Re: fvwm and post.hook problem
Cameron Hutchison wrote: > Unfortunately the new DebianMenu file does not read > main-menu-pre.hook, so it is not possible to put your own stuff at > the start of the menu, so my customised menu has my stuff after the > standard debian stuff now. This I had to do in menudefs.hook where > I used to do it in main-menu-pre.hook. The new approach is disappointing, but it looks like everything but the missing main-menu-pre.hook read can be worked around, so I filed a bug asking for its reinstatement. http://bugs.debian.org/243985 You may want to add onto this bug if you think more is called for. As an immediate workaround I, like you, copied cp /etc/X11/fvwm/system.fvwm2rc.dpkg-old ~/.fvwm/.fvwm2rc Then, to get the main-menu-pre.hook read back in, I edited /etc/menu-methods/fvwm, adding a few lines back in: -- 8< -- --- fvwm.orig 2004-04-08 14:48:28.0 -0400 +++ fvwm2004-04-15 19:30:27.0 -0400 @@ -35,6 +35,9 @@ # 3. We add a trailing newline to the lot. startmenu= "DestroyMenu \"" replacewith($section," ","_") "\"\n" \ + ifeq($section, "/Debian", \ + "AddToMenu \"/Debian\"\n" \ + "Read main-menu-pre.hook Quiet\n") \ "AddToMenu \"" replacewith($section," ","_") "\" \"" \ title() "\" Title Top\n" endmenu= "\n" -- >8 -- and then ran update-menus. Remember that the menus are automatically generated. Kirk Hilliard -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Looking for old Debian Logo Candidate (Ant)
I am trying to find one of the candidates from the Debian Logo competition of last year. It was an ant, was known as jeannette-0, and used to be available at: http://contest.gimp.org/view.cgi?month=1999-02&mode=show&graphic=jeannette-0 I am particularly looking for the ant head design, not the crawling ant. (I forget which was proposed as official and which as unofficial.) Did anyone save this? Kirk
Authenticating SMTP Server
I would like to configure my SMTP server to relay mail for a few friends who dial up and are dynamically assigned IPs. I certainly don't want an open relay, but I don't know what other options I have with Debian. A web search reveals several authenticating schemes -- some password protected, one particularly elegant one accepting outgoing mail from any IP from which the user has successfully checked POP3 mail in the last several hours -- but they seem to be for commercial servers. Neither Mail-Administrator-HOWTO nor Mail-HOWTO mention authentication. What Debian options are there? Kirk Hilliard
Re: OpenSSH <--> SSH2 interoperability
Problem solved. Brian Moore provided this concise explanation via email: -- 8< -- > I thought that various versions of ssh played well together even with > OpenSSH's default configuration, but the name "SSH2 compatibility mode" > suggests otherwise. They don't. ssh2 servers will, when receiving a connect from an ssh1 client, invoke an ssh1 server to handle it assuming one is installed. ssh2 clients will, when connecting to an ssh1 server, invoke an ssh1 client to talk to the server if one is installed. You have an ssh1 server installed and answering, the person trying to connect should install an ssh1 client so that his ssh2 client can fall back to version 1. -- >8 ------ Thanks again, Brian. Kirk Hilliard
OpenSSH <--> SSH2 interoperability
sshd(8) tells of a -V switch -V client_protocol_id SSH2 compatibility mode. When this options is specified sshd assumes the client has sent the given version string and skips the Protocol Version Identification Exchange. but I can't find this further documented either in the man page or in /usr/doc/ssh/. Does anyone here know about this, and what format "client_protocol_id" is supposed to be? All formats I try earn me a "Protocol mismatch." error. My specific situation is that I have someone who tried to connect to my sshd (OpenSSH-1.2.3, from ssh_1.2.3-2) with an ssh version 2.0.9 and was told: Disconnected; protocol version not supported. Has anyone here run experienced this? I thought that various versions of ssh played well together even with OpenSSH's default configuration, but the name "SSH2 compatibility mode" suggests otherwise. Kirk Hilliard
Re: Help Reproduce a Bash Bug
I asked: > Please help us reproduce an elusive bash bug. Ethan Benson <[EMAIL PROTECTED]> replied: > ask them if they are using kernel 2.3.99 (or maybe some earlier 2.3.* > kernel) there has been some notes about this on the kernel > list. apparently the kernel has a bug or `feature' where it returns > the wrong error. it has nothing to do with bash iirc. Bingo! Thanks Ethan. Kirk
