xprint: Xprt can't find symbol PrinterFontRegisterFpeFunctions and fails to start

[Phil Endecott]

Package: xprint
Version: 2:1.4.2-11
Severity: normal


Hi,

I've just installed xprint 2:1.4.2-11 and it fails to start:

# /etc/init.d/xprint start
/etc/init.d/xprint: Old server registry found, cleaning-up...
Stopping Xprint servers: Xprt.
Starting Xprint servers: Xprt.

No error message there, or in any log file that I can find, but there 
is no Xprt process running.  I try this:

# /usr/bin/Xprt :64
_XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6
_XSERVTransOpen: transport open failed for inet6/haiti:64
_XSERVTransMakeAllCOTSServerListeners: failed to open listener for inet6
No matching visual for __GLcontextMode with visual class = 0 (32775), nplanes = 
8
expected keysym, got XF86KbdLightOnOff: line 70 of pc
expected keysym, got XF86KbdBrightnessDown: line 71 of pc
expected keysym, got XF86KbdBrightnessUp: line 72 of pc
/usr/bin/Xprt: symbol lookup error: /usr/bin/Xprt: undefined symbol: 
PrinterFontRegisterFpeFunctions

I believe the last error is the important one.  As an aside, it would 
be really great if the init.d script could not hide this error in 
/dev/null, please; that wasted an hour.

Running nm -D on another machine suggests that perhaps the missing 
PrinterFontRegisterFpeFunctions should be defined in 
/usr/lib/libXfont.so.1:

# nm -D /usr/lib/libXfont.so.1 | grep PrinterFontRegisterFpeFunctions
d880 T PrinterFontRegisterFpeFunctions

That machine has libxfont1 version 1:1.3.3-1.
This machine where it doesn't work has libxfont1 version 1:1.4.0-1,
which seems to be the newest version, and the symbol is not defined:

# nm -D /usr/lib/libXfont.so.1 | grep PrinterFontRegisterFpeFunctions
#

So my guess is that the current libxfont1 is too new for the current 
xprint and/or the current xprint is too old for the current libxfont1.

Any thoughts?


Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.30-rc3
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages xprint depends on:
ii  libc6  2.9-8 GNU C Library: Shared libraries
ii  libdbus-1-31.2.1-1   simple interprocess messaging syst
ii  libfontenc11:1.0.1-6 X11 font encoding library
ii  libfreetype6   2.3.7-1   FreeType 2 font engine, shared lib
ii  libhal10.5.11-2  Hardware Abstraction Layer - share
ii  libpixman-1-0  0.14.0-1  pixel-manipulation library for X a
ii  libx11-6   2:1.2.1-1 X11 client-side library
ii  libxau61:1.0.1-2 X11 authorisation library
ii  libxdmcp6  1:1.0.1-2 X11 Display Manager Control Protoc
ii  libxfont1  1:1.4.0-1 X11 font rasterisation library
ii  xprint-common  2:1.4.2-11Xprint - the X11 print system (con
ii  zlib1g 1:1.2.3.3.dfsg-12 compression library - runtime

Versions of packages xprint recommends:
ii  xprint-utils 7.0.0+cvs20040414-4 utilities for Xprint, the X11 prin

-- no debconf information




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#510441: fcgi_frontend.py lacks execute permission

[Phil Endecott]

Hi,

I've also found that the trac.cgi file installed by the "deploy" 
command is not executable.  I think it should be, and trac (in CGI 
mode) didn't work until I made it executable, though I suppose this 
requirement could be something to do with my Apache config.



Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527973: initscripts needs a newish ifupdown else mountnfs fails

[Phil Endecott]

Hmmm, maybe that's not the (only) problem; NFS mounts now seem to come 
up most but not all of the time.  I need to investigate some more.



Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527973: initscripts needs a newish ifupdown else mountnfs fails

[Phil Endecott]

Package: initscripts
Version: 2.86.ds1-61
Severity: normal


I installed some other package that required a new initscripts.  Afterwards my 
NFS
filesystems stopped getting mounted at boot.

It turns out that /etc/network/if-up.d/mountnfs looks in 
/etc/network/run/ifstate, and
if the interface is not up yet [I have only lo and eth0] it stops.  With my old 
version
of ifupdown (0.6.7 I think) it seems that eth0=eth0 was not being appended to 
ifstate
until after montnfs had checked for it.  [Verified with strace -f].  Having 
upgraded to
the current version of ifupdown (0.6.8+nmu1) the problem has gone away.

The ifupdown changelog for version 0.6.8 mentions something about rewriting the 
way that
the state file is handled, and I guess that this is the key difference.

I suggest that initscripts should depend on ifupdown >= version 0.6.8.


Regards,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.30-rc3
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages initscripts depends on:
ii  debianutils   2.15.3 Miscellaneous utilities specific t
ii  e2fsprogs 1.38+1.39-WIP-2005.12.31-1 ext2 file system utilities and lib
ii  libc6 2.9-8  GNU C Library: Shared libraries
ii  lsb-base  3.2-22 Linux Standard Base 3.2 init scrip
ii  mount 2.13.1.1-1 Tools for mounting and manipulatin
ii  sysvinit-util 2.86.ds1-61System-V-like utilities

Versions of packages initscripts recommends:
ii  psmisc21.5-1 Utilities that use the proc filesy

-- no debconf information




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#402087: Would be nice to fix this...

[Phil Endecott]

I've just installed a new system and hit this confusing "no error 
message if $HOME doesn't exist" problem again.  It would be great if 
the code I proposed before, i.e.


# Warn if the user's home directory does not exist.
# It seems that $HOME is set to / after this script
# is executed, so we can't just compare $HOME with ~$USER.
if [ ! -d `eval echo ~$USER` ]
then
  xmessage "Home directory does not exist!"
fi

could be added to /etc/X11/xdm/Xstartup.  Thanks.


Phil.







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527958: mount doesn't try libblkid's SEC_TYPE when guessing fs type

[Phil Endecott]

Package: mount
Version: 2.13.1.1-1
Severity: normal


When mount has to guess a filesystem type it asks libblkid.  libblkid
may know both a TYPE and a SEC_TYPE.  mount tries the TYPE but ignores 
the SEC_TYPE.  I suggest that mount should try the SEC_TYPE if the TYPE 
fails for some reason.

Here's my scenario: I have built a kernel without vfat support (for 
political reasons; I only use fat for reading camera flash cards, and 
they only ever have short filenames).  When I try to mount an SD card, 
mount fails with "unknown filesystem type 'vfat'".  Yet the blkid 
command shows that libblkid has determined SEC_TYPE="msdos" for the 
device.  If mount tried this SEC_TYPE when the primary type failed it 
would succeed.

I note that libblkid takes precedence over the /proc/filesystems and 
/etc/filesystems lists, so even though my /proc/filesystems lists msdos 
and not vfat [once the msdos module is loaded] and irrespective of what 
I put in /etc/filesystems, I still get this error.

Regards,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.30-rc3
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages mount depends on:
ii  libblkid1 1.41.5-1   block device id library
ii  libc6 2.9-8  GNU C Library: Shared libraries
ii  libselinux1   2.0.59-1   SELinux shared libraries
ii  libuuid1  1.38+1.39-WIP-2005.12.31-1 universally unique id library

mount recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527955: console-setup: ckbcomp manpage typo

[Phil Endecott]

Package: console-setup
Version: 1.34
Severity: minor


"man ckbcomp" says:

  ckbcomp - compile a XKB keyboard description to a console font

That should surely say:

  ckbcomp - compile a XKB keyboard description to a console KEYMAP

While you're there, "a XKB" => "an XKB".

Cheers,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.30-rc3
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages console-setup depends on:
ii  console-terminus  4.28-1 Fixed-width fonts for fast reading
ii  debconf [debconf-2.0] 1.4.72 Debian configuration management sy
ii  xkb-data  1.5-2  X Keyboard Extension (XKB) configu

Versions of packages console-setup recommends:
ii  kbd   1.15-1 Linux console font and keytable ut

-- debconf information excluded




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#527619: xdm should depend on x11-xserver-utils

[Phil Endecott]

Package: xdm
Version: 1:1.1.8-5
Severity: normal


Currently xdm suggests xutils, which depends on x11-xserver-utils.
x11-xserver-utils includes the xrdb program.

If xrdb is not installed, xdm only just works.  Specifically, it will
not be able to read the contents of its configuration files in
/etc/X11/xdm.  So it will present its default appearance.  More
seriously it will not know to write a PID file in /var/run (this
is specified in /etc/X11/xdm/xdm-config).  Consequently,
/etc/init.d/xdm stop|restart will fail with a "not running" message
even though xdm is running.

I suggest that xdm should depend on x11-server-utils.


Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.30-rc3
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages xdm depends on:
ii  cpp   4:4.1.1-6  The GNU C preprocessor (cpp)
ii  debconf [debconf-2.0] 1.4.72 Debian configuration management sy
ii  libc6 2.9-8  GNU C Library: Shared libraries
ii  libice6   1:1.0.1-2  X11 Inter-Client Exchange library
ii  libpam0g  1.0.1-2Pluggable Authentication Modules l
ii  libselinux1   2.0.59-1   SELinux shared libraries
ii  libsm61:1.0.1-3  X11 Session Management library
ii  libx11-6  2:1.0.3-6  X11 client-side library
ii  libxau6   1:1.0.1-2  X11 authorisation library
ii  libxaw7   6.9.0.dfsg.1-4 X Athena widget set library
ii  libxdmcp6 1:1.0.1-2  X11 Display Manager Control Protoc
ii  libxext6  1:1.0.1-2  X11 miscellaneous extension librar
ii  libxft2   2.1.12-3   FreeType-based font drawing librar
ii  libxinerama1  2:1.0.3-2  X11 Xinerama extension library
ii  libxmu6   6.9.0.dfsg.1-6 X Window System miscellaneous util
ii  libxpm4   6.9.0.dfsg.1-4 X pixmap library
ii  libxrender1   1:0.9.4-2  X Rendering Extension client libra
ii  libxt61:1.0.2-2  X11 toolkit intrinsics library
ii  lsb-base  3.2-22 Linux Standard Base 3.2 init scrip
ii  x11-common1:7.4+1X Window System (X.Org) infrastruc
ii  x11-utils 7.4+1  X11 utilities

xdm recommends no packages.

-- debconf information:
  xdm/stop_running_server_with_children: false
  xdm/daemon_name: /usr/bin/xdm
* shared/default-x-display-manager: xdm




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#501503: udev: Griffin Powermate USB knob doesn't get /dev/input/by-id|path links

[Phil Endecott]

Kay Sievers wrote:

Does the input subsystem tell us anything specific we might want to
use to classify the device? What does:
  grep . /sys/class/input/inputX/capabilities/*
print?


