Bug#941214: Completion for mutt's -a command line switch
On 20 Jun, martin f krafft wrote: > mutt allows attaching files from the command-line: > > mutt -a /file/one /file/two /file/three -- … > > Basically, the rules are: -a takes a list of files, terminated by --. > > Zsh's completion of mutt treats the argument to -a as optional: > > '*-a[attach file using MIME]::file attachment:_files' \ In theory, this should be possible with: '*-a[attach file using MIME]:*--:file attachment:_files' \ However, _mutt also calls _arguments with -S so if it sees -- it takes that as the end of all options and only completes recipients thereafter. Oliver
Bug#891140: Please make "apt install " completion work on "local" .deb files
Chris Lamb wrote: > > Thanks for applying this. I was wonder if you could also include: > > *.changes > *.dsc > > ie. as well as *.deb. All behaviours of path-based .debs can be > inherited. Thanks in advance :) Do you mean like the following patch - considering all .changes and .dsc files alongside .deb? Or in some other context? Oliver diff --git a/Completion/Debian/Command/_apt b/Completion/Debian/Command/_apt index 645b02279..dd837379c 100644 --- a/Completion/Debian/Command/_apt +++ b/Completion/Debian/Command/_apt @@ -411,7 +411,7 @@ _apt-cmd () { /$'(install|download|source|build-dep)\0'/ -'subcmd=${match%?}' \ /$'[^\0]#\0'/ ':packages::_deb_packages -qr "\n\t\- /=" "$expl_packages[@]" avail' \# \ \( \ - // '%(.|)/%' -'[[ $subcmd = install ]]' /'[]'/ ':files:package file:_files -g "*.deb(-.)"' \| \ + // '%(.|)/%' -'[[ $subcmd = install ]]' /'[]'/ ':files:package file:_files -g "*.(changes|deb|dsc)(-.)"' \| \ /$'[^\0/=]#\/'/ /'[]'/ ':apt-releases:release name:_apt_releases' \| \ /$'[^\0/=]#='/ /'[]'/ ':apt-package-versions:package version:_apt_versions_of_binary_package' \| \ \) \
Bug#777930: ksh: ftbfs with GCC-5
I queried the ast-users mailing iist and both Arch and opensuse apparently have patches for the issue. See here http://lists.research.att.com/pipermail/ast-users/2015q3/004773.html Oliver
Bug#679824: Buggy completion with "perl -e", "python -c", and "ruby -e"
This issue was addressed upstream. The fix was included in 5.0.6 and, by extension, 5.0.7. Let me know if you think there is still something to address. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#285266: zsh: ecasound completion doesn't complete file names
Just for your information, this has now been fixed in upstream git. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#269769: zsh: sabcmd completion doesn't work
Just for your information, this has now been fixed in upstream git. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#720441: Got a fix, how to submit patch.
ольга крыжановская wrote: > What is the way to replace the patched up 2012-02-29? Find sponsor, > declare him how bad it is? I think so. One thought is that you could try to pester Jonathan Wiltshire : he did 93u+-1.2. It is probably wise to prepare the replacement first. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#720441: Got a fix, how to submit patch.
Phi Debian wrote: > On Wed, Oct 9, 2013 at 4:11 PM, Oliver Kiddle wrote: > > > > > Debian does have ksh93u. > > Ambiguous Clearly we've got ambiguity in Debian versions of ksh too: wheezy does have 2012-02-29, testing/unstable have 2012-08-01. 2012-08-01 will never go into wheezy, the best we can do is patch (or wait for jessie to be stable). As Olga is taking over ksh, you'll have to persuade her to roll a 93u+-1.3 for wheezy. What apt-get source gives you depends on how your debian installation is configured. Frustratingly, I tried very hard to have 2012-08-01 pushed into wheezy during the freeze but the lack of a sponsor combined with strict application of Debian policy meant that wheezy ended up with a patched up 2012-02-29. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#720441: Got a fix, how to submit patch.
Phi Debian wrote: > It appears after some back and forth mail with attresearch (code > owner) that the source code used by debian (and ubuntu) are both out > of date, they both have this code dump or corruptin bug that was fixed > at ksh93u (dixit David Korn). Debian does have ksh93u. Latest version at the following URL is what Debian has: http://www2.research.att.com/~gsf/download/ I think there is a more recent version that is declared a beta or alpha. This said, I wouldn't be surprised if the download link is out-of-date given that the maintainers are leaving AT&T. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#720441: Got a fix, how to submit patch.
On 4 Oct, you wrote: > How should I proceed for a patch submit, I did provide it to att > owner, so I guess future release will have the fix, but for debian > should I do something, does debian has its own set of 'fixes' (I am > ignorant here) You an include a small patch in an e-mail to this bug report. Then if someone wants to adjust the package to incorporate the patch then they can do so. Otherwise, it'll reach Debian along with the new upstream release. Thanks Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#679966: ksh: cd ../name does not work
John Paul Adrian Glaubitz wrote: > Then you should address the problem as the maintainer of ksh. Since > it's grave bug, it's blocking the release of Debian Wheezy and should > therefore addressed as soon as possible! The package has been up for adoption for some time because I don't have the time or interest to deal with this issue. Fixing it properly without either introducing new bugs or missing parts of the original bug is not trivial for me. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#679966: ksh: cd ../name does not work
John Paul Adrian Glaubitz wrote: > Works for me with version 93u+-1.1: > Therefore marked as fixed in 93u+-1.1. Note that bug 691200 was merged in as a duplicate of this one and the example in the initial description for that bug is most definitely not fixed in 93u+-1.1. Issues with cd .. are wider than the example you tried and the bug remains unfixed in 93u+-1.1. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#687779: Bug#679966: ksh: diff for NMU version 93u+-1.1
Jonathan Wiltshire wrote: > > > In my opinion, it would be > > better if the debian release goes out with either the old squeeze release > > (93s+20080202-1) or 93u+20120628-1 (which was only bug fixes). > > Well, I think the release team have already expressed that they are less > than happy with having a new upstream at this stage. I think the cleanest I wasn't aware of ever having a clear answer from the release team though I don't know who constitutes the release team. On the bug there is just the one-line sarcastic rhetorical question. > thing you could do is revert back to the Squeeze version with an epoch; That seems the less icky of the two. Rather than an epoch, I'd probably just ditch the 93u part of the version and use the date stamp. Upstream have admitted that the 93 will probably never be changed and their downloads only use the date. I might probe upstream first to see if I can get backported fixes (though it was applying backported fixes from upstream that caused much of this mess to begin with). Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#687779: Bug#679966: ksh: diff for NMU version 93u+-1.1
On 13 Oct, Jonathan Wiltshire wrote: > > I've prepared an NMU for ksh (versioned as 93u+-1.1) and > uploaded it to DELAYED/00. Please feel free to tell me if I > should delay it longer. While this addresses the direct issue that bug #679966 concerns, the cd builtin was broken in various ways in that release and I don't think it goes far enough. For example, cd /etc;cd .. leaves you in /etc. That's a grave bug as essential scripts could be broken. If you want to dig into the upstream changes feel free but it'd be hard to avoid breaking things or leaving other subtle cd issues unresolved. In my opinion, it would be better if the debian release goes out with either the old squeeze release (93s+20080202-1) or 93u+20120628-1 (which was only bug fixes). I was away last week so sorry for not having got back to you sooner. Even so, a bit of notice of your intentions would have been nice. Note that I have put the ksh package up for adoption. Oliver Kiddle -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#691200: ksh: cd .. doesn't work in many cases
Package: ksh Version: 93u+-1.1 Justification: renders package unusable Severity: grave In many situations, cd .. is broken with this version of ksh. For example, cd /etc; cd .. remains in /etc. Thank you Oliver Kiddle -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#691199: RFA: ksh -- Real, AT&T version of the Korn shell
Package: wnpp I'd be grateful if someone could take over. I'd be happy to help by answering questions etc. Despite my sponsor having been really helpful with this package it doesn't seem fair to continue to go through him or the most efficient for that matter. I think the package would benefit from someone more involved with Debian. Certain annoyances with the Debian project coupled with the fact that I now only use Debian stable (on ARM) are also behind my decision to put the package up for adoption. It'd be especially good to have someone interested in using this as the basis for packaging the full ast-open set of software. Thanks Oliver Kiddle -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#687779: unblock: ksh/93u+20120628-1