Help Reproduce a Bash Bug
Debian users, Please help us reproduce an elusive bash bug. Bug#62154 reports that Bash gives a "Permission denied" error when it should complain "No such file or directory", but neither the Bash maintainer nor I can reproduce it. Please execute the following from the command line: ./thisdoesnotexist Assuming that you do not have a file by such a name, you should see the following error message: bash: ./thisdoesnotexist: No such file or directory If, however, you see this error message: bash: ./thisdoesnotexist: Permission denied (and thisdoesnotexist does not exist) then please run the these commands: ./thisdoesnotexist ls -ld . thisdoesnotexist echo $SHELL echo $BASH_VERSION pwd mount uname -srm and paste the results into a mail message to me. Thank you, Kirk Hilliard
Re: password
Beavis <[EMAIL PROTECTED]> writes: > how do u change passwords? "man -k password" will list the name and a short description of all man pages concerning passwords. At this point, a knowledge of man page section numbers would keep you from reading all 30 lines. (See "man man".) Only 3 of the 30 are from section 1, "Executable programs or shell commands", and the one you want is: passwd (1) - change user password So read "man passwd" and then run it. Kirk Hilliard
Re: END Key in Emacs (only in Xterm)
[ The END key no longer works in "emacs -nw" in an xterm. ] Rodrigo Castro <[EMAIL PROTECTED]> writes: > The problem, in my opinion, comes when it reads xterm.elc and this > file doesn't assign correctly. That's it. The problem is with the xterm.el (and xterm.elc) files of all four flavors of emacs, because they contain the line: (define-key function-key-map "\e[4~" [select]) It looks like xemacs also breaks the HOME key. The files in question are: /usr/share/emacs/19.34/lisp/term/xterm.elc /usr/share/emacs/19.34/lisp/term/xterm.el /usr/share/emacs/20.5/lisp/term/xterm.elc /usr/share/emacs/20.5/lisp/term/xterm.el /usr/lib/xemacs-20.4/lisp/term/xterm.elc /usr/lib/xemacs-20.4/lisp/term/xterm.el.gz /usr/lib/xemacs-21.1.8/lisp/term/xterm.elc /usr/lib/xemacs-21.1.8/lisp/term/xterm.el.gz and they come from the following packages: emacs19 emacs20 emacs20-el xemacs20-support xemacs20-supportel xemacs21-support xemacs21-supportel A makeshift solution is to remove the offending .elc file and either edit the .el file or just removed it as well. (You can byte compile your edited .el file if you wish, but it is pretty small.) I sent a note about this problem to debian-emacsen. Kirk Hilliard
Re: END Key in Emacs (only in Xterm)
Branden Robinson <[EMAIL PROTECTED]> writes: > "infocmp xterm linux" reveals that kend is defined the same for both > terminal types, and the default xterm Xresources file explicitly binds the > End key to ESC [ 4 ~. If xemacs can't handle that I suggest you contact > its package maintainer. infocmp must not be telling the whole story. The END key does not work in "emacs20 -nw" in an xterm with "TERM=xterm", but it does work with "TERM=xterms" and "infocmp xterm xterms" returns no differences at all. Run "emacs20 -nw" from an xterm with TERM=xterm, type "M-x describe-bindings-internal" and 635 lines down (it's easiest to search on "home") you will find this: M-[ 4 ~ select [snip 9 lines] M-[ 1 ~ home Do this again with TERM=xterms and 631 lines down you will find: M-[ 4 ~ end [snip 5 lines] M-[ 1 ~ home The problem exists with both versions of emacs, but emacs19 doesn't have the describe-bindings-internal function. This is with an up to date potato system running xterm 3.3.6-6, emacs20 20.5a-2, and emacs19 19.34-26.3. Kirk Hilliard
Re: 80 MB ram in debian system
[Concerning passing boot parameter with a boot disk.] Ossama Othman <[EMAIL PROTECTED]> writes: > If you are just starting out and don't want to mess around > with configuration files or just want to experiment, you can manually > enter "mem=80M" (don't type the quotes) or whatever amount of RAM you have > at the "boot:" line. You would have to do this each time you boot. That > would be equivalent to 'append="mem=80M"' in your lilo.conf file. LILO > really is the better way to go, as Alex pointed out. How do I get the the "boot:" prompt? I remember seeing it a while back, but now the first thing that I see is "Loading...". Kirk Hilliard -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: less clears screen after exiting