$ grep . /sys/class/input/input10/capabilities/*
/sys/class/input/input10/capabilities/abs:0
/sys/class/input/input10/capabilities/ev:17
/sys/class/input/input10/capabilities/ff:0
/sys/class/input/input10/capabilities/key:1 0 0 0 0 0 0 0 0
/sys/class/input/input10/capabilities/led:0
/sys/class/input/input10/capabilities/msc:2
/sys/class/input/input10/capabilities/rel:80
/sys/class/input/input10/capabilities/snd:0
/sys/class/input/input10/capabilities/sw:0

The driver source is in drivers/input/misc/powermate.c.


Cheers,  Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#501503: udev: Griffin Powermate USB knob doesn't get /dev/input/by-id|path links

[Phil Endecott]

Kay Sievers wrote:

On Mon, Apr 27, 2009 at 00:31, Phil Endecott
 wrote:

So protocol==0 is not specific to this unusual device; even a keyboard or
mouse will have 0 for its non-boot interfaces.  So the current logic will
only create /dev/by-id|path nodes for boot interfaces.  Is that intentional?


These links got added to distinguish multiple keyboards on
mainframe-like servers, that's why the rules have been so strict so
far, and cover only "real" input devices. :)


Hmm, OK; so how am I supposed to identify things that may be connected 
to different sockets and enumerated in different orders each time I use them?



We see several mouse-like devices, which are not covered by that, and
we might need to find a scheme to include them in the link creation.
We need to be careful here though, because there are tons of HID
devices with "00" which have nothing to do with input devices in the
usual sense. I have a thermometer which is exactly like that. :)


Right.  I have a Wi-Spy which I think is a HID device.  Does it do any 
harm to add links for things like that?



What kind of kernel devices does this device create, I mean, if you
run "udevadm monitor --kernel" and plug it in, what does it say?


(I should say first that I re-discovered this bug because the computer 
that I was using the PowerMate with has died; I have been trying to 
remember what I had to do to make it work so that I can use it with the 
replacement.  I don't have it working properly yet, I think because 
something in X has changed.  The following info is from the new 
not-yet-working machine.)


# udevadm monitor --kernel
udevmonitor will print the received events for:
UEVENT the kernel uevent

UEVENT[1240788686.318926] add  
/devices/pci:00/:00:1d.7/usb1/1-4/1-4.1 (usb)
UEVENT[1240788686.322049] add  
/devices/pci:00/:00:1d.7/usb1/1-4/1-4.1/1-4.1:1.0 (usb)

UEVENT[1240788686.325014] add  /class/input/input15 (input)
UEVENT[1240788686.325203] add  /class/input/input15/event12 (input)
UEVENT[1240788686.325301] add  /class/usb_endpoint/usbdev1.9_ep81 
(usb_endpoint)
UEVENT[1240788686.325394] add  /class/usb_endpoint/usbdev1.9_ep02 
(usb_endpoint)
UEVENT[1240788686.325549] add  /class/usb_device/usbdev1.9 (usb_device)
UEVENT[1240788686.325651] add  /class/usb_endpoint/usbdev1.9_ep00 
(usb_endpoint)


Care to show the entire sequence?


Not sure what you mean.


Is it "inputX" only for the input devices, or do you get also a "mouseX" device?


I get only an eventX device:

# tree /dev/input
/dev/input
|-- by-id
|   |-- usb-Apple_Inc._Apple_Keyboard-event-kbd -> ../event10
|   |-- usb-Griffin_Technology__Inc._Griffin_PowerMate-event-misc -> ../event12
|   |-- usb-Kingsis_Peripherals_Evoluent_VerticalMouse_2-event-mouse -> 
../event9
|   `-- usb-Kingsis_Peripherals_Evoluent_VerticalMouse_2-mouse -> ../mouse1
|-- by-path
|   |-- pci-:00:1d.1-usb-0:1:1.0-event-mouse -> ../event9
|   |-- pci-:00:1d.1-usb-0:1:1.0-mouse -> ../mouse1
|   |-- pci-:00:1d.7-usb-0:4.1:1.0-event-misc -> ../event12
|   |-- pci-:00:1d.7-usb-0:4.2:1.0-event-kbd -> ../event10
|   |-- pci-:00:1d.7-usb-0:4.2:1.1-event- -> ../event11
|   |-- platform-i8042-serio-0-event-kbd -> ../event7
|   |-- platform-i8042-serio-1-event-mouse -> ../event8
|   |-- platform-i8042-serio-1-mouse -> ../mouse0
|   `-- platform-pcspkr-event-spkr -> ../event5
|-- event0
|-- event1
|-- event10
|-- event11
|-- event12
|-- event2
|-- event3
|-- event4
|-- event5
|-- event6
|-- event7
|-- event8
|-- event9
|-- mice
|-- mouse0
`-- mouse1


Also, care to add the full section of "lsusb -v" for this device?


Bus 001 Device 009: ID 077d:0410 Griffin Technology PowerMate
Device Descriptor:
  bLength18
  bDescriptorType 1
  bcdUSB   1.00
  bDeviceClass0 (Defined at Interface level)
  bDeviceSubClass 0
  bDeviceProtocol 0
  bMaxPacketSize0 8
  idVendor   0x077d Griffin Technology
  idProduct  0x0410 PowerMate
  bcdDevice4.00
  iManufacturer   1 Griffin Technology, Inc.
  iProduct2 Griffin PowerMate
  iSerial 0
  bNumConfigurations  1
  Configuration Descriptor:
bLength 9
bDescriptorType 2
wTotalLength   41
bNumInterfaces  1
bConfigurationValue 1
iConfiguration  4 Media Controller
bmAttributes 0xa0
  (Bus Powered)
  Remote Wakeup
MaxPower  100mA
Interface Descriptor:
  bLength 9
  bDescriptorType 4
  bInterfaceNumber0
  bAlternateSetting   0
  bNumEndpoints   2
  bInterfaceClass 3 Human Interface Device
  bInterfaceSubClass  0 No Subclass
  bInterfaceProtocol  0 None

Bug#501503: udev: Griffin Powermate USB knob doesn't get /dev/input/by-id|path links

[Phil Endecott]

Hi Kay,

Sorry for the delay; the email address that I used originally for this 
bug is now spam-filtered.  I seem to get mountains of spam to Debian 
bug addresses, for some reason.  Any random string @ chezphil.org will 
reach me.



SUBSYSTEMS=="usb", \
   ATTRS{bInterfaceClass}=="03", ATTRS{bInterfaceProtocol}=="00", \
   ENV{ID_CLASS}="misc"
I now get the desired symlinks:

/dev/input/by-id/usb-Griffin_Technology,_Inc._Griffin_PowerMate-event-misc
/dev/input/by-path/pci-:00:10.2-usb-0:1:1.0-event-misc


It's undefined if bInterfaceProtocol is 0.

What is the subclass value of the device? Like "lsusb -v" prints for a mouse:
 bNumEndpoints   1
 bInterfaceClass 3 Human Interface Device
 bInterfaceSubClass  1 Boot Interface Subclass
 bInterfaceProtocol  2 Mouse



I get:

  bInterfaceClass 3 Human Interface Device
  bInterfaceSubClass  0 No Subclass
  bInterfaceProtocol  0 None

By chance I have been learning about USB HID recently and I have a copy 
of the spec in front of me.  Apparently (sec. 4.3):


"the bInterfaceProtocol member of an Interface descriptor only has 
meaning if the bInterfaceSubClass member declares that the device 
supports a boot interface, otherwise it is 0."


So protocol==0 is not specific to this unusual device; even a keyboard 
or mouse will have 0 for its non-boot interfaces.  So the current logic 
will only create /dev/by-id|path nodes for boot interfaces.  Is that intentional?



Cheers,  Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#187021: Preventing less from believing a compressed file is empty.

[Phil Endecott]

Hi,

Can you confirm whether Steve's patch in bug 187021 message #21 was 
applied unchanged, or has it somehow been modified to eliminate the side-effects?



Regards,  Phil.











--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#510491: pngcrush sets timestamps on output file

[Phil Endecott]

I've just tried the new version; it looks like the --help output is 
corrupt for the new option:


 -ster [0 or 1]
 -text b[efore_IDAT]|a[fter_IDAT] "keyword" "text"
 -time_stamp
   -trns_array n trns[0] trns[1] .. trns[n-1]
 -trns index red green blue gray

I would still much prefer it this program behaved like every other 
program and didn't mess with the timestamps, by default.  And there's 
still no rationale for its non-standard behaviour.



Regards,  Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#514904: proj: Examples in man page and readme broken, in one case due to curly-quote

[Phil Endecott]

I have installed proj from the upstream source and I can confirm that 
the examples all do what they say they should do.  I'm pretty sure the 
problem is that, contrary to the note in README.Debian, you have not 
unpacked the proj-datumgrid-*.zip file in the source tree before building.


Phil.






--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#514904: /usr/share/proj/conus missing

[Phil Endecott]

Some more info:

According to /usr/share/doc/proj/README.Debian, "The Debian proj 
package includes built-in datum conversion.  (It was compiled with 
proj-datumgrid-1.3.zip unpacked in the nad source directory.)"


According to the proj.4 FAQ (http://trac.osgeo.org/proj/wiki/FAQ): 
"Verify that you have the binary files (eg. 
/usr/local/share/proj/conus) installed on your system."


It seems that the package does not include /usr/local/share/proj/conus, 
according to "dpkg -L proj", though it does have some other files there.


Also from the FAQ, there's a PROJ_DEBUG environment variable.  Setting 
this I get:


$ cs2cs +proj=latlong +datum=NAD27 +to +proj=latlong +datum=NAD83
pj_open_lib(proj_def.dat): call fopen(/usr/share/proj/proj_def.dat) - succeeded
pj_open_lib(proj_def.dat): call fopen(/usr/share/proj/proj_def.dat) - succeeded
-117 30
pj_open_lib(conus): call fopen(/usr/share/proj/conus) - failed
pj_open_lib(alaska): call fopen(/usr/share/proj/alaska) - failed
pj_open_lib(ntv2_0.gsb): call fopen(/usr/share/proj/ntv2_0.gsb) - failed
pj_open_lib(ntv1_can.dat): call fopen(/usr/share/proj/ntv1_can.dat) - failed
117dW   30dN 0.000

So it does seem that it's the absence of this /usr/share/proj/conus 
file that's the problem.  Has it been accidentally omitted from the 
build, perhaps?


I also consider the fact that the program silently continues and prints 
a wrong answer when it has detected that the file is missing to be a bug.



Phil.







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#495349: Dependency & symlink problems

[Phil Endecott]

I've also just encountered this bug.  Having installed geotiff-bin, 
when I try to run listgeo I get the error "listgeo: error while loading 
shared libraries: libgeotiff.so: cannot open shared object file: No 
such file or directory".  If I understand correctly - and it's quite 
possible that I don't - there are two problems.  Firstly, geotiff-bin 
doesn't depend on libgeotiff1.2, so there isn't much chance of it 
finding a libgeotiff.so unless you've previously installed that 
package.  Secondly, if you manually install libgeotiff1.2 you still 
don't actually get a /usr/lib/libgeotiff.so.  I would have thought that 
making a symlink from /usr/lib/libgeotiff.so to libgeotiff.so.1 would 
be appropriate, and adding that link manually makes listgeo work.  Or 
maybe listgeo should be compiled to expect the .so.1?