Mehdi Dogguy wrote: > So intead of reverting a patch [1] and re-open a severity "normal" bug, > you uploaded a new upstream release [2] ? I would have classified the original bug as grave too. The upstream release is a bug fix release: they didn't even change the version number hence the appended date. And it seemed to be the approach that would best limit the number of times I'd need to bother my sponsor. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#687779: unblock: ksh/93u+20120628-1
Package: release.debian.org Severity: normal User: release.debian@packages.debian.org Usertags: freeze-exception Hi, Please unblock package ksh/93u+20120628-1 which fixes #679966 (severity grave). Frankly it would be better to exclude ksh from the next Debian release than include it with this bug because it breaks very basic functionality in the shell. The fix was done upstream so I don't really know the details. Thanks Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#679966: Possible security issues and a temporary fix
Zoltan Hidvegi wrote: > More info: this is not an upstream issue, it's caused by the > handle-removed-working-dir.patch which is an attempt to fix > #667038. Besides chopping off the last path component of any cd > ../name command, it also does sfprintf(shp->strbuf,oldpwd) which is a > problem if oldpwd contains printf formatting escapes, which probably > could be exploited. Workaround is to remove > handle-removed-working-dir.patch which would then reopen #667038. The patch was received from upstream. There's just been a new upstream release which resolves both issues. So either the patch I got from upstream relied on other changes since the release or the second issue was found and fixed later. Anyway, I've prepared a package of the new release and depending on how busy my sponsor is it'll hopefully be uploaded before long. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#677540: "send" fails with "post" errors
James wrote: > After this morning's upgrade to nmh 1.5-release-0.2, "send" fails with > > post: message has no From: header > post: See default components files for examples > post: re-format message and try again > send: message not delivered to anyone This is the result of a change in the upstream nmh and not a Debian packaging bug. You likely have your own custom components and replcomps etc files and these need updating to include a From: header. If you want the generated for your from the username and local hostname, look at /usr/lib/mh/components. Otherwise, you can simply include a hardcoded >From header. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#667038: ksh93 fails to detect renamed current or parent directory
Below is a report made via the Debian bug tracking system. I'm guessing that when handling `..', ksh is trying to use what it previously knew to be the path to the current directory whereas the other shells aren't specifying a path. Oliver --- Forwarded Message Date:Tue, 03 Apr 2012 16:57:12 +0200 From:Vincent Lefevre To: Debian Bug Tracking System Subject: Bug#667038: ksh93 fails to detect renamed current or parent directory Package: ksh Version: 93u-1 Severity: normal ksh93 fails on the following script: ypig:~> cat tst.sh mkdir tmpdir1 cd tmpdir1 mv ../tmpdir1 ../tmpdir2 cd .. rmdir tmpdir2 ypig:~> ksh93 tst.sh tst.sh[4]: cd: ..: [No such file or directory] rmdir: failed to remove `tmpdir2': No such file or directory dash, bash, zsh, pdksh, mksh and posh all behave correctly on this example. Note: similar problem when renaming a parent directory. - -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experi mental') Architecture: amd64 (x86_64) Kernel: Linux 3.2.0-2-amd64 (SMP w/8 CPU cores) Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages ksh depends on: ii libc6 2.13-27 ksh recommends no packages. ksh suggests no packages. - -- no debconf information --- End of Forwarded Message -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#649968: ksh: traps and 'set -e' set inside a function are ignored
On 25 Nov, Blazej Krzeminski wrote: > Traps set inside KSH functions and 'set -e' (exit on error) command > are ignored when the function is used in a command substitution > and the command substitution $( ) is contained in a || or && command, > or is following an 'if'. This problem does not occur in pdksh. This is intended behaviour. ksh is following the POSIX standard here whereas pdksh is not. Conditionals such as || && and if are would be useless together with set -e if the trap were triggered. So the standard says that set -e is not honored during conditonals like if, && and ||. The fact that it occurs inside a function and with command substitution makes no difference. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#452785: Can not input UTF-8 chars in prompt
I'm still unable to reproduce this. It'd be good to have a clearer ide of your environment. What edit mode is being used - vi or emacs? Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#433026: ksh: setting & unsetting emacs mode causes segfault
Can you still reproduce this? I'm still be unable to. Any chance of a core file? Without one, I don't really have a hope of fixing it. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#633524: ksh fc command calls /bin/ed (doesn't exist)
On 11 Jul, auto58326...@hushmail.com wrote: > Problem: > When issuing the fc command, the program searches for /bin/ed which > doesn't exist in recent default debian installs. I'm concerned that changing the default is unwise because some user's may be used to it being the default. I always uninstall nano on my machines. At least ed will quit if you type q which was my first guess. Nano keystrokes are all a bit strange and it requires a curses capable terminal. Any idea if something from /etc/alternatives is appropriate. On my system editor points to something that requires X which isn't ideal. ex might be a reasonable choice. Does that exist on a minimal system. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#636580: Fwd: ${-#pattern} fails in set -u mode
Following is a bug reported to the Debian bug tracking system. --- Forwarded Message Date:Thu, 04 Aug 2011 18:42:18 +0900 From:SATOH Fumiyasu To: sub...@bugs.debian.org Subject: Bug#636580: ksh: ${-#pattern} fails in set -u mode Package: ksh Version: 93u-1 Severity: important ksh treats ${-#pattern} (and others) as an error with set -u mode. $- is always set by ksh, thus ${-#pattern} must NOT fails whether set -u or not. $ /bin/ksh --version version sh (AT&T Research) 93u 2011-02-08 $ /bin/ksh -c 'set -u; echo $-; echo ${-#dummy}' chsuB /bin/ksh: line 1: -: parameter not set $ /bin/ksh -c 'set -u; echo $-; echo ${-##dummy}' chsuB /bin/ksh: line 1: -: parameter not set $ /bin/ksh -c 'set -u; echo $-; echo ${-%dummy}' chsuB /bin/ksh: line 1: -: parameter not set $ /bin/ksh -c 'set -u; echo $-; echo ${-%%dummy}' chsuB /bin/ksh: line 1: -: parameter not set $ /bin/ksh -c 'set -u; echo $-; echo ${-/dummy1/dummy2}' chsuB /bin/ksh: line 1: -: parameter not set - -- - -- Name: SATOH Fumiyasu (fumiyas @ osstech co jp) - -- Business Home: http://www.OSSTech.co.jp/ - -- Personal Home: http://www.SFO.jp/blog/ --- End of Forwarded Message -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#571796: console-tools: Virtual terminals no longer visible
I'm also experiencing exactly this problem on a fully updated Debian testing system. It broke at about the point the kernel was upgraded from 2.6.32 to 2.6.38. I've not got any non-default grub options that I'm aware of. I've got a fairly old PCI Matrox G450 graphics card so no support for kernel mode setting. Have you got any idea how I might fix it or diagnose the problem? Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#605630: Unreasonable and big use of resources by: while(true)
The following has been sent to the Debian bug tracker for ksh. --- Forwarded Message Date:Wed, 01 Dec 2010 23:56:39 +0100 From:=?UTF-8?Q?Pawe=C5=82?= Kalinowski To: Debian Bug Tracking System Subject: Bug#605630: ksh: Unreasonable and big use of resources by: while(true) ; do true; done Package: ksh Version: 93s+20080202-1 Severity: normal `while(true); do true; done` uses all the RAM, one whole core of CPU and makes system almost irresponsible(after filling all the RAM) Replacing true with other commands do not change anything The same command works correctly in all other shells(constant and small use of memory, limited to ~15-20% use of CPU) Version without spawning additional shell, `ksh -c "while true; do true; done"` works properly - -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing'), (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35.7 (SMP w/2 CPU cores) Locale: LANG=en_CA.UTF-8, LC_CTYPE=en_CA.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages ksh depends on: ii libc6 2.11.2-7 Embedded GNU C Library: Shared lib ksh recommends no packages. ksh suggests no packages. - -- no debconf information --- End of Forwarded Message -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#616027: postinst script fails for halevt