Tomihisa Welsh <[EMAIL PROTECTED]> complained that: > Ever since I installed debian 1.3.1, less will unwantingly clear > the screen when after viewing a document (within an xterm). I can get rid > of this problem with the "-X" option (disables termcap before and after > calling less) however this option also disables my arrowkeys. Any ideas? and G John Lapeyre <[EMAIL PROTECTED]> replied: > > My hamm distribution doesn't do it, in xterm nor rxvt. > I don't know if its my .bashrc thats different or hamm vs. stable. Really? It's not hamm since I am running an up to date unstable system and I have the same behavior that Tomihisa describes. Are you doing anything unusual in your .bashrc? This situation should be figured out and either fixed (so that the arrow keys work with ``less -X'') or a work-around put on the auto-faq. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: X Windows
Adam Heath <[EMAIL PROTECTED]> writes: I wrote: > > If you have a ps2 mouse, run gpm with the -R repeater switch, and > tell X to use > > Protocol"MouseSystems" > Device "/dev/gpmdata" > > in the "Pointer" section of your XF86Config so that you will still > be able to use gpm when you visit the console for old time's sake. and Adam Heath <[EMAIL PROTECTED]> replied: > I don't have to do this. I run unstable (does it make a > difference?), and both 2.0.33 and 2.1.86. When I switch to text > mode, the mouse immediately works. When I go back to X, it > sometimes takes a few seconds to start functioning(the mouse). Strange. I just checked and I still need it on unstable running 2.0.32. (Symptom: mouse does not move at all in X. Hardware: Toshiba Tecra 730CDT laptop.) Do you have a PS/2 mouse? I understood that the problem was that the kernel did not know how to share /dev/psaux between two processes, and that this was on someone's 2.1 todo list, but I never heard if anything had been done yet. When I go back to X, it also takes about a second for the mouse to start functioning, so I don't think that this is a symptom. At first I was skeptical about gpm -R, but I am quite happy with it as it seems to work transparently. Kirk -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: X Windows
Hi Pete, > Does anyone think X windows is something I should have for my linux > system? All I do is program. Would it help me in any way? This depends on your personal taste and on your hardware. For me, my 486-33 with a non-accelerated ISA video board did not have enough performance to justify X for anything that I could also do from the console. I found it very distracting the way that an xterm lagged when I scrolled text full speed in less. It is wonderful that with linux you have the option of using multiple VCs with scroll back and gpm, giving you many, but not all of the advantages you gain with X. I now primarily run a P-150 laptop (so X performance is no longer an issue) with a 12.1 1024x768 screen, and with the "fixed" font (6x13) I can get two 80x56 xterms side-by-side without overlap. This no-overlap criteria was what had me hold out until 1024x768 became available. (Everyone who looks at my laptop remarks on how small the font is, but I don't know if it is the sharpness and high contrast of the LCD display or what, but I find it easier to read than same font on much larger CRTs.) With X, you will be able to set up your xterms with several thousand line scroll-back buffers, and you won't lose the scroll-back data in one xterm when you look at another, like with the VCs. In addition to having multiple xterms (or emacs windows) visible at once, you will still be able to shift between desktops (or between pages on a desktop if that is what you like) with the same ALT-arrow ease that you use to shift between VCs. Personally, I run fvwm2 (as my window manager) and set myself up with one page per desktop, but with a two dimensional layout of desktops (achieved by making ALT-up/down change the desktop number by 1000). I hardly ever have windows overlapping each other, but instead have a couple of dozen windows distributes across a dozen or so desktops. I will sometimes make an emacs window sticky (so that it stays in place as I change desktops) and arrange several xterms in neighboring desktops so that I can switch between them as I cut and paste code from several other .h or .c files or manpages to the emacs window which remains visible throughout the process. Here are a some of mouse related survival tips for X: X expects a three button mouse, and you can emulate the third button on a two button mouse by pressing both buttons simultaneously, but in X, button 