Cheers,  Phil.







--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#510491: pngcrush sets timestamps on output file

[Phil Endecott]

Package: pngcrush
Version: 1.6.7-1
Severity: normal


Hi,

It seems that pngcrush sets the timestamps on the file that it creates to
be the same as the input file.  I've confirmed this by looking at the output
of strace.  There is no rationale for this behaviour in the documentation
nor any way to disable it, as far as I can see.

This caught me because neither Subversion nor my backup system noticed that
the files had changed.   I believe that the behaviour should be made 
consistent with just about every other command-line program, i.e. the 
operating system should be allowed to set appropriate timestamps.

Regards,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages pngcrush depends on:
ii  libc6  2.7-10GNU C Library: Shared libraries
ii  libpng12-0 1.2.15~beta5-1PNG library - runtime
ii  zlib1g 1:1.2.3.3.dfsg-12 compression library - runtime

pngcrush recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#502122: closed by [EMAIL PROTECTED] (Steinar H. Gunderson) (Bug#502122: fixed in nfs-utils 1:1.1.4-1)

[Phil Endecott]

[EMAIL PROTECTED] wrote:

Source: nfs-utils
Source-Version: 1:1.1.4-1

We believe that the bug you reported is fixed in the latest version of
nfs-utils, which is due to be installed in the Debian FTP archive:


Be careful, there was some debate about whether the fix was actually correct:

http://thread.gmane.org/gmane.linux.nfs/22843

Quoting J. Bruce Fields' last messge in that thread:

  "So this patch *does* break stuff.
  Oops!

  So should we revert it and do something else, or patch statd to create
  a new state file if necessary?"

This doesn't seem to have been discussed further yet, AFAICT.


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#501847: kexec-tools: Please document the changed reboot behaviour.

[Phil Endecott]

Hi Khalid,

My request was that the behaviour be documented somewhere.

For example, "man kexec" tells me to kexec -l and then kexec -e.  But 
actually I should, depending on a debconf setting, do something like 
kexec -l; reboot.  The kexec man page might not be the right place to 
say that, but it's the only documentation you ship.


Khalid Aziz wrote:

Please see 
that led to this change in behavior. kexec-tools 20080227 added a
debconf option to ask the user if they want kexec to handle reboots or
not.


Not documented either, as far as I can see.


Current version in leeny has this debconf option. This should
address your concern. If kexec reboots are enabled, a kexec reboot will
load the kernel /vmlinuz. Unless a user has built custom kernels,


A significant group of users for fast reboot are those who are in the 
edit-run-debug cycle with custom kernels.



the
last kernel installed by a debian package would update /vmlinuz link to
point to the kernel installed by it which is typically the latest kernel
and which is what most typical users want to boot. This default can be
changed in /etc/default/kexec.


It would be great to get a menu.  In practice, I find that the time 
saved by kexec does not make up for the convenience of the grub menu.  
I'll see if I can come up with something.



Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#327584: hwclock.sh & hwclockfirst.sh clarifications

[Phil Endecott]

Hi LaMont, thanks for your reply.

LaMont Jones wrote:

On Mon, Oct 13, 2008 at 06:18:56PM +0100, Phil Endecott wrote:
I'm also curious to know why reading the clock takes a second anyway.   
Note that we have to use --directisa on this hardware to avoid a random  
lock-up.  It looks as if the CPU is busy for the second that it's  
running - in a busy-loop maybe?  Is there anything that can be done to  
accelerate this?


With directisa, it waits for the realtime clock to transition.  That is
going to take up to 1 second.


OK.  We should really try to get the normal access fixed, if that 
doesn't need to spin in that way.



As for why two copies, it's at least partially a result of rtc being a
module instead of a built-in.  See also bug #443487.


Hmm.  Well, since I'm not using adjtime and since I have no modules, it 
would be good to avoid the second invocation somehow.  Is there a tidy 
way to do this?



Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#502122: nfs-utils: Patch to make sm-notify faster when there's no-one to notify

[Phil Endecott]

Package: nfs-utils
Version: 1.1.3-1
Severity: normal
Tags: patch


Hi,

I use NFS only rarely on my laptop.  During boot, sm-notify takes several
seconds to run - whether or not I'm using NFS.  On investigation, this
is because sm-notify indirectly calls sync():

/* Get and update the NSM state. This will call sync() */
nsm_state = nsm_get_state(opt_update_state);

which causes lots of pending writes from the boot process to be written
out, and slows other boot processes that are trying to read.  It takes 2-3
seconds on my system.  And I believe it's unnecessary in the (common for me)
case where I didn't use NFS in the last session and sm-notify actually has
nothing to do.

I propose therefore that sm-notify should first check whether its list of
hosts is empty, and terminate immediately in this case.  The following simple
patch implements that:

diff -ur nfs-utils-1.1.3.orig/utils/statd/sm-notify.c 
nfs-utils-1.1.3/utils/statd/sm-notify.c
--- nfs-utils-1.1.3.orig/utils/statd/sm-notify.c2008-07-27 
22:01:45.0 +0100
+++ nfs-utils-1.1.3/utils/statd/sm-notify.c 2008-10-13 19:02:54.0 
+0100
@@ -169,6 +169,10 @@
backup_hosts(_SM_DIR_PATH, _SM_BAK_PATH);
get_hosts(_SM_BAK_PATH);
 
+   if (!hosts) {
+   return 0;
+   }
+
/* Get and update the NSM state. This will call sync() */
nsm_state = nsm_get_state(opt_update_state);
set_kernel_nsm_state(nsm_state);


This accelerates boot by the expected couple of seconds on my machine, but
I have not tested whether it breaks something when there are hosts to notify.
An "expert" should look at this before you consider applying it.

(BTW, I had a look at the sourceforge mailing list for NFS-utils and it seems
to be 100% spam.  Is there a better upstream bug reporting channel?)


Regards,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#327584: hwclock.sh & hwclockfirst.sh clarifications

[Phil Endecott]

the documentation should explain more or less verbosely, why two
scripts doing virtually the same in short distance are needed.


Yes please!  I'm currently trying to improve the boot time on my ASUS 
Eee; you may have heard of Arjan van de Ven's demo where he gets this 
machine to boot in only 5 seconds - Debian is currently about 5 times 
worse than that.  And it spends two whole seconds (gasp!) setting the 
clock.  If it would set it only once, that would save a whole valuable 
second!  So I'm tempted to hack out one or other of these scripts - but 
I fear doing so, since I don't know why we have two in the first 
place.  Could we please have a few words of explanation, dear experts?


I'm also curious to know why reading the clock takes a second anyway.  
Note that we have to use --directisa on this hardware to avoid a random 
lock-up.  It looks as if the CPU is busy for the second that it's 
running - in a busy-loop maybe?  Is there anything that can be done to 
accelerate this?


Cheers,  Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#501847: kexec-tools: Please document the changed reboot behaviour.

[Phil Endecott]

Package: kexec-tools
Version: 20080324-2
Severity: normal


I've just installed kexec-tools and was surprised to find that, rather than just
packaging the kexec executable, you have installed a script so that kexec is 
used
when I reboot.  I didn't want that.  This feature - and how to disable it - are
not described in the documentation as far as I can see.  Can you please add some
documentation describing this?  I would also vote for disabling it by default.

It also seems that the kernel which you run will not in general be the one that
was running before the reboot.  Is that what users would have expected?

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#501503: udev: Griffin Powermate USB knob doesn't get /dev/input/by-id|path links

[Phil Endecott]

Package: udev
Version: 0.103-1
Severity: normal


Hi,

I have a Griffin Powermate, which is a USB "knob" that can be used as 
e.g. a volume control, for scrolling through video/audio, etc.  No
/dev/input/by-* links are created for it.  This is because
persistent-input.rules only sets ENV{ID_CLASS} for keyboards and mice
when it recognises them from their ATTRS{bInterfaceProtocol}.  This
device has ATTRS{bInterfaceProtocol}==0.  I have added this, to give
it class "misc"; I'm not sure if there is an official USB-spec-defined
meaning for protocol==0 that I should be using:

SUBSYSTEMS=="usb", \
ATTRS{bInterfaceClass}=="03", ATTRS{bInterfaceProtocol}=="00", \
ENV{ID_CLASS}="misc"
I now get the desired symlinks:

/dev/input/by-id/usb-Griffin_Technology,_Inc._Griffin_PowerMate-event-misc
/dev/input/by-path/pci-:00:10.2-usb-0:1:1.0-event-misc

Cheers,  Phil.

-- Package-specific info:
-- /etc/udev/rules.d/:
/etc/udev/rules.d/:
total 16
lrwxrwxrwx 1 root root   20 2006-04-06 20:13 020_permissions.rules -> 
../permissions.rules
lrwxrwxrwx 1 root root   19 2006-09-14 23:31 025_libgphoto2.rules -> 
../libgphoto2.rules
lrwxrwxrwx 1 root root   22 2006-09-30 23:28 025_logitechmouse.rules -> 
../logitechmouse.rules
-rw-r--r-- 1 phil phil  949 2007-08-10 22:24 chezphil.rules
lrwxrwxrwx 1 root root   13 2006-04-06 20:13 udev.rules -> ../udev.rules
lrwxrwxrwx 1 root root   25 2006-04-06 20:13 z20_persistent-input.rules -> 
../persistent-input.rules
lrwxrwxrwx 1 root root   19 2006-04-06 20:13 z20_persistent.rules -> 
../persistent.rules
-rw-r--r-- 1 root root 5953 2007-12-14 16:44 z25_persistent-cd.rules
-rw-r--r-- 1 root root  573 2007-08-03 00:27 z25_persistent-net.rules
lrwxrwxrwx 1 root root   33 2006-12-21 20:26 z45_persistent-net-generator.rules 
-> ../persistent-net-generator.rules
lrwxrwxrwx 1 root root   12 2006-04-06 20:13 z50_run.rules -> ../run.rules
lrwxrwxrwx 1 root root   16 2006-04-06 20:13 z55_hotplug.rules -> 
../hotplug.rules
lrwxrwxrwx 1 root root   19 2006-04-18 22:43 z60_alsa-utils.rules -> 
../alsa-utils.rules
lrwxrwxrwx 1 root root   15 2006-04-06 23:41 z60_hdparm.rules -> ../hdparm.rules
lrwxrwxrwx 1 root root   33 2006-05-21 12:25 z60_xserver-xorg-input-wacom.rules 
-> ../xserver-xorg-input-wacom.rules
lrwxrwxrwx 1 root root   29 2006-04-06 20:13 z75_cd-aliases-generator.rules -> 
../cd-aliases-generator.rules