Package: halevt Version: 0.1.6.2-1.4 Severity: normal The post installation script appears to be failing when I install halevt. The problem is with the following line: invoke-rc.d --quiet hal restart This is returning status 100 because there is no /etc/init.d/hal Is there a missing package dependency perhaps? Or has the hal package changed so as not to use an init script? Oliver -- System Information: Debian Release: wheezy/sid APT prefers testing APT policy: (700, 'testing'), (600, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-5-amd64 (SMP w/4 CPU cores) Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages halevt depends on: ii adduser 3.112+nmu2 add and remove users and groups ii debconf [debconf-2.0 1.5.38 Debian configuration management sy ii dpkg 1.15.8.10 Debian package management system ii hal 0.5.14-5Hardware Abstraction Layer ii install-info 4.13a.dfsg.1-6 Manage installed documentation in ii libboolstuff-0.1-0 0.1.12-3library for operating on boolean e ii libc62.11.2-11 Embedded GNU C Library: Shared lib ii libdbus-1-3 1.4.1-2 simple interprocess messaging syst ii libdbus-glib-1-2 0.88-2.1simple interprocess messaging syst ii libglib2.0-0 2.24.2-1The GLib library of C routines ii libhal1 0.5.14-5Hardware Abstraction Layer - share ii libxml2 2.7.8.dfsg-2GNOME XML library ii lsb-base 3.2-27 Linux Standard Base 3.2 init scrip ii passwd 1:4.1.4.2+svn3283-3 change and administer password and halevt recommends no packages. halevt suggests no packages. -- 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#599202: ksh: env. var. set prefix for a command is not working
"Jeffrey B. Green" wrote: > My assumption is that two forms said to be equivalent imply the results > of the forms are identical, they can be substituted for each other. > Otherwise, they not not equivalent, just similar. What is your > definition of equivalent? But the man page doesn't just say "equivalent", it says "equivalent (as far as the above execution of cmd is concerned". And the whole paragraph is clearly just talking about the environment passed to an executed program. The evaluation of a command's arguments is all done as a separate thing. If you can think of some better wording which doesn't clutter it up and detract from the main point that is being made then I'll put it forward to upstream. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#599202: ksh: env. var. set prefix for a command is not working
"Jeffrey B. Green" wrote: > The question is then: what do users do with the explanation on the > manpage? I.e. they are not equivalent since the results are not the same. But they are equivalent "as far as the above execution of cmd is concerned". When you run XXX="hello" /bin/echo $XXX, the echo command will have XXX in it's environment. That's why I used printenv to show that the manpage is correct. There's nothing wrong with the explanation on the manpage. Your attempt to verify the example on the manpage failed because it relied on false assumptions about how other things work - i.e. variable expansions. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#599202: ksh: env. var. set prefix for a command is not working
Jeff Green wrote: > > % XXX="hello" /bin/echo $XXX When the shell see's this line, it expands $XXX itself first. It is not /bin/echo doing the expansion. Note that if you do the following, it will print hello. XXX="hello" printenv XXX Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#597776: issues with colour prompt
The following issue has been reported for the Debian package. After completion, ksh appears to repeat just the last line of the PS1 prompt. In the case of this coloured prompt, it is outputting: [0;33m[[0m156[0;33m][0m$ cl So why the duplicated escape characters? Thanks Oliver --- Forwarded Message Date:Wed, 22 Sep 2010 22:14:42 +0100 From:Peter T To: Debian Bug Tracking System Subject: Bug#597776: ksh: issues with colour prompt Package: ksh Version: 93s+20080202-1 Severity: normal Background - on my previous computer running Lenny I did almost all my console stuff in gnu screen which is not effected by this bug. On my new computer using Squeeze I've opted for a tiling window manager and lots of terminals in place of gnu screen. At the moment I am being forced to use a mono prompt PS1=`uname -n`':${PWD/#$HOME/\~} [!]\$ ' because of issues with my colour prompt PS1="^[[0;33m`uname -n`^[[0m:^[[1;33m"'${PWD/#$HOME/\~} ^[[0m ^[[0;33m[^[[0m!^[[0;33m]^[[0m\$ ' This color prompt works fine most of the time but when using tab completion I end up with something like host:~ [1]$ cl 1) cls 2) /usr/bin/clear 3) /usr/bin/clear_console 4) /usr/bin/cleanlinks host[0m:~ [0;33m[[0m1][0m$ cl hitting enter returns the prompt to normal I do not observe this behaviour in gnu screen is being run nor if the prompt is being run on a remote computer over ssh or on the linux console (no X). This behaviour has been observed using rxvt-unicode (urxvt) and vte based terminals (roxterm and sakura). Now it could be that I've made mistakes in my prompt if so I apologise for wasting your time but it may also be a bug. - -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Kernel: Linux 2.6.35-4.dmz.2-liquorix-686 (SMP w/1 CPU core; PREEMPT) Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages ksh depends on: ii libc6 2.11.2-5 Embedded GNU C Library: Shared lib ksh recommends no packages. ksh suggests no packages. - -- no debconf information --- End of Forwarded Message -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#569040: build of webkitimageqt on amd64
I tried rebuilding from source because I have an amd64 system. You're missing a build dependency on libqt4-dev which it needs for moc-qt4. With that installed, it build and installed on amd64. Still can't get the WMS plugin to work though - just red boxes containing "Exception occurred". What do you put in WMS Plugin Preferences, I only have landsat at http://onearth.jpl.nasa.gov/wms.cgi?request=GetMap&layers=global_mosaic&styles=&format=image/jpeg&; and that doesn't work. What's the URL for yahoo? Thanks Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#577716: uzbl: bookmark scripts fail with dash as /bin/sh
Package: uzbl Version: 0.0.0~git.20100403-1 Severity: normal Tags: upstream The default config file and scripts for inserting bookmarks rely on the -e option to echo. I'm not quite sure why they want to interpret backslash escapes but in any case, the -e option is not defined in POSIX and should not be used. My bookmarks file ends up having lines such as: -e http://www.independent.co.uk/ news Thank you Oliver Kiddle -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (700, 'testing'), (600, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-trunk-amd64 (SMP w/4 CPU cores) Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages uzbl depends on: ii libatk1.0-0 1.30.0-1The ATK accessibility toolkit ii libc62.10.2-6Embedded GNU C Library: Shared lib ii libcairo21.8.8-2 The Cairo 2D vector graphics libra ii libfontconfig1 2.8.0-2 generic font configuration library ii libfreetype6 2.3.11-1FreeType 2 font engine, shared lib ii libglib2.0-0 2.22.4-1The GLib library of C routines ii libgtk2.0-0 2.18.9-2The GTK+ graphical user interface ii libpango1.0-01.26.2-1Layout and rendering of internatio ii libsoup2.4-1 2.29.91-1 an HTTP library implementation in ii libwebkit-1.0-2 1.1.17-2Web content engine library for Gtk ii libx11-6 2:1.3.3-2 X11 client-side library ii python 2.5.4-9 An interactive high-level object-o ii python-gobject 2.20.0-1+b1 Python bindings for the GObject li ii python-gtk2 2.16.0-2Python bindings for the GTK+ widge ii python-support 1.0.7 automated rebuilding support for P uzbl recommends no packages. Versions of packages uzbl suggests: ii dwm-tools 31-1 dynamic window manager (tools) ii python-simplejson 2.1.0-1simple, fast, extensible JSON enco ii socat 1.7.1.2-3 multipurpose relay for bidirection ii xclip 0.12-1 command line interface to X select ii zenity2.28.0-2 Display graphical dialog boxes fro -- 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#551901: ksh: setting EDITOR env var causes input prob for an alias w/ here file
> It seems to be fixed in the 2010 version, e.g. > It looks like you can close this bug. Thanks for checking. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#551901: ksh: setting EDITOR env var causes input prob for an alias w/ here file
On 20 Oct, Jeff Green wrote: > $ cati > > So are you saying that on that line, no key other than newline does anything? > The blank input line for the 2nd cati invocation basically is > ignoring all input except a newline. Once the newline is entered, > the shell recognizes input again. I've not been able to reproduce this at all. Does it still occur with the 20100309 release of ksh? Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#528861: ksh_93t+-1(mips/unstable): FTBFS on mips. libast not built ?