2 (traditionally the middle one) is used for more important functions than button 3 -- such as for pasting text. So you might want to swap buttons 2 and 3 by putting this in your .xsessions: xmodmap -e "pointer = 1 3 2" When emacs is run in its own window, text is pasted at the mouse location, not at point. This is likely to drive you batty if you have used gpm a lot, so you can override this by putting this line in your .emacs: (setq mouse-yank-at-point t) You can still move point to the mouse location before pasting (if that is what you want) by pressing mouse-1. Also, use alternate selection in emacs to avoid moving point when selecting text. Do this by holding down Meta (alt) while using the mouse. Because of an evil thing done in the system.fvwm2rc which causes fvwm2 to intercept Meta-mouse events (see Bug #17668), you may need to put this line in your ~/.fvwm2/post.hook: Mouse 0 W M - If you have a ps2 mouse, run gpm with the -R repeater switch, and tell X to use Protocol"MouseSystems" Device "/dev/gpmdata" in the "Pointer" section of your XF86Config so that you will still be able to use gpm when you visit the console for old time's sake. Enjoy, Kirk (a programmer who was quite happy with his VCs, but who loves the added power of X) P.S. Did I mention that you will be able to run ediff-merge with the two source files side by side. It makes comparisons much easier. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: how to update
Martin Pelikan <[EMAIL PROTECTED]> aksed: > > what is the fastest way to update from libc5 to libc6? and Jens Ritter <[EMAIL PROTECTED]> responded: > This script can: > http://taz.net.au/autoup/ > > It downloads all required packages and installs them in the right order. Craig Sanders' autoup script is the way to go, but first make sure that you read Scott Ellis' Mini-HOWTO at: http://www.debian.org/doc/libc5-libc6-Mini-HOWTO/libc5-libc6-Mini-HOWTO.html also available as a link from: http://www.debian.org/developers_corner.html You should do this not only so that you know what process is being automated, but also because it mentions some of the ramifications of the upgrade which aren't discussed in the script. The autoup script does not yet have its own stand-alone documentation. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: cdwrite and cd writers
Hi Stefan, > cdrecord (including 1.6a9) is using illegal code, which makes it impossible > to compile on all versions of linux. > > small example, found in lib/format.c > > > va_list rargs; > > rargs = va_arg(args, va_list); > > > this is possible on linux/x86 because va_list is a void * > > on linux/ppc however, va_list is a struct, and handled by va_ppc.h > and this makes compilation fail. > > format() will need to be rewritten without a recursive use of va_list. Is this being done (by cdrecord's upstream author J. Schilling, by cdrecord's Debian package maintainer Erik Andersen, by one of the Debian developers working on the ports to non-i386 platforms)? What do these people have to say about the situation? You should open a bug against this package so that this problem can be tracked. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: hamm
Hi Florian! > > Is it already possible to upgrade to the unstable release of debian??? > How??? The process is explained at: http://www.debian.org/doc/libc5-libc6-Mini-HOWTO/libc5-libc6-Mini-HOWTO.html (also accessible as a the link "libc5-libc6-Mini-HOWTO" from http://www.debian.org/developers_corner.htm) Most all of this process is automated by the autoup.sh script found at: http://taz.net.au/autoup/ You should still read over the Mini-HOWTO even if you use the script. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Hamm fvwm2 -- new Focus behavior
I am still confused about how some functions that I define to emulate the old Focus function (which also Raised) cause focus to automatically shift back to under the mouse, but others don't. Anyhow, here is an improvement on what I last posted, and it still behaves the way the old Focus did. >From the tail of my ~/.fvwm2/post.hook: # The Focus command no longer raises the window automatically. # Here is a function that emulates the old Focus behavior. DestroyFunc focus-and-raise AddToFunc focus-and-raise "I" Focus + "I" Raise # And here are the new bindings. Key Tab A M Next [CurrentDesk !iconic] focus-and-raise Key Tab A MS Prev [CurrentDesk !iconic] focus-and-raise Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: adduser in hamm dumps core !