-- /sys/:
/sys/block/hda/dev
/sys/block/hda/hda1/dev
/sys/class/drm/card0/dev
/sys/class/input/input0/event0/dev
/sys/class/input/input0/mouse0/dev
/sys/class/input/input1/event1/dev
/sys/class/input/input3/event2/dev
/sys/class/input/mice/dev
/sys/class/misc/agpgart/dev
/sys/class/misc/cpu_dma_latency/dev
/sys/class/misc/network_latency/dev
/sys/class/misc/network_throughput/dev
/sys/class/misc/psaux/dev
/sys/class/misc/rtc/dev
/sys/class/misc/snapshot/dev
/sys/class/sound/controlC0/dev
/sys/class/sound/pcmC0D0c/dev
/sys/class/sound/pcmC0D0p/dev
/sys/class/sound/pcmC0D1c/dev
/sys/class/sound/pcmC0D1p/dev
/sys/class/sound/timer/dev
/sys/class/usb_device/usbdev1.1/dev
/sys/class/usb_device/usbdev1.4/dev
/sys/class/usb_device/usbdev2.1/dev
/sys/class/usb_device/usbdev3.1/dev
/sys/class/usb_device/usbdev3.2/dev
/sys/class/usb_device/usbdev3.4/dev
/sys/class/usb_device/usbdev4.1/dev
/sys/class/usb_device/usbdev4.3/dev
/sys/class/usb_device/usbdev4.5/dev
/sys/class/usb_device/usbdev4.6/dev
/sys/class/usb_endpoint/usbdev1.1_ep00/dev
/sys/class/usb_endpoint/usbdev1.1_ep81/dev
/sys/class/usb_endpoint/usbdev1.4_ep00/dev
/sys/class/usb_endpoint/usbdev1.4_ep81/dev
/sys/class/usb_endpoint/usbdev2.1_ep00/dev
/sys/class/usb_endpoint/usbdev2.1_ep81/dev
/sys/class/usb_endpoint/usbdev3.1_ep00/dev
/sys/class/usb_endpoint/usbdev3.1_ep81/dev
/sys/class/usb_endpoint/usbdev3.2_ep00/dev
/sys/class/usb_endpoint/usbdev3.2_ep02/dev
/sys/class/usb_endpoint/usbdev3.2_ep03/dev
/sys/class/usb_endpoint/usbdev3.2_ep81/dev
/sys/class/usb_endpoint/usbdev3.2_ep82/dev
/sys/class/usb_endpoint/usbdev3.2_ep83/dev
/sys/class/usb_endpoint/usbdev3.4_ep00/dev
/sys/class/usb_endpoint/usbdev3.4_ep02/dev
/sys/class/usb_endpoint/usbdev3.4_ep81/dev
/sys/class/usb_endpoint/usbdev4.1_ep00/dev
/sys/class/usb_endpoint/usbdev4.1_ep81/dev
/sys/class/usb_endpoint/usbdev4.3_ep00/dev
/sys/class/usb_endpoint/usbdev4.3_ep81/dev
/sys/class/usb_endpoint/usbdev4.5_ep00/dev
/sys/class/usb_endpoint/usbdev4.5_ep03/dev
/sys/class/usb_endpoint/usbdev4.5_ep81/dev
/sys/class/usb_endpoint/usbdev4.5_ep82/dev
/sys/class/usb_endpoint/usbdev4.6_ep00/dev
/sys/class/usb_endpoint/usbdev4.6_ep02/dev
/sys/class/usb_endpoint/usbdev4.6_ep81/dev
/sys/class/usbmon/usbmon0/dev
/sys/class/usbmon/usbmon1/dev
/sys/class/usbmon/usbmon2/dev
/sys/class/usbmon/usbmon3/dev
/sys/class/usbmon/usbmon4/dev
/sys/devices/pci:00/:00:10.0/usb1/1-1/dev
/sys/devices/pci:00/:00:10.0/usb1/dev
/sys/devices/pci:00/:00:10.1/usb2/dev
/sys/devices/pci:00/:00:10.2/usb3/3-1/dev
/sys/devices/pci:00/:00:10.2/u

Bug#499459: glabels: New card type request

[Phil Endecott]

Package: glabels
Version: 2.1.3-3
Severity: wishlist


Hi,

I have some business cards that describe themselves as 
"DECAdry OCC-3342" which glabels doesn't seem to know about.  
However, as far as I can see they have the same dimensions 
as type "OCB-3327", so you could duplicate that menu entry 
for them.

(BTW I tried to submit this to the glabels bug tracker at
sourceforge, but it doesn't like me for some reason.)

Thanks,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497220: acpi-support-base: Needs to depend on finger

[Phil Endecott]

Bart Samwel wrote:


Nope. I use gdm, and I get:

$ w -hs
root tty1 - 2:19  -bash
root tty2 - 2:19  -bash
bsamwel  tty7 :00.00s x-session-manager
root pts/1:0.0  2:09m gnome-terminal
bsamwel  pts/2:0.0  0.00s w -hs

The first two lines are from virtual terminals, the third one is for
tty7 which is what my X is running on, and the final two ones are for
terminals.


OK; so xdm and gdm (gdm==x-session-manager???) are inconsistent about 
what they put in the wtmp records (or whatever they are).  That's 
unfortunate, but I bet we're not the first to discover it.  I wouldn't 
be surprised if KDE did something different again


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497220: acpi-support-base: Needs to depend on finger

[Phil Endecott]

Julien Cristau wrote:

On Thu, Sep  4, 2008 at 15:04:14 +0100, Phil Endecott wrote:

No this doesn't work for me.  You're looking for :0 in the FROM column,  
right?  I have it in the TTY column:


$ w -hs
phil tty1 -17:19  -bash
root tty2 -15:31  -bash
phil pts/0egypt.chezphil.o  0.00s sshd: phil [priv]
phil pts/2egypt.chezphil.o  1:54  nano libskyline/src/compute_skyline.cc
phil :0   -?xdm?  -:0

Very peculiar.  I'm not really sure what to suggest; I think that  
understanding what w, who, finger etc. are really trying to tell us WRT  
X sessions would be a good start.  I doubt there's anything useful in  
the man pages



Your wtmp entry comes from xdm, Bart's probably comes from a terminal
emulator.  I have this:
USER TTY  FROM  LOGIN@   IDLE   JCPU   PCPU WHAT
julien   :0   -02:54   ?xdm?  14:13m  0.04s -:0 
julien   pts/0:0.0 02:54   25:03m  0.58s  0.58s bash

(first xdm, then xterm)

I'd say looking at the tty is the right thing to do.


Ah yes; I did wonder about that.  For some reason I'm not seeing lines 
in w, who or finger output for terminal emulators running inside my X 
session, though I have seen them in the past.  If you did have them 
they could in principle be for different users than the user who owns 
the X session itself.  The TTY=:0 line is the right one to use.  But 
presumably Bart is not seeing a line like that, right?



Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497220: acpi-support-base: Needs to depend on finger

[Phil Endecott]

Bart Samwel wrote:

getXuser() {
w -hs | while read -r THIS_USER THIS_TTY THIS_DISPLAY DUMMY_REMAINDER; 
do
if [ "$THIS_DISPLAY" = "$displaynum" ] ; then
user=$THIS_USER
break
fi
done
if [ x"$user" = x"" ]; then
startx=`pgrep -n startx`
if [ x"$startx" != x"" ]; then
user=`ps -o user --no-headers $startx`
fi
fi
if [ x"$user" != x"" ]; then
userhome=`getent passwd $user | cut -d: -f6`
export XAUTHORITY=$userhome/.Xauthority
else
export XAUTHORITY=""
fi
export XUSER=$user
}


No this doesn't work for me.  You're looking for :0 in the FROM column, 
right?  I have it in the TTY column:


$ w -hs
phil tty1 -17:19  -bash
root tty2 -15:31  -bash
phil pts/0egypt.chezphil.o  0.00s sshd: phil [priv]
phil pts/2egypt.chezphil.o  1:54  nano libskyline/src/compute_skyline.cc
phil :0   -?xdm?  -:0

Very peculiar.  I'm not really sure what to suggest; I think that 
understanding what w, who, finger etc. are really trying to tell us WRT 
X sessions would be a good start.  I doubt there's anything useful in 
the man pages



Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497220: acpi-support-base: Needs to depend on finger

[Phil Endecott]

Bart Samwel wrote:

Hi Phil,

Phil Endecott wrote:

I am trying to get the lid event to suspend on my Eee 901 and have encountered
the following problems:

1. getXuser() in /usr/share/acpi-support/power-funcs uses finger, but the package 
does not depend on finger (and I didn't have it installed).


Thanks for reporting, added!


I've just spotted detect_x_display() in 
/usr/share/eeepc-acpi-scripts/functions.sh from package 
eeepc-acpi-scripts which does a similar thing by parsing the output of 
"who", rather than "finger".  "who" has the advantage of being provided 
by coreutils, which is a "Priority: required" package, while finger is 
"Priority: standard".  There is also "w" from procps.



the format provided by finger is (:$displaynum) or (:$displaynum.screennum).


Err, no; mine doesn't have the ():

$ finger
Login Name    Tty  Idle  Login Time   Office Office Phone
phil  Phil Endecott  *tty114:51  Sep  2 16:40
phil  Phil Endecott   pts/0  Sep  4 12:09 (egypt.chezphil.org)
phil  Phil Endecott  *:0 Sep  4 12:29
root  root   *tty213:02  Sep  3 21:40

Obviously yours does and I'm sure I've seem that notation somewhere or 
other; I don't know if the () mean something or whether it's a finger 
version thing, or what.



Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#496690: xdm: want to use Tab to switch from Login to Password field

[Phil Endecott]

I have also configured Tab to have the finish-field behaviour.  A 
mis-feature of this configuration, however, is that pressing tab while 
in the Password field will not swap back to the username field, but 
will instead try to logon.  So if the goal is consistency with other 
logon screens or other GUI apps, it is only partially successful.  I'm 
not aware of any way to fix this.


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497220: acpi-support-base: Needs to depend on finger

[Phil Endecott]

Package: acpi-support-base
Version: 0.109-5
Severity: normal


I am trying to get the lid event to suspend on my Eee 901 and have encountered
the following problems:

1. getXuser() in /usr/share/acpi-support/power-funcs uses finger, but the 
package 
does not depend on finger (and I didn't have it installed).

2. getXuser() is called from CheckPolicy() in 
/usr/share/acpi-support/policy-funcs and $displaynum is not set.  Do you expect 
that $displaynum is set by the caller of CheckPolicy(), i.e. in this case 
eeepc-acpi-scripts' /etc/acpi/actions/lid.sh?  Or should it be set to 0 in 
CheckPolicy()?

(Most users won't notice this second problem as the grep in getXuser will look 
for ':' when displaynum is not set and match either the idle time or the login 
time.  In fact, even with a valid display number, the pattern ':0' will 
frequently match the idle or login time in the second finger|grep|awk.  You 
need 
to take care to match only in the Tty column of the finger output.)

Cheers,  Phil 

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#497200: [Debian-eeepc-devel] Bug#497200: ITP: rt2860-source -- source for RT2860 wireless adapter kernel module

[Phil Endecott]

Damyan Ivanov wrote:

There may be some licensing problems and this is why I CC debian-legal.
All the sources are licensed under GPL-2+, except one file,
include/firmware.h, which is generated from a binary blob and contains
the following notice:


[snip]

Looks bad to me.