I can't reproduce this problem using qemu to run a mips system on an amd64 system. So I really need some help to pin down the cause. The build continues to fail for mips and mipsel with newer upstream releases of ksh. Judging from subtleties in the build logs, I think you have dash and not bash as /bin/sh on the build machines so I've repeated the build (under qemu) with that. It's not something else like busybox is it? Is there anything else slightly unusual in the build machine's setup which I might be able to replicate in qemu? The first differences in the build logs that I see are the following: /build/buildd-ksh_93t+20100301-1-mips-sjFIE0/ksh-93t+20100301/src/lib/libast/string/strerror.c: In function '_ast_strerror': /build/buildd-ksh_93t+20100301-1-mips-sjFIE0/ksh-93t+20100301/src/lib/libast/string/strerror.c:78: error: 'sys_nerr' undeclared (first use in this function) /build/buildd-ksh_93t+20100301-1-mips-sjFIE0/ksh-93t+20100301/src/lib/libast/string/strerror.c:78: error: (Each undeclared identifier is reported only once /build/buildd-ksh_93t+20100301-1-mips-sjFIE0/ksh-93t+20100301/src/lib/libast/string/strerror.c:78: error: for each function it appears in.) /build/buildd-ksh_93t+20100301-1-mips-sjFIE0/ksh-93t+20100301/src/lib/libast/string/strerror.c:79: error: 'sys_errlist' undeclared (first use in this function) mamake [lib/libast]: *** exit code 1 making strerror.o The preceding configure style checks have correct results. So any idea why these errors are occurring? Thank you Oliver Kiddle -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#528861: ksh_93t+-1(mips/unstable): FTBFS on mips. libast not built ?
On 16 May, you wrote: > Package: ksh > Version: 93t+-1 > Severity: serious > > There was an error while trying to autobuild your package: I can't reproduce this on armel or amd64. Is there anything unusual about your build machine. In particular, what shell in /bin/sh? Not working with pdksh derived shells is a known problem. Particularly suspicious in the build log is: sh: gcc: command not found failed to open configuration file `/nonexistent/.dpkg.cfg' for reading: Permission denied Any idea why your bug seems to have stopped any build attempts on other platforms. This is irritating. Oliver -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#478019: zsh: Should handle non-breaking space as word separator
Samuel Thibault wrote: > Clint Adams, le Sat 26 Apr 2008 16:05:48 +0100, a écrit : > > Having locale-based (and multibyte) word separators sounds like a nightmare > > to me, but maybe someone has some ideas. > > iswspace() The characters used for word splitting are contained in the IFS variable. So you can add character 0x82 to that and in a suitable locale, that will work. In a UTF-8 locale, you might have problems; I don't know whether multi-byte chars are supported in IFS. For what it's worth, I don't think it is right for non-breaking spaces to be in IFS by default. I've seen people use them precisely because they are not. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#469235: ksh: printf %s in UTF-8 is not POSIX-compliant
Vincent Lefevre wrote: > As you can see, the ksh93 printf builtin doesn't behave like the > coreutils printf, and this is ksh93 which is wrong. Indeed, the > field width and the precision are number of bytes, not number of > characters. Thanks for the report but in this case, I really think the standard writers have got it wrong. Either that or they wrote this before multibyte locales were commonplace. This is a shell and treating the argments as characters is far more useful. In the rare case where someone really wanted this behaviour, they could use the C locale. Feel free to raise the issue on ast-users if you feel strongly. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#451382: i18n is NOT so easy!
Clint Adams wrote: > On Thu, Dec 06, 2007 at 05:56:12PM +0100, Dr. Markus Waldeck wrote: > > Please correct _cut as mentioned in my mail from Fri, 16 Nov 2007 10:16:44 > > +0100. > > Okay, but it is certainly ugly. You can use $'\u00e4' which wouldn't be quite so ugly. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#452701: Time to upgrade
You wrote: > > Package: ksh > Version: 93s+20070628-1 > Severity: wishlist > http://www.research.att.com/~gsf/download/release.2007-11-05.2007-03-28.html No, you may not have it yet but 20071105 has already been uploaded to sid. Would be nice if you could upgrade to it and verify that bug 304365 is still an issue. I still can't work out how to reproduce it. Are you sure you haven't got something in a profile file. Could you have another look at the reply I sent on 13/4/2005. Thanks Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#433026: ksh: setting & unsetting emacs mode causes segfault
On 13 Jul, Diab Jerius wrote: > Setting and unsetting emacs mode in a particular manner causes a > segmentation fault. My .profile file contains the single line > > set -o emacs I can't reproduce this but I don't have access to an amd64 system. So I'll need your help to resolve it. Do you have any other setup files, i.e. is the ENV variable set? Does this still occur if you run ksh as ksh -f. Is ksh definitely linked to ksh93 through the alternates system? Is the problem still there with the latest 93s+ version? The ideal would be if you could apt-get source the latest ksh-93s+20070628 version from unstable, build it from source with debug enabled and get a core file from the segmentation fault. Let me know if you need any help with that. Thanks Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#431206: ksh s+ 2007-06-28 released, please update
On 30 Jun, "Josh Hurst" wrote: > 06-10-15 The remaining math functions from C99 were added for any system > that supports them. > Please compile the package with -std=gnu99 to get this math support Are you sure that -std=gnu99 is necessary? There are definitely new math functions available with the new build without using this option. I think it is just finding the functions in libc and using them regardless of compiler support for C99. If you believe it is necessary, can you perhaps please name a specific math function so that I can verify that it makes a difference. Thanks Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#430472: Fwd: Bug#430472: zsh litters /usr with files which it does not claim ownership of
"Adam Megacz" wrote: > You should include a dummy file in the archive for each of these. The > dummy file can be empty (ie zero bytes long) if necessary. Then make > sure that the process which creates them during postinst doesn't > complain about the fact that they're already there. Is that really the only way? Surely, that would mean that the checksum for the files would be incorrect so running debsums would flag the files invalid. I'd have thought it'd be easier to simply generate the files when the package was built instead of when it is installed. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#304365: Fixing this bug
--- Josh Hurst <[EMAIL PROTECTED]> wrote: > This bug can be fixed by updating to ksh s+, add -DSHOPT_SYSRC to the > build flags and add a file called /etc/ksh.kshrc with the content > cut cut cut > set -o gmacs Is this really the problem in this case? The up arrows and tab key work fine for me even after `set -o vi'. ksh defaults to vi mode editing and I think it would be wrong to change that default whether I think it is an appropriate default or not. Juhapekka: can you perhaps verify this bug again please? Enabling SHOPT_SYSRC is something I don't mind doing if you want. ksh s+ is in unstable by the way. Oliver _ Inbox cluttering up with junk? Clean up with Yahoo! Mail. http://uk.docs.yahoo.com/mail/isp_targeting.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#182445: ITP: xmlroff -- XSL Formatter