Hi Joost!
I also upgraded to the 5.004.04-4 perl and perl-base packages this
morning. While adduser still does not segfault for me, perl still
segfaults and dumps core when I run either my couple hundred line mail
preprocessing script on one particular input file, or when I run
Vitaly Belostotsky's four line script given in:
http://www.debian.org/Bugs/db/17/17393.html
Have you tried to run this on you hamm system:
$ perl -e 'use POSIX; $m{'a'}=0; $n=strftime( "%b",0,0,0,0,0,0 );'
Segmentation fault (core dumped)
$ perl -e 'use POSIX; $n=strftime( "%b",0,0,0,0,0,0 );'
$
Kirk
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: Ans: Re: Debian max file size is 1GB instead of 2GB?
Steve Hsieh <[EMAIL PROTECTED]> writes: > To anyone interested in this -- > > The problem was not related to the kernel, but to ulimit. > /etc/login.defs has ULIMIT set to 1GB for some reason. However, it is > not clear to me why some accounts use this value, while other accounts > ignore it. In any case, if you comment out the ULIMIT line, all accounts > no longer have the 1GB restriction and you can create files up to the real > 2GB limit. Aha! Then this might be fixed in hamm: On my machine running bo: $ ulimit 1048576 On my machine running hamm: $ ulimit unlimited I ran out of space at 1GB on the bo machine, but I don't have space to check this on the hamm machine. Did anyone here run into a 1GB limit on a hamm machine? Kirk Hilliard P.S. There doesn't seem to be sufficient difference between /etc/login.defs on the two machines to account for this. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: adduser in hamm dumps core !
Oleg Krivosheev <[EMAIL PROTECTED]> writes:
> On Mon, 26 Jan 1998, Neilen Marais wrote:
>
[snip]
> > You can make sure by typing ldd /usr/bin/perl. If you see any mention
> > of libc5, you know something is causing perl to be linked to the worng
> > library
>
> that's the case - perl is linked against both libc6 and libc5.
>
> But i have latest perl bundle from hamm !!!
[adduser has not segfaulted for me yet, but other Perl scripts have]
I also have the latest perl bundles, and I show.
~ $ ldd /usr/bin/perl
libnsl.so.1 => /lib/libnsl.so.1 (0x4000e000)
libgdbm.so.1 => /usr/lib/libgdbm.so.1 (0x40014000)
libdb.so.2 => /lib/libdb.so.2 (0x4001a000)
libdl.so.2 => /lib/libdl.so.2 (0x40028000)
libm.so.6 => /lib/libm.so.6 (0x4002b000)
libc.so.6 => /lib/libc.so.6 (0x40044000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x4000)
libcrypt.so.1 => /lib/libcrypt.so.1 (0x400e9000)
What do you show?