I did not yet check if this code is actually linked in the GPL-2+
module, but have a bad feeling it it does. Would a compiled GPL source,
including firmware.h be even distributable?

Perhaps the module can be changed to load its firmware from external
file or even not need that nasty firmware.h (there are traces of
support to other hardware and that firmware may be for them).


There's some code in common/rtmp_init.c that's inside #ifdef 
BIN_IN_FILE that tries to get the firmware from a file, and falls back 
to the contents of firmware.h.  BIN_IN_FILE is not defined, and so it 
unconditionally uses firmware.h.  I have had a quick try at enabling 
BIN_IN_FILE, and defining RTMP_FIRMWARE_FILE_NAME to point to 
common/rt2860.bin, and according to the debug output it worked.  I 
can't be certain though, as it still has firmware.h compiled in as a 
fallback and unless you power-cycle you can't be sure that the previous 
download has been lost.


Someone keen could probably convert it to use the proper kernel 
firmware loading infrastructure.  The code that does this seems to be 
quite localised in common/rtmp_init.c.



Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#426294: anacron: fails to start if on battery

[Phil Endecott]

anacron deliberately doesn't run if the machine is running on battery 
power.  This is fine if the only jobs are "housekeeping" tasks that 
it's safe to skip, but if you have critical jobs like backups then it's 
not so good.  This is a dangerous situation as users are unlikely to 
notice that this feature exists until it's too late, e.g. when they 
want to recover something from a backup.


I suggest:

1. This feature needs to be documented, at the very least in 
README.Debian.  Actually looking at bug #83334 message #40 a patch was 
submitted that added "anacron will avoid running while the system is 
battery powered" to that file, but that either never made it in to the 
package or it has subsequently been removed.  But I don't think 
documentation alone is sufficient, as users familiar with anacron from 
non-laptops will be able to set up their jobs without referring to any 
documentation and won't learn about this feature (until it's too late).


2. There needs to be some way to enable or disable this feature, and 
I'd suggest that it should be disabled by default.  A debconf question 
and/or /etc/default/anacron variable would be a possible method.


3. Better than a global enable/disable setting would be to indicate 
that some jobs are deferrable, and should only be run on ac power, and 
others are critical and should be run unconditionally.  This could be 
done using different directories or a naming convention.


A related problem is that some jobs may require network connectivity 
while others don't.  I can imagine a scheme where job names include 
tags, e.g.


/etc/cron.daily/man_db--acpower  - only run when on ac power, doesn't 
need network
/etc/cron.daily/backup--net  - run even when on battery, only if 
network is up


You can specify a job pattern on the anacron command line to specify 
the tags to match in the current run.



Regards,  Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#495932: gpsbabel: man page formatting snag

[Phil Endecott]

Package: gpsbabel
Version: 1.3.5-1.1
Severity: normal


/usr/share/man/man1/gpsbabel.1 contains this, near the start:

br
INTYPE and OUTTYPE must be one of the file types listed below, and
br
may include options valid for that file type.  For example:
br
'gpx', 'gpx,snlen=10' and 'ozi,snlen=10,snwhite=1'
br
(without the quotes) are all valid file type specifications.
br

This comes out as:

   INTYPE and OUTTYPE must be one of the file types listed below, and
   may include options valid for that file type.  For example:
   (without the quotes) are all valid file type specifications.

Note that the line containing the examples has been lost.

I know next to nothing about roff (if that's what it is), but my guess is
that having a ' as the first character of the line has caused the problem.


Cheers,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages gpsbabel depends on:
ii  libc6  2.7-10GNU C Library: Shared libraries
ii  libexpat1  1.95.8-3.4XML parsing C library - runtime li
ii  libusb-0.1-4   2:0.1.12-2userspace USB programming library
ii  zlib1g 1:1.2.3.3.dfsg-12 compression library - runtime

gpsbabel recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#495842: Starting gpsd on hotplug seems broken

[Phil Endecott]

Package: gpsd
Version: 2.37-6
Severity: normal


Please forgive me if I have misunderstood how this is supposed 
to work.

I want gpsd to start when I plug in the USB device.

debconf asks me this question:

  Start gpsd automatically on boot?

This isn't what I want so I say no.  It asks no other questions 
and saves this in /etc/default/gpsd:

  START_DAEMON="false"
  USBAUTO="false"

As expected, /etc/init.d/gpsd looks at this and doesn't start 
gpsd at boot.

When I hotplug the device, /lib/udev/gpsd.hotplug.wrapper is 
called and does this:

  [ "$START_DAEMON" = "true" ] || exit 0

Since START_DAEMON is false, it exits.


So I think there are two problems here.  Firstly, if I say "no" 
to starting on boot, debconf should ask me about hotplugging and 
set USBAUTO based on my answer.  Secondly, 
/lib/udev/gpsd.hotplug.wrapper should not exit immediately when 
START_DAEMON is false (I'm unsure whether the rest of the hotplug 
script should or shouldn't run in the case when the daemon was 
started at boot).

I have tested this on my system by (a) editting /etc/default/gpsd 
to set START_DAEMON=false and USBAUTO=true and (b) reversing the 
sense of the test in /lib/udev/gpsd.hotplug.wrapper, and it now 
seems to do what I wanted.

Regards,  Phil.




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#485994: latencytop: Wish package description mentioned kernel config requirement

[Phil Endecott]

Package: latencytop
Version: 0.4
Severity: wishlist


I just installed latencytop and discovered the kernel configuration 
requirement.  I only installed it out of curiosity and I'm not going 
to build a new kernel just to try it out.  Maybe next time that I'm 
building a kernel anyway...

It would be great if this requirement were mentioned in the package 
description, as seen at packages.debian.org.  If it had been I would
have been saved some effort.

Thanks,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages latencytop depends on:
ii  libc6 2.7-10 GNU C Library: Shared libraries
ii  libglib2.0-0  2.12.4-2   The GLib library of C routines
ii  libncursesw5  5.6+20080531-1 Shared libraries for terminal hand

latencytop recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#482273: initscripts: mountall.sh doesn't wait for kernel to scan for USB storage devices

[Phil Endecott]

Package: initscripts
Version: 2.86.ds1-13
Severity: normal


My machine has its root filesystem on an IDE disk and a second filesystem on a 
USB 
device.  At boot, the kernel scans for and finds the USB device concurrently 
with other 
activities; this process includes a deliberate pause "waiting for device to 
settle" 
before the SCSI system takes over and looks for partitions.  Normally the 
device is ready by 
the time that mountall runs and the USB filesystem is mounted OK; however, 
about one time in 
four, the device is not ready and it is not mounted.  Ideally, initscripts 
would wait 
for an extra second or two until the kernel had created the device and retry 
the mount.

There are a couple of related cases to compare this with:

- If an NFS mount attempt fails, a mount process will be backgrounded to keep 
retrying 
it.  (I think this functionality is inside mount.  See man 5 nfs.)  (Actually 
this can be
the wrong thing to do if the failure was e.g. "mount point does not exist" as 
it hides 
the actual error, see bug #389055.)

- If a USB storage device provides the root filesystem, the kernel and/or the 
initramfs 
code will wait for it to be ready.


Regards,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages initscripts depends on:
ii  debianutils   2.15.3 Miscellaneous utilities specific t
ii  e2fsprogs 1.38+1.39-WIP-2005.12.31-1 ext2 file system utilities and lib
ii  libc6 2.7-10 GNU C Library: Shared libraries
ii  lsb-base  3.2-10 Linux Standard Base 3.2 init scrip
ii  mount 2.12r-10   Tools for mounting and manipulatin

initscripts recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#477838: bluez-utils: hidd is deprecated, use "input service" instead

[Phil Endecott]

Package: bluez-utils
Version: 3.30-3
Severity: wishlist


This report is based on the following message from the bluez users mailing list:
http://thread.gmane.org/gmane.linux.bluez.user/13828/focus=13831

I had set up my bluetooth keyboard based on the instructions in 
/usr/share/doc/bluez-utils/README.Debian, i.e. starting hidd --server by 
enabling it in 
/etc/default/bluetooth, running hcitool scan, and then hidd --conect .  
This method is 
now deprecated; instead we should use something called the "input service" over 
dbus.  See 
the following page, which contains examples of dbus-send commands to do this:
http://wiki.bluez.org/wiki/HOWTO/InputDevices

As a humble user, it looks to me as if the old commands were much mmore usable. 
 However, 
the bluez developers no doubt have their reasons for this change and the Debian 
package will 
inevitably have to adapt eventually when the old commands change from being 
"deprecated" to 
"gone".

I have not actually tried any of these new dbus-send commands myself

Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages bluez-utils depends on:
ii  dbus  1.2.1-1simple interprocess messaging syst
ii  libbluetooth2 3.30-1 Library to use the BlueZ Linux Blu
ii  libc6 2.7-10 GNU C Library: Shared libraries
ii  libdbus-1-3   1.2.1-1simple interprocess messaging syst
ii  libglib2.0-0  2.12.4-2   The GLib library of C routines
ii  libusb-0.1-4  2:0.1.12-2 userspace USB programming library
ii  lsb-base  3.2-10 Linux Standard Base 3.2 init scrip
ii  makedev   2.3.1-80   creates device files in /dev
ii  module-init-tools 3.2.2-2tools for managing Linux kernel mo
ii  udev  0.103-1/dev/ and hotplug management daemo

Versions of packages bluez-utils recommends:
pn  bluez-audio(no description available)

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#477335: bluez-utils: Suggested changes to README.Debian

[Phil Endecott]

Package: bluez-utils
Version: 3.30-3
Severity: wishlist


I'm typing this using a Bluetooth keyboard which it has taken me a long time to 
get 
working.  Here are a couple of suggestions for README.Debian that could have 
saved 
me a bit of time:

hidd --connect 
...
For keyboards, type the PIN on the Bluetooth keyboard and hit
enter, and then type it in on a cable-attached keyboard to pair.

I was typing it at the hidd program, which was ignoring me.  Maybe this is 
obvious 
if you're running gnome (etc) that pops up a dialog, but I wasn't.  So I 
suggest 
replacing that last part of that sentence with:

...and then supply it to the passkey agent (using a cable-attached 
keyboard).

Regarding the passkey-agent program: is there a reason for not including a 
binary? 
It's only 8532 bytes.  Example usage would also be helpful:

$ ./passkey-agent --default 1234

BTW, the "unsupported" /var/lib/bluetooth/.../pincodes feature didn't work for 
me.  
Maybe I was doing something wrong (strace tells me that the file was being read 
by 
hcid).  If this has now broken for some reason you might want to remove that 
section.