--- "W. Borgert" <[EMAIL PROTECTED]> wrote: > On Wed, Apr 18, 2007 at 07:51:31PM +0200, Daniel Leidert wrote: > > What is the current state of the ITP of xmlroff, Oliver? What is As far as I was concerned, I had handed it to Wolfgang. > There was a long time of no visible activity upstream, but there > seems to be new movement lately. The current state in itself is > IMHO not really useful. So packaging should only be done in > anticipation that xmlroff will become a useful FO processor. It > certainly has the potential. Nevertheless, real life limits my > availibility for Debian too much for this task, ATM. Given that you (Wolfgang) don't seem interested in packaging it at the moment, I'll take another look at it once 0.5 is released. 0.5 is a significant reorganisation so will affect the Debian packaging so packaging 0.4 probably doesn't make sense. What you say about xmlroff's current state is true. At the moment, packaging it is likely to help xmlroff more than Debian but xmlroff could do with some help. Oliver ___ Yahoo! Answers - Got a question? Someone out there knows the answer. Try it now. http://uk.answers.yahoo.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#394577: zsh-beta: Completion of date does not work
On 14 Apr, Clint wrote: > On Sun, Oct 22, 2006 at 12:22:37AM +0200, Jörg Sommer wrote: > > % date --ver > > --version-- Versionsinformation anzeigen und beenden > > % date --ver > > > > but the text is not completed. > > This improves things slightly. This is what _guard is for. Use something like: ': :_guard "^--*" format or date' Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#379132: ksh93 should be installed as /bin/ksh
On 21 Jul, "Josh Hurst" wrote: > RFE: ksh93 should be installed as /bin/ksh by default to match SuSE, Can I close this bug? As far as I'm concerned, this is "default" because ksh93 is the highest priority package providing /bin/ksh via the alternates system (see my earlier reply). Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384797: ksh stops issuing error messages for bad commands
--- Norman Ramsey <[EMAIL PROTECTED]> wrote: > David Korn reports that this bug is fixed upstream in the 'ksh 93r+' > version on the web. A Debian update would be lovely. That's an alpha release though isn't it? There's a few other fixes in that but probably at least as many new bugs. I was planning to wait till 93s. Have you tried the alpha? If this bug is irritating you, I could be persuaded to package it. (I packaged an earlier alpha when there was a FTBFS on amd64.) Oliver ___ To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#182445: Will you package xmlroff?
--- "W. Borgert" <[EMAIL PROTECTED]> wrote: > are you working on the Debian packaging of xmlroff? If not, I > would like to package xmlroff with the XML/SGML group as > maintainer, if the team does not object. Of course, you could > join the team, btw. I was working on it. I had a package before the structure got rearranged with pangoxsl becoming part of the same package. The manpage I wrote has been accepted upstream. I was just about to look at the new release and was going to decline your request but my PC died last night. It's probably just the power supply but I don't know yet. Anyway, if you'd like to package xmlroff, go ahead. Please let me know how you progress and I'll be happy to have a look at any initial package you produce. And if you change your mind, let me know and I'll get a package ready just as soon as my PC is working again. Oliver ___ Try the all-new Yahoo! Mail. "The New Version is radically easier to use" The Wall Street Journal http://uk.docs.yahoo.com/nowyoucan.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#382240: ksh: FTBFS: bashisms ?
--- Julien Danjou <[EMAIL PROTECTED]> wrote: > ksh fails to build, however I guess this is caused by /bin/sh linked > to /bin/dash. > > If you really need bash and don't use only POSIX compliant syntax, > please call /bin/bash instead of sh. Actually, more likely this is dash failing to be POSIX compliant. dash is pdksh based right? Those ksh build scripts are fairly complicated and they work with ksh88/ksh93/bash/zsh(with sh emulation) and the /bin/sh on various Unix systems but not with pdksh. I even had to apply a fix for it to work with bash 3.1. It would be useful if you could attempt to track down which part of the build script is not running correctly under dash. The alternative might be detecting /bin/sh being dash and trying zsh/bash whatever instead. Oliver ___ All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine http://uk.docs.yahoo.com/nowyoucan.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#379132: ksh93 should be installed as /bin/ksh
Josh Hurst wrote: > RFE: ksh93 should be installed as /bin/ksh by default to match SuSE, > Fedora, Darwin, SCO Unix, and AIX5L which all install ksh93 as > /bin/ksh. Under unstable/testing, /bin/ksh is a symbolic link that is handled by the alternates system. ksh93 is the highest priority package providing /bin/ksh. I think this is the best way to handle /bin/ksh because it gives the user control. zsh provides half reasonable ksh emulation and it wouldn't be unreasonable for a zsh user to choose to use zsh as /bin/ksh to allow ksh scripts to run without installing extra programs. > (Yes, there was pdksh long ago - but that project is dead, the last > major update was released in 2001 and since then nothing happened) According to the popularity contest pdksh still has many more installations than ksh93. In addition a pdksh fork named mksh exists that has seen some development. Any idea how we might convince more people to swap pdksh for ast-ksh? Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#368903: zsh: completion for svn diff
On 25 May, Joerg Sommer wrote: > > can you add completion support for svn diff? If I understand it right, it > should be the same rule like commit -- all files they are known by svn > are able to be used with diff. The patch below should fix this and bug 342348. The diff subcommand can also take URLs so it isn't quite the same as for commit. I've made the URL completion a bit cleverer using svn list too. Clint: I'd appreciate if you can get this into upstream CVS as I'm currently without access to a net connection that doesn't block ssh. Oliver --- _subversion.origThu Jun 1 10:53:09 2006 +++ _subversion Thu Jun 1 18:03:19 2006 @@ -49,6 +49,11 @@ '*:file:_files -g ".svn(/e:_svn_deletedfiles:)"' ) ;; + diff) +args+=( + '*: : _alternative "files:file:_files -g \*\(e:_svn_status:\)" "urls:URL:_svn_urls"' + ) + ;; help) args+=( '*::sub command:_svn_commands' @@ -60,6 +65,11 @@ '*:file:_files -g "*(e:_svn_controlled:)"' ) ;; + resolved) +args+=( + '*:file:_files -g "*(e:_svn_conflicts:)"' +) + ;; revert) args+=( '*:file:_files -g "(.svn|*)(/e:_svn_deletedfiles:,e:_svn_status:)"' @@ -141,6 +151,11 @@ [[ -f ${(M)REPLY##*/}.svn/text-base/${REPLY##*/}.svn-base ]] } +(( $+functions[_svn_conflicts] )) || +_svn_conflicts() { + [ -n $REPLY.(mine|r<->)(N[1]) ] +} + (( $+functions[_svn_deletedfiles] )) || _svn_deletedfiles() { # Typical usage would be _files -g '.svn(/e:_svn_deletedfiles:)' @@ -167,15 +182,32 @@ (( $+functions[_svn_urls] )) || _svn_urls() { - local expl + local expl remfiles remdispf remdispd suf ret=1 - if [[ -prefix *: ]]; then -_urls + if [[ -prefix *: ]] && ! _urls && + zstyle -T ":completion:${curcontext}:" remote-access + then +remfiles=( ${(f)"$(svn list $IPRFIX${PREFIX%%[^./][^/]#} 2>/dev/null)"} ) +compset -P '*/' +compset -S '/*' || suf=file +remdispf=(${remfiles:#*/}) +remdispd=(${(M)remfiles:#*/}) +_tags files +while _tags; do + while _next_label files expl ${suf:-directory}; do +[[ -n $suf ]] && compadd "$@" "[EMAIL PROTECTED]" -d remdispf $remdispf && ret=0 +compadd ${suf:+-S/} "$@" "[EMAIL PROTECTED]" -d remdispd \ +${remdispd%/} && ret=0 + done + (( ret )) || return 0 +done else compset -S '[^:]*' _wanted url-schemas expl 'URL schema' compadd -S '' - \ -file:// http:// https:// svn:// svn+ssh:// +file:// http:// https:// svn:// svn+ssh:// && ret=0 fi + + return ret } (( $+functions[_svn_commands] )) || -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#358948: FTBFS on AMD64: md5 file doesn't exist