The reason that I posted this instead of just emailing Oleg is that I
hope that someone here can riddle me this:
~ $ ls -il /usr/bin/{perl,perl.dist,perl5.00404}
115161 -rwxr-xr-x 2 root root 492108 Dec 9 20:17 /usr/bin/perl*
115161 -rwxr-xr-x 2 root root 492108 Dec 9 20:17 /usr/bin/perl.dist*
114698 -rwxr-xr-x 1 root root 492108 Dec 9 20:17 /usr/bin/perl5.00404*
~ $ diff /usr/bin/{perl,perl5.00404}
~ $
Why do I have two distinct copies of perl?
Kirk Hilliard
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: How to properly set TERM variable
Britton <[EMAIL PROTECTED]> writes: > I like to set TERM=xterm-color, and don't have a lot of patience with > programs that can't deal with this. When telneting, it is good to put > TERM=vt100 or some such thing in your .bashrc or it's equivelent on the > remote system. Note that, as things stand now, in hamm there is no xterm-color. /etc/terminfo/x/xterm includes the color definition, and there is a /usr/share/terminfo/x/xterm-old.[1] (There are also color_xterm, xterm-16color, and xterm-pcolor, but no xterm-color.) For some details see: http://www.debian.org/Bugs/db/14/14366.html Kirk Hilliard [1] I hate dselect[2] in a color xterm since the whitespace surrounding the descriptive text of the information section is not colored, so big blocks of reverse video text jump around as I scroll through packages. Hence, I always run: # TERM=xterm-old dselect [2] I have noticed that, in hamm, dselect no longer seems to handle xterm resizing. Also, stray characters sometime appear after rapidly scrolling through packages. Has anyone else noticed this, and is it a bug in dselect, or in ncurses, or what? -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: adduser in hamm dumps core !
Oleg Krivosheev <[EMAIL PROTECTED]> writes: > > have all hamm system and just found that > adduser dumps core independently of > any given input. Hi Oleg! Does this include when it is given no input, as in: # adduser adduser: I need a name to add. # If so, then it will be easier for people to test it without actually adding users. Anyhow, I have an all hamm system (up to date against my mirror as of 24 hours ago, giving me adduser version 3.8) and in response to your query I ran "adduser ", answered the questions, and it worked. adduser is a perl script, and bug#17393 reports perl segfaulting and dumping core on a simple script. Whereas Vitaly Belostotsky provided a four line script which does this consistently, I have a couple hundred line script which does this much more sporadically and is highly dependent on its input data (and is even sensitive to print statements that I added for debugging purposes). Since adduser reads system information, it could be that the data it gets from your system is just right to trigger this bug. I bumped the severity of bug#17393 up to important yesterday. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Hamm fvwm2 -- new Focus behavior
Hi Jameson! > I also found the fvwm changes troublesome since I had to spend a > few hours (or was that days) redoing post.hook. I also feel > something like should raise and focus on successive > windows. I found the default fvwm configuration anemic. [snip] > HERE IS SOME OF MY CODE FROM post.hook: [snip] > # I make cycle and popup windows. > # I believe this was the default before, but something has changed in fvwm2. > # CirculateUp is obsolete. > Key Tab A C Function next-and-raise > AddToFunc next-and-raise I Next [CurrentPage !iconic] Focus > + I Current Raise > + I Current WarpToWindow 5p 5p > > Key Tab A SCFunction previous-and-raise > AddToFunc previous-and-raise I Prev [CurrentPage !iconic] Focus > + I Current Raise > + I Current WarpToWindow 5p 5p [snip] That mostly does it, but I would rather not WarpToWindow -- I want the cursor to stay put. But if I leave WarpToWindow out, then focus returns to the window under the cursor. This is not the case with when just Focus is bound to the key (such as in /etc/X11/fvwm2/system.fvwm2rc) -- either with the good old behavior where Focus also raised and with the new behavior where it doesn't. It appears that this "refocus under mouse" behavior is brought on by multi-line function definitions using "+". Aha! I just discovered that if I add the Focus and Raise commands on separate AddToFunc lines without the "+", it works as I wanted. Also, it seems that when the "+" is used, then the function takes longer to execute. I don't understand why this is, do you? I am now happy with this tail of my ~/.fvwm2/post.hook: # FVWM2 Focus command no longer raises the window automatically. # Here are some functions to compensate for that. DestroyFunc next-and-raise AddToFunc next-and-raise I Next [CurrentDesk !iconic] Focus AddToFunc next-and-raise I Current Raise DestroyFunc previous-and-raise AddToFunc previous-and-raise I Prev [CurrentDesk !iconic] Focus AddToFunc previous-and-raise I Current Raise Key Tab A M Function next-and-raise Key Tab A MS Function previous-and-raise Thanks Jameson, Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Debian max file size is 1GB instead of 2GB?