Many thanks, Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.25
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages bluez-utils depends on:
ii  dbus  1.2.1-1simple interprocess messaging syst
ii  libbluetooth2 3.30-1 Library to use the BlueZ Linux Blu
ii  libc6 2.7-10 GNU C Library: Shared libraries
ii  libdbus-1-3   1.2.1-1simple interprocess messaging syst
ii  libglib2.0-0  2.12.4-2   The GLib library of C routines
ii  libusb-0.1-4  2:0.1.12-2 userspace USB programming library
ii  lsb-base  3.2-10 Linux Standard Base 3.2 init scrip
ii  makedev   2.3.1-80   creates device files in /dev
ii  module-init-tools 3.2.2-2tools for managing Linux kernel mo
ii  udev  0.103-1/dev/ and hotplug management daemo

Versions of packages bluez-utils recommends:
pn  bluez-audio(no description available)

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#439284: closed by Stig Sandbeck Mathisen <[EMAIL PROTECTED]> (Re: Bug#439284: munin-node: Suggest default plugin timeout is > 10 secs)

[Phil Endecott]

[EMAIL PROTECTED] wrote:

On Thu, Aug 23, 2007 at 11:13:49PM +0100, phil wrote:

It seems that by default the munin-node program waits for 10 seconds for
a plugin to respond.  This value is present in the source of munin-node and
can be over-ridden by a setting in munin-node.conf.  I suggest that the
default munin-node.conf should set a higher value.


Increasing the default will create other problems.


Could you elaborate on that a bit please?


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#469914: scrot: Wish man page said "screenshot"

[Phil Endecott]

Package: scrot
Version: 0.8-7
Severity: wishlist


I need a screenshot about once a year or so, and I can never remember the
name of this program.  But when I type "man -k screenshot", nothing is
found.  It would be great if the word screenshot were included in the
header of the man page, e.g.

NAME
   scrot - Capture a screenshot using imlib2

Cheers,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages scrot depends on:
ii  giblib1  1.2.4-4 wrapper library for imlib2, and ot
ii  libc62.3.6.ds1-4 GNU C Library: Shared libraries
ii  libfreetype6 2.2.1-5 FreeType 2 font engine, shared lib
ii  libimlib21.2.1-2 powerful image loading and renderi
ii  libx11-6 2:1.0.3-6   X11 client-side library
ii  libxext6 1:1.0.1-2   X11 miscellaneous extension librar
ii  zlib1g   1:1.2.3-13  compression library - runtime

scrot recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#443767: libstdc++6-4.1-dev: libstdc++-dev doesn't need to depend on g++; recommend or suggest is sufficient

[Phil Endecott]

Hi Matthias,

Matthias Klose wrote:

won't change that.


Would you like to explain why?

Regards,

Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#458743: boost: Please fix thread safety bug

[Phil Endecott]

Package: boost
Version: 1.34.1-4
Severity: normal


Hi,

Apparently, Boost.Function is not thread safe in 1.34.*.  Boost.Thread uses 
Boost.Function, and so Boost.Thread is itself not thread-safe.  Since a 
non-thread-safe threads library is about as useful as the proverbial chocloate 
teapot, 
it would be great to fix this.

I understand that it will be fixed in 1.35, which may appear soon.  On the 
other hand, 
we have already been waiting a lng time for 1.35 to come out and I'm not 
holding 
my breath.  The recommended fix [see the news item on the boost.org web page] 
is to 
use the current Boost.Function from trunk and to re-compile the libraries that 
depend 
on it (i.e. Boost.Thread).

I don't expect to see many people demanding this fix, since the nature of 
thread-safety problems is that they manifest themselves rarely and are hard to 
reproduce and debug.  However, this is a real bug, and a fix does exist.  I'm 
surprised that Boost have not decided to release a 1.34.2 incorporating the 
fix, but 
that doesn't need to stop Debian from fixing it.

Regard,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#457654: Should be fixed in 1.35

[Phil Endecott]

I am told that this will not be a problem in Boost version 1.35 - due 
"real soon now" - as a result of a redesign of Boost.Serialization.  See:


http://thread.gmane.org/gmane.comp.lib.boost.devel/169668

Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#446668: genisoimage: Please add option to not cross filesystems

[Phil Endecott]

[EMAIL PROTECTED] wrote:

The original mkisofs includes the feature you like via the build in
find(1) since 1.5 years:


Thanks Joerg, that's useful to know.


Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#446821: libjpeg: Decoding faster with -Os

[Phil Endecott]

Package: libjpeg
Version: 6b
Severity: minor


Hi,

Playing around with the upstream libjpeg, it seems to me that decoding is about
15% faster when the library is compiled with -Os than with -O4.  I think that's
significant enough to consider mentioning, since jpeg can be the bottleneck in
some programs.  You would want to see if you can reproduce it on a variety of
different hardware and with a variety of images though.

Just thought you might be interested...

(And no, I haven't actually looked at whether you have changed the default in
the Debian package.  But the djpeg that I've built with -O4 runs at exactly the
same speed as the one from your package, so I'm guessing that you've not tried
-Os.)

Regards,  Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#446668: genisoimage: Please add option to not cross filesystems

[Phil Endecott]

Package: genisoimage
Version: 9:1.1.0-1
Severity: wishlist


Hi,

Many tools that recurse through directory structures have an option to
not cross filesystems.  As far as I can see, genisoimage does not have
such an option.  Could one be added?  I have been trying to backup
a system and found that genisoimage was trying to include /proc!

Thanks, Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages genisoimage depends on:
ii  libc62.3.6.ds1-4 GNU C Library: Shared libraries
ii  libmagic14.17-1  File type determination library us
ii  zlib1g   1:1.2.3-13  compression library - runtime

genisoimage recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#443767: libstdc++6-4.1-dev: libstdc++-dev doesn't need to depend on g++; recommend or suggest is sufficient

[Phil Endecott]

Package: libstdc++6-4.1-dev
Version: 4.1.1-14
Severity: normal


Hi,

Currently libstdc++6-4.1-dev depends on g++-4.1, and vice-versa.

While I think it's reasonable for g++ to depend on libstdc++-dev, I don't see a 
good reason why libstdc++-dev must depend on g++, and I suggest that this is 
reduced to a Recommends: or Suggests:.

In my case I wanted to install libstdc++-dev to use with a cross-compiler on 
another machine, with the libstdc++ include files visible via an NFS mount.  
See 
this thread for the details: 
http://thread.gmane.org/gmane.linux.debian.devel.gcc/25092

Thanks,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages libstdc++6-4.1-dev depends on:
ii  g++-4.1  4.1.1-14The GNU C++ compiler
ii  gcc-4.1-base 4.1.1-14The GNU Compiler Collection (base 
ii  libc6-dev2.3.6.ds1-4 GNU C Library: Development Librari
ii  libstdc++6   4.1.1-14The GNU Standard C++ Library v3

libstdc++6-4.1-dev recommends no packages.

-- no debconf information




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#443576: Strict aliasing problem

[Phil Endecott]

I think I found a bug in gcc-4.2



int i, j;
printf("%d %d\n", j, (void *)(j));


This looks like a strict-aliasing issue to me; you're casting from an 
int to a void*, which is undefined.  Did you get any warnings?  (Did 
you compile with warnings enabled? -Wstrict-aliasing?)  Investigate the 
-f[no]strict-aliasing options.


Regards,

Phil.







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#436695: libagg-dev: Copyright file needs updating as agg is now GPL

[Phil Endecott]

Package: libagg-dev
Version: 2.5-1
Severity: normal


Hi,

According to 
http://packages.debian.org/changelogs/pool/main/a/agg/agg_2.5-1/libagg-dev.copyright

agg is licensed as:

  Anti-Grain Geometry - Version 2.0
  Copyright (C) 2002-2004 Maxim Shemanarev (McSeem)

  Permission to copy, use, modify, sell and distribute this software 
  is granted provided this copyright notice appears in all copies. 
  This software is provided "as is" without express or implied
  warranty, and with no claim as to its suitability for any purpose.

However, my understanding is that since version 2.5 AGG is now licensed 
under the GPL; see http://www.antigrain.com/license/index.html

I suggest that the copyright file should be updated.

You might also like to clarify whether the "GPC" code is included in 
this package.  It seems to me that it is included.  See the antigrain 
license page linked above.  The GPC code has a "free for non-commerical 
use" license, which I suspect is not Debian-compatible.

Regards, Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#429269: cpufrequtils: Please support VIA C7-M e_powersaver in /etc/init.d/loadcpufreq

[Phil Endecott]

Mattia Dongili wrote:

By the way, did the e_powersaver module appear recently?


Quite likely, but not sure.


Should I also
test for some particular kernel version or there was simply no support
for VIA C7 before?


I suspect that there was no support.


Is it ok to default to longhaul for CPU_FAMILY=6 and only use
e_powersaver if the MODEL=10 ?


Probably.

Sorry my answers are vague!  Maybe the kernel change log would have 
better answers?


Cheers,

Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#429269: cpufrequtils: Please support VIA C7-M e_powersaver in /etc/init.d/loadcpufreq

[Phil Endecott]

Package: cpufrequtils
Version: 002-4
Severity: normal


Hi,

/etc/init.d/loadcpufreq includes some logic to guess the appropriate 
frequency scaling module for the current CPU.  For VIA CPUs it has:

CentaurHauls*)
# VIA
if [ $CPU_FAMILY == 6 ]; then
MODULE=longhaul;
fi
;;

'longhaul' is the correct module for the VIA C3, but for the VIA C7 the 
correct module is 'e_powersaver'.

I believe that it's necessary to look at the 'model' field from 
/proc/cpuinfo to determine whether the processor is a C3 or a C7.  
Here's the contents of /proc/cpuinfo on a C3 box:

processor   : 0
vendor_id   : CentaurHauls
cpu family  : 6
model   : 9
model name  : VIA Nehemiah
stepping: 8
cpu MHz : 1002.329
cache size  : 64 KB
fdiv_bug: no
hlt_bug : no
f00f_bug: no
coma_bug: no
fpu : yes
fpu_exception   : yes
cpuid level : 1
wp  : yes
flags   : fpu vme de pse tsc msr cx8 sep mtrr pge cmov pat 
mmx fxsr sse rng rng_en ace ace_en
bogomips: 2006.99

and here it is for a C7 box:

processor   : 0
vendor_id   : CentaurHauls
cpu family  : 6
model   : 10
model name  : VIA Esther processor 1200MHz
stepping: 9
cpu MHz : 400.004
cache size  : 128 KB
fdiv_bug: no
hlt_bug : no
f00f_bug: no
coma_bug: no
fpu : yes
fpu_exception   : yes
cpuid level : 1
wp  : yes
flags   : fpu vme de pse tsc msr pae mce apic sep mtrr pge 
cmov pat clflush acpi mmx fxsr sse sse2 tm nx up pni est tm2 rng rng_en 
ace ace_en ace2 ace2_en phe phe_en pmm pmm_en
bogomips: 800.64
clflush size: 64


Regards,

Phil.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#426944: apt-cross: Silently fails with "-S stable" without --update

[Phil Endecott]

Package: apt-cross
Version: 0.1.0
Severity: normal


Hi,

I've just tried apt-cross for the first time having previously used dpkg-cross. 
 
