Bug#376180: kdevelop3 has wrong name in KDE's K-menu

2006-06-30 Thread Patrik Arvhult
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Package: kdevelop3
Version: 3.2.0-1

Recently I did install kdevelop3 on my Debian Sarge thingy. The installation 
seems broken: kdevelop does appear in KDE:s K-menu but if I click it nothing 
interesting happens.

The menu entry (for KDE/CPP dev) is:  kdevelop --profile KDECppIDE %u
I would like it to be:  kdevelop3 --profile KDECppIDE %u
or maybe a symbolic link could do it.

Best regards,
 Patrik Arvhult

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFEpW0ajauAa1f6YKoRAru7AKCCcth64oYrxIRigtZmWGP9EzAV6gCgs3Vj
5Gvjwx+Y1m5Od7q01RRkinY=
=z2Ag
-END PGP SIGNATURE-


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



Bug#350635: Using debootstrap: dpkg needs leading / to be able to find archives

2006-01-30 Thread Patrik Arvhult
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Package: debootstrap
Version: 0.3.3

debootstrap, debootstrap somewhat patched as described for bug # 350583.

If running:

Puter#~/> debootstrap --arch i386 sarge /mnt/debinst \
http://ftp.se.debian.org/debian

on a Slackware Linux 1.2.3 system,

the debootstrap package,  downloads  packages properly and executes this 
command:

chroot /mnt/debinst dpkg --force-all --install \
var/cache/apt/archives/

and fails.

if trying manually, i find that:

chroot /mnt/debinst dpkg --force-all \
--install /var/cache/apt/archives/

does work. So a leading slash does help.

Similar: "chroot /mnt/debinst ls bin/ls"
fails. But "chroot /mnt/debinst ls /bin/ls" does work..

Looking into scripts sarge and sarge.build I see these paths are generated by 
the script named "functions", and is emitted at row 369 by a utility function 
"debfor".

change:
script named functions, row 369 was:

echo $path

changed to

echo /$path


This seems to help in my particular case. debootstrap moves on unpacking and 
configuring packages.

But in general I suspect it to be a bad solution. Maybe the need of this 
extra / should be user-supplied in some better way or the need of it beeing 
calculated/tested.

Best regards,
Patrik Arvhult
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFD3n/+jauAa1f6YKoRAjX9AKC+uaQ2hk9rP/dw7L3n8uk4Kba/FgCgqkkZ
F9nOwNeLT9Ra59dEyQa3HPg=
=BIPM
-END PGP SIGNATURE-


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



Bug#350583: Yes, it was related to a gnu sed extension

2006-01-30 Thread Patrik Arvhult
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Found my problem was related to sed:

A script named "sarge" in the debootstrap package contains at row 115:

  sed -ne 's/^Version: *//Ip' | head -n 1

looking at the flags Ip: I is some gnu extension for ignore-case, thereof sed 
will fail on all systems lacking this extension.

I changed row 115 to:

  sed -ne 's/^[Vv][Ee][Rr][Ss][Ii][Oo][Nn]: *//p' | head -n 1

and did also do the same change for the script sarge.build at row 77.

Above is a regex accepting the string Version beeing upper/lower-cased in 
whatever possible way like veRSIon or VERSION, without using the I flag. I 
think this makes sense in a boostrap package. (Or, is there some old regex 
that dont accept re:s with brackets []?)


Woohoo! Now it works, so far..., best regards,
Patrik Arvhult

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFD3mCIjauAa1f6YKoRAmvRAKCv9SRk8H71PlFj4HhOOhJ19gxTqwCfSkcw
zMEg/DjOoEEaSQaUc9hsecE=
=pwPG
-END PGP SIGNATURE-


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



Bug#350583: debootstrap 0.3.3

2006-01-30 Thread Patrik Arvhult
Package: debootstrap
Version: 0.3.3

I am trying to run a bootstrap from Linux 1.2.3 (maybe it is slackware 96) to 
debian Sarge. (No CDROMS, only floppies/network avail. Using the 
debootstrap_0.3.3.tar.gz bootstrapper I found refered to by the debian 
webbpages. Actually I didnt find any information on what version to use, 
0.3.3 seems to be the newest one.

Found these two bugs:

1)
debootstrap_0.3.3.tar.gz -> file pkgdetails.c

there is a int "any" declared in a new style. Not very portable and not 
feasible if bootstrapping from some old system. Annoying, but good thing 
there were only one of these and not zillions ;)

Change row 43 from "int any=0;"  to just "any=0;"
Insert "int any=0;" betwen row 23 and 24.

((Probably the first initializations isnt really needed, I didnt take the time 
check, just did put in somthing that should work.))

2)

Something similar to a tar -O bug seems to be present in this debootstrap 
package, at a glance it seems to be the same error as mentioned at: 
http://lists.debian.org/debian-boot/2003/05/msg00450.html, bug 193788. That 
is another package, but the error seems very similar:

running "debootstrap --verbose --arch i386 sarge /mnt/debinst \ 
http://ftp.debian.org/debian";

All packages is downloaded, when unpacking starts it stops with this message:

  dpkg: parse error, in file '/var/lib/dpkg/status' near line 2
package 'dpkg': error in Version string '': version string is empty"

Investigating /mnt/debinst/var/lib/dpkg/status shows that the message is 
right: the version is missing: the label is there but its value-field is 
leaved empty. I can fill this one in manually, but the same will happen for 
all packages, and debootstrap does always start over from the beginning 
erasing my changes.