--- Martin Michlmayr <[EMAIL PROTECTED]> wrote: > This package fails to build on AMD64. I cannot quite figure out why, > but it seems the md5 file is never generated. I notice that, looking at the build logs, there is a successful build later on. See: http://buildd.debian.org/build.php?&pkg=ksh&ver=93r-1&arch=amd64&file=log Any idea what was changed for that to work? What version of bash are you using as /bin/sh? Recent versions had a bug in the parsing of backquotes. The fix may have broken the workaround I put in. Using ksh or zsh as /bin/sh might help. Unfortunately, I don't have access to an amd64 machine. Oliver ___ Win a BlackBerry device from O2 with Yahoo!. Enter now. http://www.yahoo.co.uk/blackberry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#353685: spamprobe 1.4b released
Nicolas Duboc wrote: > I have actually packaged this 1.4b version this week-end. I'm currently > testing it on my own configuration. I planned to upload tomorrow if I > find no problem. Excellent. That's good to hear. Thanks Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#353685: spamprobe 1.4b released
Package: spamprobe Version: 1.2a-1 Severity: wishlist http://prdownloads.sourceforge.net/spamprobe/spamprobe-1.4b.tar.gz There's been a number of new releases since the debian package was last updated. Would be nice to have an update to the package. Oliver Kiddle This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#352822: ITP: pangoxsl -- additional XSL atttributes for Pango
Package: wnpp Severity: wishlist * Package name: pangoxsl Version : 1.6.0.1 Upstream Author : Tony Graham <[EMAIL PROTECTED]> * URL : http://sourceforge.net/projects/pangopdf/ * License : LGPL Description : additional XSL atttributes for Pango PangoXSL defines attributes for several XSL inline properties that are not currently implemented by Pango. This is primarily useful for xmlroff (http://xmlroff.org/) which I'm also considering packaging. For pangoxsl, I have already prepared an initial package (producing libpangoxsl-1.0-0 and libpangoxsl-dev packages). A new minor upstream release is coming soon so I'm waiting for that. xmlroff needs other work like writing a manpage to be done first. This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#279812: O: sfio -- Enhanced library for managing I/O streams.
If there is any interest in sfio, let me know. It would be best handled as part of the same package as ksh. Ksh (also from AT&T) uses sfio and the source is largely distributed as a single tarball. Being a new maintainer, I wanted to keep things simple by just packaging ksh but all the AT&T utilities (such as nmake etc) could be built from one source package. Oliver Kiddle This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#340611: flowscan depends on pdksh but /bin/ksh may not be pdksh
Package: flowscan Version: 1.006-8 Severity: normal This package depends on pdksh and as far as I can tell, uses /bin/ksh. However, since version 5.2.14-18 of pdksh, pdksh is installed as /bin/pdksh and the alternates system used to handle /bin/ksh. Upstream suggests that either real ksh or pdksh can be used. I'd be very suprised if either of mksh or zsh (emulating ksh as it does when linked from /bin/ksh) didn't work. I'd therefore suggest that the dependency be changed to ksh | mksh | pdksh | zsh. An alternative is to ensure that a more specific path such as /bin/ksh93 is assigned to SHELL in the Makefile. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#340610: kimberlite: depends on pdksh but /bin/ksh may not be pdksh
Package: kimberlite Version: 1.1.0-3.3 Severity: normal kimberlite depends on pdksh and includes ksh scripts in /usr/lib/kimberlite that rely on #!/bin/ksh to run the scripts. However, since version 5.2.14-18 of pdksh, pdksh is installed as /bin/pdksh and the alternates system used to handle /bin/ksh. This means that installing kimberlite can add pdksh as a dependency but a different ksh is then used. One solution would therefore be to change the dependency to ksh | mksh | pdksh | zsh. Alternatively, the scripts should be changed to use #!/bin/pdksh. The advantage of changing the dependency is that it reduces the chance that someone needs to install an extra package. The problem is that the scripts may break in one of the other shells. Looking through the scripts closely, I noticed one problem. Function definitions are of this form: stopDevices() { typeset TMPFILE=/tmp/$MYNAME.stopDevices.$$ ... } In the real Korn shell, the two function declaration forms have different semantics and variable definitions will not have local scope. The fix is to use the syntax: function stopDevices { } This should really be changed upstream: other distributions have changed their ksh to the real ksh. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#228903: nis: yppush fails on master which isn't also a client
Mark Brown wrote: > > > The solution was to temporarily run ypbind. It seems that yppush trys > > to find out the list of slaves from ypbind. As this is a master and is > > not a NIS client itself, this fails. It should be able to just look in > > /var/yp/ypservers. > > Is there some reason why you actually need to disable ypbind? I don't as such need to disable it. Enabling it is an effective workaround for the bug but it remains a bug. Following the howto document, it is quite logical to setup NISCLIENT=false on a server. The workaround isn't necessarily obvious so it could cause other people problems. The fix would be to make yppush access /var/yp/ypservers directly if the machine is a master for the domain. yppush's man page states: It is normally run only on the NIS master by /var/yp/Makefile after the master databases are changed. That implies that someone might want to run yppush on some machine other than a master so it probably still needs to be able to use NIS to get the list of slaves. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#321117: Default umask seems dangerous
On 30 Sep, Clint Adams wrote: > > I'll put "umask 022" in /etc/skel/.zshrc, I suppose. It would be better to put that in one of zprofile or zlogin. It is exported to child processes and it could be a nuisance (or even a security risk) if child shells override the setting. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#330883: zsh: For svn revert, _subversion completes deleted files incorrectly
You wrote: > The subversion case is a bit special. Even if the removed item was a > directory, the suffix should be a space, not a slash, because AFAIK, > doing "svn revert" on an item inside a removed directory has not much > sense. We could try passing -S ' ' to _files. Perhaps try adding that in yourself: find this bit and add -S " " -r '/\-' after _files. revert) args+=( '*:file:_files -g "(.svn|*)(/e:_svn_deletedfiles:,e:_svn_status:)"' ) ;; This is probably also relevant to the delete subcommand. What this will break is completing a slash after real existing directories. But reverting directories is perhaps more common than reverting a file from a subdirectory. The more common use of fake files is with things like Cygwin's cygdrive where the fake files are directories. I don't have access to cygwin but that may work due to the fact that they will exist when stat is run. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#330883: zsh: For svn revert, _subversion completes deleted files incorrectly
You wrote: > For "svn revert", when I complete a deleted file, the space isn't > added after the filename. There's no such problem with modified > files (marked with 'M') for instance. This isn't actually an _subversion issue but a more fundamental file completion issue. Filename completion adds a slash for directories and a space for other files. For files that don't actually exist it adds no suffix because it is unknown whether the file is a normal file or a directory. This occurs for the fake-files style as well as files added using the e glob modifier (which is what _subversion does). Communicating file type information with fake files would be a very difficult thing to do. Perhaps we should make it default to a removable space suffix for non-existing files. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#327224: nis.config fails if hostname not configured