On 24 Jan, Steve Hsieh <[EMAIL PROTECTED]> wrote: > > It seems that on Debian, the maximum single file size on ext2fs is 1GB and > not 2GB. Can someone confirm this, and suggest how to fix the problem, > if possible? To which Stephen P. Ryan <[EMAIL PROTECTED]> replied: > > I no longer have the space to test this, but I did have a 2GB file in > a 2.5GB partition on a Debian 1.3 system several months ago. Well, I've got a bunch of free space here. Let's see: $ # Try to make a 1.5GB file. $ perl -le '$_=6x1023; print while ++$i<=1024*512*3' > big $ ls -l big -rw-r--r-- 1 kirk kirk 1073741824 Jan 24 22:53 big $ echo "hello" >> big $ ls -l big -rw-r--r-- 1 kirk kirk 1073741824 Jan 24 22:55 big $ wc big 1048576 1048576 1073741824 big Exactly 1GB. FWIW, this is a 486/33 running stock Debian 1.3 using the kernel from a 1997-06-20 rescue disk. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: [Q] dselect-mountable install does not find packages
Hi peter! I don't understand your response. > > > For instance I don't have "Packages.gz" in 'hamm/hamm' but only > > > in 'hamm/hamm/binary-i386'. Same to others. > > > > Packages.gz is supposed to be in binary-i386. > > (there's a link from one step down to binary-i386) Where? I just scoped out ftp.debian.org, and it, like my mirror, has Packages.gz in binary-i386. No symlinks there. > > Oh! Are you not mirroring binary-all? A lot of the "packages" in > > binary-i386 are actually links to binary-all. > > If I didn't mirror it I wasn't able to install most of the stuff. :) Huh? Are you saying that of course you are mirroring binary-all since you installed a lot of packages which require it? Have you checked to see that some of the packages dselect is complaining about are in place on your mirror, either in binary-i386 or as a symlink from there to binary-all? If so, then I don't know what is going on. Have you tried dselect's "mounted" method? The "mountable" method is great, but this could provide a good debugging test. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: [Q] dselect-mountable install does not find packages
Hi Peter! > Well, perhaps -- I cannot tell really -- it's related to that I'm not > using a full mirror. For instance I don't have "Packages.gz" in > 'hamm/hamm' but only in 'hamm/hamm/binary-i386'. Same to others. Seems > that it looks for the files in the wrong directory... Packages.gz is supposed to be in binary-i386. Oh! Are you not mirroring binary-all? A lot of the "packages" in binary-i386 are actually links to binary-all. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: [Q] dselect-mountable install does not find packages
grin <[EMAIL PROTECTED]> writes: > The 'mountable' method of dselect is great -- you don't have to wait for > thousands of packages telling you "it's already installed". Agreed! BZ to its author (Andy Mortimer, I think). > But while from the local disk mirror it worked flawlessly, I failed > to use it through NFS. I am using it succesfully via NFS against a mirror of Hamm. > I have a partial mirror of hamm on the NFS, containing only the > directories you see below: > > I set it up to use NFS mount in fstab, entered the dirs (hamm/binary-i386, > contrib/binary-i386, non-free/binary-i386 and non-us/hamm), and update > worked fine. After selecting I went on to [I]nstall, but it told me OK, just curious here -- what did you enter non-us/hamm as -- your local arcive? This is not related to your problems, but I was surprised that I was only asked for main, contrib, non-free, and local -- no mention of non-us. > "Warning: package XX, version X.Y-Z not available for installation. > Skipping." I don't know about this, but ... > and some > > "Error: package XX does