Checking out tar on my Linux 1.2.3 I find that -O and --stdout functionality 
is present and seems to work. Maybe this is something related to sed? 

Also running "chroot /mnt/debinst tar --help | less" does show support for -O 
and --stdout. And some tests using the chrooted tar with -O does work. So the 
downloaded tar seems to be ok too. Strange.


Best regards,
Patrik Arvhult   -  fighting windmills
Sweden




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



Bug#341712: I take my bug-report back, sorry for disturbing.

2005-12-02 Thread Patrik Arvhult
I take my bug-report back, sorry for disturbimg. I found that lost mail i was 
writing when my harddisk went full. So kmail did save it some way or did keep 
the window somewhere, not loosing anything. Good. Very good.

Best regards,
Patrik Arvhult
Sweden
 


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



Bug#341712: Kmail shuts down if out of storage-space, without saving work

2005-12-02 Thread Patrik Arvhult
Package: kmail
Version: 4:3.3.2-3

Using ix86 Debian Sarge distro.

Hi!

If kmail gets out of space (here it happened when I was moving some mails, and 
also was writing a mail) it shuts down showing a message-box telling its out 
of harddisk space and thereof shuts down. No options! So - I did just hope my 
mail - about one hour of writing - did exist in some temporary file. But it 
seems not.

When kmail gets out of space: i would like it to wait until space is avail. 
Maybe not regarding all functionality, but at least for being able to rescue 
the mail(s) I was currently editing - and then shut down.

Or maybe something like for example how Pine, or the Jove editor behaves: if 
something goes wrong, theres often a backup around.

Also NetBSD:s great feature of automagical buffering of "lack-of-disk-space" 
into memory comes to my mind: kmail or entire kde could for example somewhat 
emulate this by saving files in two passes: the first pass writes an image in 
memory. A second pass tries to copy this image to the final storage device. 
If theres a failure: ask someone to free some space - then try again. Of 
course this would not be needed if running NetBSD; but in that case the user 
need to be informed about that virtual storage space is used and that the 
storage really is full and need attention. ((Even better would be if this 
great feature of NetBSD is put into the Linux kernel: makeing avail space on 
a harddisk going negative, keeping the spill in memory until someone notice 
and solves the storage problem or its self-solved due to an allocation-peak 
going back.))

Kmail could also ask me for some location I would like it to try to save the 
current work to. This for each mail-editing window. Not really a 
rocket-science feature. Often people have several storage devices available, 
even portable and hotpluggable ones.

Another way is to show a big blue screen stating "General error, rescue not 
implemented. Please calm down." and some nice special selected hexadecimal 
digits along with it, at least making it a great shutdown. ;)

Best regards,
Patrik Arvhult


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



Bug#341687: Kmail goes sigsegv due to timing of user actions

2005-12-02 Thread Patrik Arvhult
Package: kmail
Version: 4:3.3.2-3

Using ix86 Debian Sarge distro.

Hi!

When I start up my kmail it shows the default 3 sections  (mailfolders, list 
of mails in inbox, and the newest mail in the inbox). The first mail it shows 
happens to be gpg-encrypted.

My puter is kinda old and slow (it kinda takes its clock from the 50Hz 
mains-line ;) and thereof it takes some time for the decryption-password box 
to show up ((its the debian sarge standard distro Kgpg agent)). So before it 
comes up i click some other mail.

However this doesn't make kmail to stop the decryption procedure: the password 
box shows up and I think: "kk - whatever u want, stupid software - just dont 
crash on me!" and fills in my secret key password.

When I complete the password entry hitting carriage-return, kmail goes 
sigsegv.

For me, this bug is able to show up because my computer is slow and because 
kmail apparently is able to enter some bad state due to timing of user 
actions.

But I guess it also can, will and has, happen(ed) on heavily loaded, faster 
systems.

Instead of this behavior I want kmail to do either: 1 - not accept user 
commands for displaying another mail before the process of drawing (including 
decrypting and such) of the current mail is done. Or 2: In some way kill / 
shut down / recall its routines (threads or calls or whatever) of decrypting 
a mail when a user selects it to do somthing else. Or 3: Do what it is now 
doing in a proper way, not going sigsegv - but i regard this last variant as 
a waste of computer resources and users time because apparently the user 
doesnt want the mail displayed (and also not decrypted).

Best regards,
Patrik Arvhult


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



Bug#316900: Varning should be included for the Webcollage screensaver

2005-07-04 Thread Patrik Arvhult
Package: xscreensaver
Version: 4.21-3

Hi! I suggest a warning-message (or mail) is added to the installation of 
xscreensaver, this warning should tell xscreensaver contains webcollage which 
can show adult material.

(After upgrading my debian woody to sarge I configured kde to select 
screensavers at random - maybe something was new..  I was curios..)

Yesterday I found some screensaver showing a women sucking at a dogs penis.

My first thought was - this cannot really be good for debian. Over time this 
kinda stuff will show up, and at the wrong places: at least a warning would 
be good. Hell - I got childrens and parents visiting here! On the other hand 
it had been a cool view in my boss:s office.

Best regards,
 Patrik Arvhult
  


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