Package: nis Version: 3.14-2 Severity: normal On a newly installed machine, I had not yet configured a hostname properly. "hostname --fqdn" returns status 1 and an error message. This results in installation of nis failing as follows: Setting up nis (3.14-2) ... hostname: Host name lookup failure dpkg: error processing nis (--configure): subprocess post-installation script returned error exit status 1 The nis.config script should be adjusted to cope with this situation. It should allow for "hostname --fqdn" to fail. fqdn=`hostname --fqdn 2>/dev/null || hostname` would do the job. hostname alone returns "debian" and no error. Thanks Oliver Kiddle This message has been scanned for viruses by MailControl - www.mailcontrol.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#322731: ksh93 echo builtin does not respect (suppress trailing newline)
You wrote: > It appears that ksh93 is not respecting the \c (suppress trailing > newline) in an echo statement - for example: > > #!/bin/ksh > echo "enter a number => \c" > output is: > enter a number => \c This is by design. See the man page. In particular: echo [ arg ... ] When the first arg does not begin with a -, and none of the arguments contain a \, then echo prints each of its arguments separated by a space and terminated by a new-line. Otherwise, the behavior of echo is system dependent and print or printf described below should be used. See echo(1) for usage and description. echo has always varied between systems. It is an old System V/BSD difference. If you have access to Solaris, look at the difference between /usr/ucb/echo and /usr/bin/echo. This is why the print command was introduced. printf is also a good option if you want consistent behaviour. > Note that "echo -n "enter a number => " DOES work, as does coreutils > echo. As far as I can tell, coreutils behaves the same as ksh: $ /bin/echo "enter a number => \c" enter a number => \c Perhaps it is also system dependant. bash's echo also behaves like this by default. zsh defaults to the opposite. But both bash and zsh have an option. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#322573: zsh: subversion completion breaks on svn ci
You wrote: > This version of zsh breaks subversion completion for svn ci: > > % svn ci > _svn:42: unknown file attribute Oops. That's my fault. Sorry about that. Fix is below: Index: _subversion === RCS file: /cvsroot/zsh/zsh/Completion/Unix/Command/_subversion,v retrieving revision 1.14 diff -u -r1.14 _subversion --- _subversion 8 Aug 2005 08:54:21 - 1.14 +++ _subversion 11 Aug 2005 14:38:08 - @@ -40,7 +40,7 @@ ) ;; commit) -args-+=( +args+=( '*:file:_files -g "*(e:_svn_status:)"' ) ;; This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#314578: ksh: FTBFS (amd64/gcc-4.0): static declaration of '_sfds
On 17 Jun, you wrote: > Package: ksh > Version: 93q-3 > Severity: normal > Tags: patch > > When building 'ksh' on amd64/unstable with gcc-4.0, > I get the following error: This has apparently been fixed upstream. Without access to the correct fix, I'll wait for the release of 93r. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#314578: ksh: FTBFS (amd64/gcc-4.0): static declaration of '_sfds
On 17 Jun, you wrote: > > With the attached patch 'ksh' can be compiled > on amd64 using gcc-4.0. I'm a little reluctant to simply apply this patch without understanding what's going on better. Is this a gcc-4.0 or amd64 issue, or something that really is specific to both. What affect does the patch actually have? Could it break compilation elsewhere? Would it perhaps be better to post to the ast-users list (i.e. upstream) with a clearer explanation? Thanks Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#310872: zsh can't be a ksh replacement (can't trap ERR).
Bart wrote: > (Google indicates that ksh does run > on UNICOS; I wonder what it does with ERR in that case.) I tried asking on ast-users and received the following answer from Glenn Fowler: | ksh does not currently handle this | but I talked with David Korn and the next release will do the following | | the "SIG" prefix is optional for trap and kill, e.g., "HUP" == "SIGHUP" | the change will be to do only a signal lookup when a "SIG" prefix is specified, | and a shell trap name lookup followed by a signal lookup otherwise | | so, on systems like unicos that have SIGERR, the trap name "SIGERR" refers | to the signal, and "ERR" refers to the shell trap on error | "SIGERR" on systems with no SIGERR signal will fail | | this approach will handle all future non-posix signal/shell-trap clashes, | e.g., if some system invents SIGEXIT That would seem a far better solution to me than having a separate ZERR, especially as it deals with future conflicts. Would it be hard for us to do that same in zsh? We can always keep ZERR for backward compatibility. Oliver This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#310872: zsh can't be a ksh replacement (can't trap ERR).
> IMHO, this can only lead to confusion. if one installs zsh as ksh, he > might use commands that are not supported by "historical" ksh, > and "historical" scripts may not run. Quite frankly there is no shell in Debian which perfectly emulates "historical" ksh. The newer versions of AT&T ksh could break things in much subtler ways than this (such as the change from dynamic to static variable scoping) (see /usr/share/doc/ksh/COMPATIBILITY.gz). pdksh is also little better. I'd also bet that tcsh doesn't do a perfect job when it handles the csh alternative but it works for me. Given that zsh adjusts it's behaviour based on the name it was invoked under, it is useful for user's to be able to take advantage of this and have the ksh symlink point to it. Why install a ksh package if you use zsh and zsh can run all your ksh scripts? And given the low priority of the alternative, users aren't going to do this without some idea of what they're doing. Certainly, it would make sense to enable ERR for ksh compatibility mode. Or perhaps it'd be better to insist on "SIGERR" on UNICOS. Oliver __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
[EMAIL PROTECTED] wrote: > > HOTPLUG_RC_$SUBSYSTEM > If set to "no", disables coldplugging for the $SUBSYSTEM. Thanks but this isn't really the point of why I raised a wishlist item in the bug tracker. I had already done "update-rc.d -f hotplug remove" which presumably has the same effect. As the udev web page clearly states, there is no fundamental dependency on the hotplug scripts so it seems a reasonable wish for the Debian package to not have a dependency. If I want udev but not hotplug scripts, I shouldn't have to install the hotplug scripts and then read their documentation to work out how to disable them. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
[EMAIL PROTECTED] wrote: > > Again, why can't you stop wasting everybody's time instead of RTFM like > I suggested in my first reply? I have RTFM and I don't see how it helps. The readme file you pointed to doesn't even mention udev. I'll quote the udev web page: It is recommended that you also have the Linux Hotplug scripts installed, but it is not necessary for it to work properly. (http://www.kernel.org/pub/linux/utils/kernel/hotplug/udev.html) Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
Greg KH wrote: > > See http://www.gentoo.org/doc/en/udev-guide.xml > > hotplug-base is NOT the hotplug/coldplug startup stuff. You want that > file. Why would you not? That's what I was assuming. The trouble is that in Debian there is no "hotplug-base". udev depends on the "hotplug" package and that contains all the scripts I'm complaining about. Now if udev can work without those scripts as the Gentoo howto implies then it'd be nice to be able to have such a setup on Debian. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
[EMAIL PROTECTED] wrote: > > > It is also annoying because hotplug depends on usbutils and my machine > > doesn't even have USB. > It does not: > > Depends: module-init-tools | modutils (>= 2.4.2), debconf (>= 0.2.26), > procps, grep (>= 2.5.1.ds1-2), sed (>= 4), bash (>= 2.05b-1) Sorry, I was looking at "apt-cache rdepends usbutils" which does list hotplug. Perhaps that's not the same. aptitude did install it without my asking: perhaps it does that for recommends. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