It looks like it should be a very helpful tool, and my first attempt worked:

# apt-cross -a arm --install librecode-dev

(One suggestion is that it could accept "install" as well as "--install", for 
more consistency with apt-get.  And a -s for simulate would also be 
appreciated.)

So that was OK but I soon realised that it had loaded librecode-dev for 
unstable, 
which has the new ABI!  So I uninstalled it, which worked, and tried to install 
an older version, from stable:

# apt-cross -a arm --remove librecode-dev
# apt-cross -a arm -S stable --install librecode-dev

This simply printed:

One package to install.

straight away and did nothing else.

With "-S testing" it did seem to work: it downloaded the package files and 
then downloaded and installed the requested library and dependencies.

I eventually realised that I needed to do "apt-cross -a arm -S stable --update" 
first.  After that the --install worked fine.

So the bug is that for unstable and testing it seemed to automatically download 
the package files while for stable it didn't.  And/or, that without the package 
files it printed only a very terse message, not "no packages files, please run 
apt-cross --update".

Here is the contents of my sources.list, exluding commented-out lines, in case 
it is relevent :

deb http://ftp.uk.debian.org/debian stable main
deb http://ftp.uk.debian.org/debian/ testing main contrib
deb http://ftp.uk.debian.org/debian/ unstable main
deb-src http://ftp.uk.debian.org/debian/ unstable main
deb http://security.debian.org/ testing/updates main
deb http://tonelli.sns.it/pub/mplayer/etch ./
deb http://mirrors.ecology.uni-kiel.de/debian/debian-multimedia unstable main

Please excuse me if I am doing something stupid (or if you consider my 
expectations unreasonable!), and thank you for this excellent tool.

Regards,

Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages apt-cross depends on:
ii  apt   0.6.43.3   Advanced front-end for dpkg
ii  binutils  2.17-1 The GNU assembler, linker and bina
ii  dpkg-cross1.28   tools for cross compiling Debian p
ii  dpkg-dev  1.13.22package building tools for Debian
ii  libfile-homedir-perl  0.64-1 Get the home directory for yoursel
ii  libwww-perl   5.805-1WWW client/server library for Perl
ii  sudo  1.6.8p12-5 Provide limited super user privile

apt-cross recommends no packages.

-- no debconf information



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#423888: dhcp3-relay: dhcrelay not listed in "man -k DHCP"

[Phil Endecott]

Package: dhcp3-relay
Version: 3.0.4-12
Severity: minor


Hi,

It would be helpful if "man -k DHCP" listed dhcrelay.  It doesn't, 
because the man page begins

NAME
   dhcrelay - Dynamic Host Configuration Protocol Relay Agent

I suggest changing this to:

   dhcrelay - Dynamic Host Configuration Protocol (DHCP) Relay Agent

The same seems to apply to dhclient.

Regards,  Phil.



-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages dhcp3-relay depends on:
ii  debconf [debconf-2.0]1.4.72  Debian configuration management sy
ii  debianutils  2.15.3  Miscellaneous utilities specific t
ii  dhcp3-common 3.0.4-12Common files used by all the dhcp3
ii  libc62.3.6.ds1-4 GNU C Library: Shared libraries

dhcp3-relay recommends no packages.

-- debconf information excluded



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#416911: Please let me know if I can help

[Phil Endecott]

Hi,

Please let me know if there is anything that I can do to help get 
DirectFB 1.0 into Debian.


Many thanks,

Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#402087: xdm: "no home directory" test for Xstartup

[Phil Endecott]

Brice Goglin wrote:

Hi Phil,

Did you have a chance to look at this bug? Julien provided an url to
look at the current code of /etc/X11/xdm/Xstartup.


Yes, see my message with Date: Fri, 02 Mar 2007 22:30:15 +

(You posted a URL for the current Xstartup first; I replied to that 
message; after my reply, Julien posted the same URL again.)


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#416911: directfb: DirectFB 1.0 is available

[Phil Endecott]

Package: directfb
Severity: wishlist


Hi,

DirectFB version 1.0 has been available for a few weeks now.

It would be great to see a Debian package.

Many thanks,

Phil.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#413657: dns-browse seems to need wish, not tcl|wish

[Phil Endecott]

Package: dns-browse
Version: 1.9-4
Severity: normal


I have tcl installed but not wish.  This satisfies dns-browse's 
dependecy:

  Depends: tcl8.4 | wish, dnsutils

However, I believe that wish really is needed:

$ dns_browse 
bash: /usr/bin/dns_browse: /usr/bin/wish: bad interpreter: No such file 
or directory

I see that the dns_tree program, which is included in the package, 
doesn't need wish.  Maybe you wrote the dependency that way because of 
this?  I'm not sure what the Debian rules are for this sort of 
situation.  I would say that either (a) have two packages, or (b) make 
wish a dependency.  Or perhaps it is possible to Suggest or Recommend 
wish?

Regards,  Phil.

-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages dns-browse depends on:
ii  dnsutils  1:9.3.4-2  Clients provided with BIND
ii  tcl8.48.4.12-1   Tcl (the Tool Command Language) v8

dns-browse recommends no packages.

-- no debconf information



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#413371: directfb: Please package ++dfb

[Phil Endecott]

Package: directfb
Severity: wishlist


Hi,

I look forward to diretfb 1.0 packages soon!  It would be great if we 
could also have Debian packages for ++dfb, the "alternative" C++ binding 
for directfb, as well as dfb++.

(I thought that I had already made this request, but searching the bug 
database it seems not.  Appologies if I have missed an earlier request.)

Regards, Phil.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#402087: xdm: "no home directory" test for Xstartup

[Phil Endecott]

Brice Goglin wrote:

http://git.debian.org/?p=pkg-xorg/app/xdm.git;a=blob;h=7e775a3de2fe86d8011f5f5d22ee006e8dba1739;hb=dd2c9398e08e8539d82533207a5eddfffa4ea694;f=debian/local/Xstartup


Thanks.  Hmm, I suspect that the lines that I described were actually 
ones that I added in an earlier attempt to make this work - sorry.  
Anyway, when I investigated this at the start of December I did 
understand it all and got it working correctly, and I'd like to 
encourage you use the Xstartup file below which files the following problems:


1. A "no home directory" message is displayed (this bug).

2. /etc/nologin works with the /etc/nologin message being displayed, 
rather than simply displaying "login incorrect" as currently happens.  
(bug #118677).  For this to work you also need to prevent PAM from 
doing the nologin test itself; see my comment in bug #118677 where I 
point out which line to remove.


Xstartup file follows.


#!/bin/sh
#
# $Id: Xstartup 189 2005-06-11 00:04:27Z branden $
#
# This script is run as root after the user logs in.  If this script 
exits with

# a return code other than 0, the user's session will not be started.

PATH="$PATH:/usr/bin/X11"

if [ -e /etc/nologin ]; then
  # always display the nologin message, if possible
  if [ -s /etc/nologin ] && which xmessage > /dev/null 2>&1; then
xmessage -file /etc/nologin -geometry 640x480
  fi
  if [ "$(id -u $USER)" != "0" ] && \
 ! grep -qs '^ignore-nologin' /etc/X11/xdm/xdm.options; then
exit 1
  fi
fi

# Warn if the user's home directory does not exist.
# It seems that $HOME is set to / after this script
# is executed, so we can't just compare $HOME with ~$USER.
if [ ! -d `eval echo ~$USER` ]
then
  xmessage "Home directory does not exist!"
fi

# Insert a utmp entry for the session.
if grep -qs '^use-sessreg' /etc/X11/xdm/xdm.options \
  && which sessreg >/dev/null 2>&1; then
exec sessreg -a -l "$DISPLAY" -u /var/run/utmp \
 -x /etc/X11/xdm/Xservers "$USER"
  # NOTREACHED
fi

exit 0

# vim:set ai et sts=2 sw=2 tw=0:







--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#402087: xdm: "no home directory" test for Xstartup

[Phil Endecott]

Brice Goglin wrote:

Hi Phil,

It might be too late for my eyes to work fine, but I can't find the
lines your talking about in /etc/X11/xdm/Xstartup (in 1:1.0.5-2). Does
your proposal still apply with this version of xdm?


Hi Brice,

I don't want to update xdm just to see how the file has changed.  Is it 
visible online somewhere?  If not, if you could email me a copy then 
I'll compare it with what I have.


Phil.






--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#394058: xserver-xorg-core: X mouse manpage is in section 4x

[Phil Endecott]

Package: xserver-xorg-core
Version: 1:1.0.2-9
Followup-For: Bug #394058


I was going to report the same issue.

"mouse(4)" is a man page describing the pinout and protocol used by
serial mice.  The page relevent to the X mouse driver is "mouse(4x)"
which you can view using "man 4x mouse".

The xorg.conf manpage SEE-ALSO section should be updated to refer
to "mouse(4x)" rather than "mouse(4)".  The same may be true of the
other driver pages referred to there.

Appologies if this has been fixed recently.  I'm not keen to upgrade
my X server just to see if a man page typo has been fixed!

Phil.


-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages xserver-xorg-core depends on:
ii  libc6 2.3.6.ds1-4GNU C Library: Shared libraries
ii  libfontenc1   1:1.0.1-6  X11 font encoding library
ii  libgcc1   1:4.1.1-14 GCC support library
ii  libxau6   6.9.0.dfsg.1-4 X Authentication library
ii  libxdmcp6 6.9.0.dfsg.1-4 X Display Manager Control Protocol
ii  libxfont1 1:1.0.0-4  X11 font rasterisation library
ii  x11-common1:7.0.19   X Window System (X.Org) infrastruc
ii  xserver-xorg-input-evdev  1:1.1.2-1  X.Org X server -- evdev input driv
ii  xserver-xorg-input-kbd [x 1:1.0.1.3-2X.Org X server -- keyboard input d
ii  xserver-xorg-input-mouse  1:1.0.4-3  X.Org X server -- mouse input driv
ii  xserver-xorg-video-via [x 1:0.1.33.2-3   X.Org X server -- VIA display driv
ii  zlib1g1:1.2.3-13 compression library - runtime

Versions of packages xserver-xorg-core recommends:
ii  xkb-data  0.9-4  X Keyboard Extension (XKB) configu

-- no debconf information



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#407125: inkscape and inkview binaries are very large

[Phil Endecott]

Package: inkscape
Version: 0.44.1-1
Severity: normal


Hi,

The inkscape and inkview binaries are some of the largest on my machine:

$ ls -l /usr/bin | sort -rn -k 5 | head
-rwxr-xr-x 1 root   root8689772 2006-08-28 22:33 hugin
-rwxr-xr-x 1 root   root7599308 2006-03-18 20:00 AbiWord-2.4
-rwxr-xr-x 1 root   root7175116 2006-08-28 22:33 nona_gui
-rwxr-xr-x 1 root   root7138380 2006-08-28 22:33 nona
-rwxr-xr-x 1 root   root6182732 2006-01-16 15:56 mplayer
-rwxr-xr-x 1 root   root5026708 2006-09-25 13:28 inkscape
-rwxr-xr-x 1 root   root5017492 2006-09-25 13:28 inkview
-rwxr-xr-x 1 root   root4016828 2006-05-04 09:11 Xprt
-rwxr-xr-x 1 root   root3372768 2006-03-29 13:46 gs-gpl
-rwxr-xr-x 1 root   root3085008 2006-01-30 21:09 gimp-2.2

I'm particularly sensitive to disk space as I'm running from a 4GB flash drive.