Greg KH wrote: > > Reading various FAQs etc on the web, it seems clear that udev can work > > fine without hotplug. > > Huh? That is not true. Please point me at those faqs. The kernel > _MUST_ support hotplug for udev to work properly. I'm talking about the hotplug scripts not CONFIG_HOTPLUG in the linux kernel. So in particular, I meant the Debian hotplug package. Sorry for not being clear. See http://www.gentoo.org/doc/en/udev-guide.xml Quoting from that: udev will install hotplug-base as one of it's dependencies. You do not need to install hotplug unless you want your modules automatically loaded when you plug devices in. hotplug also handles the automated bringup of network devices and firmware downloading. That's not the only FAQ I've seen. If you want, I can dig out more. It is also annoying because hotplug depends on usbutils and my machine doesn't even have USB. Thanks Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307781: udev: why is hotplug dependency necessary
Package: udev Version: 0.056-2 Severity: wishlist Reading various FAQs etc on the web, it seems clear that udev can work fine without hotplug. It also seems that other distributions allow udev to be used without hotplug. I really don't want to use hotplug: it takes ages at bootup, loads modules for hardware I very rarely use and doesn't do anything useful (any modules I want loaded at boot are compiled in). I'd very much appreciate it if the udev package only recommended hotplug and made it easy to use it without hotplug. Thanks Oliver -- System Information: Debian Release: 3.1 APT prefers testing APT policy: (650, 'testing') Architecture: i386 (i686) Kernel: Linux 2.6.11.8 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Versions of packages udev depends on: ii hotplug 0.0.20040329-22 Linux Hotplug Scripts ii initscripts 2.86.ds1-1 Standard scripts needed for bootin ii libc62.3.2.ds1-21GNU C Library: Shared libraries an ii makedev 2.3.1-77creates device files in /dev ii sed 4.1.2-8 The GNU sed stream editor -- no debconf information __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#304365: History and Tab completion do not work at all
You wrote: > > Does your terminal simply generate a Ctrl-I for tab? > > It moves my cursor about 8 characters. Ctrl-L generates ^L . Ctrl-D exits > from that shell. UpArrow generates ^[[A . If you run cat -T, then press , does it print ^I. If ^[[A for up arrow doesn't work then it sounds like something else is wrong. > I use terminaltype xterm-16color in my xterms. It does not help, if I > start ksh like this: Is this normal xterm and not something else like aterm, pterm, rxvt etc? > > You're only going to get filename completion in ksh if it is > > unambiguous. No nice zsh like lists of matches. Try <=> to list > > possible matches. > > That does not work. Have you modified /etc/profile? Is the ENV environment variable set to anything? Can you try the following: % ksh93 -p $ mkdir tmpdir $ cd tmpdir $ :> foo $ cat f > Yeah. You may need to add something like this to /etc/profile : > > case $SHELL in > *ksh*) Given that /etc/profile is only run for login shells, that wouldn't especially help. At the moment, there isn't anything I'd want to put in such a system startup file anyway. The pdksh package has managed well enough without one. Oliver __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#304365: History and Tab completion do not work at all
You wrote: > Package: ksh > Version: 93q-1 > Severity: normal > > If I hit ArrowUp -key in ksh93, it do not bring the last command from > history. If I hit or , it do not work as Tab > completion at all. They work for me. Does your terminal simply generate a Ctrl-I for tab? You're only going to get filename completion in ksh if it is unambiguous. No nice zsh like lists of matches. Try <=> to list possible matches. > You may also need some good examples. You'll find some of them from > dotfiles.com , for example this: > > http://www.dotfiles.com/files/5/197_.kshrc Does the keyboard trap in that example fix completion for you? I plan to add an example .kshrc in /usr/share/doc and I can include a keyboard trap there. As ksh doesn't support a system-wise kshrc file it is harder to force any fix on users. Oliver __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#304357: wishlist: use update-alternatives for ksh-symlink
Juhapekka Tolvanen wrote: > > Package: zsh > Version: 4.2.5-1 > Severity: wishlist > > Packager of pdksh realised that and made Or to be more precise, Clint raised a bug on pdksh. > Hence, I propose also zsh will use update-alternatives for managing > that ksh-symlink. I also propose that a package called ksh will do so, > too. Thanks for the suggestion. Everything's already under control. Clint uploaded a 4.2.5-2 yesterday which apparently does this. I'll be doing a new ksh package once I've got my hands on the the updated pdksh. Thanks Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#202667: nmh: rcvstore segfault when a certain file is in ~/Mail/inbox
This patch should fix it. Oliver Index: sbr/folder_read.c === RCS file: /cvsroot/nmh/nmh/sbr/folder_read.c,v retrieving revision 1.2 diff -u -r1.2 folder_read.c --- sbr/folder_read.c 2 Jul 2002 22:09:14 - 1.2 +++ sbr/folder_read.c 23 Feb 2005 13:52:38 - @@ -70,7 +70,7 @@ adios (NULL, "unable to allocate storage"); while ((dp = readdir (dd))) { - if ((msgnum = m_atoi (dp->d_name))) { + if ((msgnum = m_atoi (dp->d_name)) && msgnum > 0) { /* * Check if we need to allocate more * temporary elements for message names. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#294563: ITP: ksh -- The real AT&T Korn shell
Package: wnpp Severity: wishlist * Package name: ksh Version : 93q * URL : http://www.kornshell.com/ * License : Common Public License (CPL) Version 1.0 Description : The real AT&T Korn shell The latest version of AT&T Software Technology's (David Korn's) ksh has changed licence and now goes under the CPL (previous licence had a few issues). The KornShell is a UNIX command interpreter (shell). This version is a much newer version of the ksh distributed with common commericial Unix variants (ksh88) and also newer than the original 1993 version. It primarily adds powerful features for scripting such as compound variables, static scoping, associative arrays and namespaces. The licence change also applies to the other software in the ast-base distribution if there is interest in them. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#289664: zsh: completion for gnome-gv
You wrote: > > Shouldn't that be in Completion/X/Command? > > I don't know. I was thinking that C/X/C was for commands with X-style > arguments, but I guess that doesn't make sense. The original idea was that people could exclude directories they weren't interested in and X is anything which requires an X server to run. Certainly many of the other things in there (gqview, dcop, mplayer) don't take "X style arguments". Just on the basis that we have far more files in the Unix directory, I'd be tempted to say it should go in X. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#289748: zsh: menu navigation is suboptimal
You wrote: > The effect of this is that it can make it very hard to select the last > item in a long multi-column menu of completions. For date-based > organization of data in a directory, this can be particularly > frustrating. It is possible to get straight to the last item from the first by using the reverse-menu-complete editor command. I bind that to shift-tab: bindkey '\e[Z' reverse-menu-complete For the behaviour you want, with the up and down keys, you can try the following bindings: bindkey -M menuselect '\e[B' menu-complete bindkey -M menuselect '\e[A' reverse-menu-complete Note that your terminal may generate different escape sequences from those keys. Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#289664: zsh: completion for gnome-gv
Clint Adams wrote: > This should do the trick. > > Index: Completion/Unix/Command/_gnome-gv Shouldn't that be in Completion/X/Command? (I'm not familiar with gnome-gv but am assuming it is a GUI postscript viewer). Oliver -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]