I suspect that the inkscape and inkview binaries are > 90% identical.  Would it 
be 
possible to eliminate one of them, and thereby save nearly 5MB?  Possible 
methods:

- Two links to the same executable which checks argv[0].
- An executable for inkscape and a script for inkview which calls inkscape with 
a 
  view-only flag.
- Two small executables and a shared library.

I appreciate that not everyone shares my desire to optimise disk space and will 
understand if you choose to ignore this report.

Thanks,  Phil.



-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages inkscape depends on:
ii  libatk1.0-0   1.12.3-1   The ATK accessibility toolkit
ii  libbonobo2-0  2.14.0-1   Bonobo CORBA interfaces library
ii  libc6 2.3.6.ds1-4GNU C Library: Shared libraries
ii  libcairo2 1.2.4-1The Cairo 2D vector graphics libra
ii  libdbus-1-3   1.0.1-2simple interprocess messaging syst
ii  libfontconfig12.4.1-2generic font configuration library
ii  libfreetype6  2.2.1-5FreeType 2 font engine, shared lib
ii  libgc1c2  1:6.7-2conservative garbage collector for
ii  libgcc1   1:4.1.1-14 GCC support library
ii  libgconf2-4   2.14.0-1   GNOME configuration database syste
ii  libglib2.0-0  2.12.4-2   The GLib library of C routines
ii  libglibmm-2.4-1c2a2.8.2-2.1  C++ wrapper for the GLib toolkit (
ii  libgnomevfs2-02.14.0-1   GNOME virtual file-system (runtime
ii  libgtk2.0-0   2.8.13-1   The GTK+ graphical user interface 
ii  libgtkmm-2.4-1c2a 1:2.6.5-1  C++ wrappers for GTK+ 2.4 (shared 
ii  liblcms1  1.13-1 Color management library
ii  libloudmouth1-0   1.1.4-2Lightweight C Jabber library
ii  liborbit2 1:2.14.0-1 libraries for ORBit2 - a CORBA ORB
ii  libpango1.0-0 1.14.8-2   Layout and rendering of internatio
ii  libpng12-01.2.15~beta5-1 PNG library - runtime
ii  libpopt0  1.10-3 lib for parsing cmdline parameters
ii  libsigc++-2.0-0c2a2.0.16-3   type-safe Signal Framework for C++
ii  libstdc++64.1.1-14   The GNU Standard C++ Library v3
ii  libx11-6  6.9.0.dfsg.1-6 X Window System protocol client li
ii  libxcursor1   1.1.3-1X cursor management library
ii  libxext6  6.9.0.dfsg.1-6 X Window System miscellaneous exte
ii  libxfixes31:4.0.1-4  X11 miscellaneous 'fixes' extensio
ii  libxft2   2.1.8.2-3  FreeType-based font drawing librar
ii  libxi66.9.0.dfsg.1-6 X Window System Input extension li
ii  libxinerama1  6.9.0.dfsg.1-4 X Window System multi-head display
ii  libxml2   2.6.27.dfsg-1  GNOME XML library
ii  libxrandr26.9.0.dfsg.1-4 X Window System Resize, Rotate and
ii  libxrender1   1:0.9.0.2-1X Rendering Extension client libra
ii  libxslt1.11.1.18-1   XSLT processing library - runtime 
ii  zlib1g1:1.2.3-13 compression library - runtime

Versions of packages inkscape recommends:
ii  imagemagick 7:6.2.4.5.dfsg1-0.13 Image manipulation programs
pn  libwmf-bin (no description available)
pn  perlmagick (no description available)
pn  pstoedit   (no description available)

-- no debconf information



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#407126: hugin-tools: nona and nona_gui binaries are very large

[Phil Endecott]

Package: hugin-tools
Version: 0.6.1-1
Severity: normal


Hi,

The nona and nona_gui binaries are some of the largest on my system:

$ ls -l /usr/bin | sort -rn -k 5|head
-rwxr-xr-x 1 root   root8689772 2006-08-28 22:33 hugin
-rwxr-xr-x 1 root   root7599308 2006-03-18 20:00 AbiWord-2.4
-rwxr-xr-x 1 root   root7175116 2006-08-28 22:33 nona_gui
-rwxr-xr-x 1 root   root7138380 2006-08-28 22:33 nona
-rwxr-xr-x 1 root   root6182732 2006-01-16 15:56 mplayer
-rwxr-xr-x 1 root   root5026708 2006-09-25 13:28 inkscape
-rwxr-xr-x 1 root   root5017492 2006-09-25 13:28 inkview
-rwxr-xr-x 1 root   root4016828 2006-05-04 09:11 Xprt
-rwxr-xr-x 1 root   root3372768 2006-03-29 13:46 gs-gpl
-rwxr-xr-x 1 root   root3085008 2006-01-30 21:09 gimp-2.2

I'm particularly sensitive to disk space as I'm running from a 4GB flash drive.

I suspect that the nona and nona_gui binaries are > 90% identical.  Would it be 
possible to eliminate one of them, and thereby save nearly 7MB?  Possible 
methods:

- Two links to the same executable which checks argv[0].
- An executable for nona and a script for nona_gui which calls nona with a 
  "gui" flag.
- Two small executables and a shared library.

I appreciate that not everyone shares my desire to optimise disk space and will 
understand if you choose to ignore this report.

Thanks,  Phil.




-- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.19
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages hugin-tools depends on:
ii  hugin-data0.6.1-1data files for hugin
ii  libboost-thread1.33.1 1.33.1-4   portable C++ multi-threading
ii  libc6 2.3.6.ds1-4GNU C Library: Shared libraries
ii  libgcc1   1:4.1.1-14 GCC support library
ii  libjpeg62 6b-13  The Independent JPEG Group's JPEG 
ii  libpano12-0   2.8.3-1panorama tools library
ii  libpng12-01.2.15~beta5-1 PNG library - runtime
ii  libstdc++64.1.1-14   The GNU Standard C++ Library v3
ii  libtiff4  3.8.0-3Tag Image File Format (TIFF) libra
ii  libwxbase2.6-02.6.3.2.1.5wxBase library (runtime) - non-GUI
ii  libwxgtk2.6-0 2.6.3.2.1.5wxWidgets Cross-platform C++ GUI t
ii  zlib1g1:1.2.3-13 compression library - runtime

Versions of packages hugin-tools recommends:
ii  hugin 0.6.1-1a Panorama Tools GUI to make panor

-- no debconf information



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#402142: flash-kernel: Minor correction to comment in source code

[Phil Endecott]

Package: flash-kernel
Version: 0.7
Severity: minor


Current comment:

# See http://www.nslu2-linux.org/wiki/Info/BootFlash -- the NSLU2 uses a
# 16 byte MTD header, the first two bytes (big endian) give the length of
# the remainder of the image, and the remaining bytes are zero.  Generate
# this header.

should say "first *four* bytes".



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#402087: xdm: "no home directory" test for Xstartup

[Phil Endecott]

Package: xdm
Version: 1:1.0.4-1
Severity: normal


/etc/X11/xdm/Xstartup includes a couple of commented-out lines that seem 
to have been intended to report a "no home directory" condition to the 
user.  I propose the following in their place:

# Warn if the user's home directory does not exist.
# It seems that $HOME is set to / after this script
# is executed, so we can't just compare $HOME with ~$USER.
if [ ! -d `eval echo ~$USER` ]
then
  xmessage "Home directory does not exist!"
fi


(In NFS environments, missing home directories are not uncommon.  Console 
logon reports the condition loudly.  In an Xsession the results can be 
odd, e.g. lost preferences.  Showing a message is important.)


Phil.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#118677: How to fix xdm and /etc/nologin

[Phil Endecott]

In order to make xdm work with /etc/nologin, there are two issues to fix:

First, by default, nologin is checked by PAM thanks to this line in 
/etc/pam.d/xdm:

authrequisite   pam_nologin.so

Presumably xdm just gets a "no!" reply from PAM, with the result that 
the user sees an unfriendly "login incorrect" message.  I suggest 
removing this from PAM so that xdm can handle it itself.


xdm's attempt to handle nologin is in the first few lines of 
/etc/X11/xdm/Xstartup:

if [ -e /etc/nologin ]; then
  # always display the nologin message, if possible
  if [ -s /etc/nologin ] && which xmessage > /dev/null 2>&1; then
xmessage -file /etc/nologin -geometry 640x480
  fi
  if [ "$(id -u)" != "0" ] && \
 ! grep -qs '^ignore-nologin' /etc/X11/xdm/xdm.options; then
exit 1
  fi
fi

This fails because this script is run as root (see the comment at the 
start), so $(id -u) will always return 0.  The fix is instead to test 
the $USER variable, which xdm passes; see the "STARTUP PROGRAM" section 
of the xdm man page.  Something like this:


  if [ "$(id -u $USER)" != "0" ] ...

This seems to work for me.

I'm using xdm 1:1.0.4-1.





--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#401288: tftpd: Suggest mentioning no tsize in package description

[Phil Endecott]

Package: tftpd
Severity: minor


tftpd does not support "tsize".  I don't know what that is, but it is 
needed by the PXE network bootloader.  atftpd and tftpd-hpa do have this 
feature.

Since network booting is the most common application of tftp, I suggest 
that it is worth mentioning this in the ftfpd package description.  For 
example:

   Tftpd is a server which supports the Internet Trivial File Transfer 
   Protocol (RFC 783). The server is normally started by inetd(8).
   Tftpd is not suitable for use with the PXE bootloader; for that,
   use atftpd or tftpd-hpa.

(See the note at the start of section 4.6 of the Debian Installer 
documentation.)


Phil.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#401175: debian-installer: Could offer to set a label on swap partitions

[Phil Endecott]

Package: debian-installer
Severity: minor


The "partition disks" dialog presents an option to set a partition label 
if "use as" is set to ext3 filesystem.  If "use as" is set to "swap 
area" the label field is not included.  It is possible to set partition 
labels on swap partitions using "mkswap -L".  Labels on swap partitions 
are useful so that /etc/fstab can refer to the swap partition as 
/dev/disk/by-label/swap.  I suggest the D.I. allows the user to set a 
label for their swap partition, and furthermore offers a default label 
of "swap".



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#400262: cupsys: Cannot rename printer from the web interface

[Phil Endecott]

Package: cupsys
Version: 1.2.5-1
Severity: normal


The CUPS web interface does not provide a way to rename a printer.
So, having added a printer using this interface, I am stuck with the 
cryptic name that it chose based on the make and model.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]