not have a filename! Skipping." this sounds like bug#8054 against dpkg-mountable. If a package fails to install, dselect (using the mountable method) looses track of it (Andy thinks that some strange behavior of dpkg itself is to blame for the actual problem) and you have to rerun "1. [U]pdate" in dselect. Give that a try and see what happens. It may be that some other problem is keeping the packages from installing, giving this bug a chance to rear its head. Let me know what happens, as I am very interested in seeing this wonderful method working well for the hamm release. Andy, do you think that bug#8054 will be resolved prior to the release of hamm? Do you need someone to help debugging? Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Hamm fvwm2 -- new Focus behavior
The behavior of the Focus command has changed in the version of fvwm2 provided in hamm. /usr/doc/fvwm2/ChangeLog.gz says: 2.0.46 [snip] 37. Focus command no longer raises - use an interpreted focus-and-raise function if that's desired /etc/X11/fvwm2/system.fvwm2rc (in both bo and hamm) contains the lines: Key Tab A M Next [CurrentDesk !iconic] Focus Key Tab A MS Prev [CurrentDesk !iconic] Focus I prefer the old behavior where Alt-Tab raised the window that it shifted focus to, so I added the following lines to my ~/.fvwm2/post.hook: DestroyFunc focus-and-raise AddToFunc focus-and-raise + "I" Raise + "I" Focus Key Tab A M Next [CurrentDesk !iconic] focus-and-raise Key Tab A MS Prev [CurrentDesk !iconic] focus-and-raise This works as long as my mouse is not in a window, but if it is, then focus shifts back to the window that the mouse is in after it finishes raising the next one. How can I duplicate the behavior of Focus in bo where the focus stays with the newly raised window regardless of mouse position? (This behavior is independent of SloppyFocus, which I normally use.) Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
pcmcia Megahertz XJ10BT troubles
Hi all! I have two PCMCIA ethernet cards, both on the list of supported cards for pcmcia-cs, and one works fine but the other does not seem to be recognized. The details: Computer: Toshiba Tecra 730CDT Cards: Linksys EthernetCard Model No. EC2T (works) Megahertz PCMCIA Ethernet Model XJ10BT or XJ10BC (It says XJ10BT on the box. I think that XJ10BC is the same card but with a dongle for 10Base2.) Kernel: 2.0.32 with Tecra patch, built with default Debian options. Software: pcmcia-cs 2.9.12-2 pcmcia-modules-2.0.32 2.9.12-2 The Linksys card works fine, but when I insert the Megahertz card I get two beeps, one high, one low, followed by the following message on the xconsole: Jan 22 19:54:37 globe cardmgr[88]: initializing socket 1 Jan 22 19:54:37 globe cardmgr[88]: unsupported card in socket 1 Jan 22 19:54:37 globe cardmgr[88]: version info: "Meg", "", "@", "" Jan 22 19:54:37 globe cardmgr[88]: manfid: 0x0128, 0x0103 I had the same behavior in bo (using pcmcia-cs_2.9.6-3 and pcmcia-modules-2.0.29-7_2.9.6-3) except that I only got one beep (low) before the above message appeared. /usr/doc/pcmcia-cs/SUPPORTED.CARDS.gz contains the following entry in the list of supported cards: Megahertz XJ10BT, CC10BT Ethernet I checked out David Hinds' site http://hyper.stanford.edu/HyperNews/get/pcmcia/forums.html and didn't find any problems reported against this card. Any suggestions? Kirk Hilliard P.S. The Linksys card is physically damaged and I need to prop popsicle sticks under the connector to make it work, so I would like to get the Megahertz working. Besides, I want to see how well that cool looking XJack works in practice. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: emacs, X, and Alt key
Hi Mark! > I, like Thomas Vaughan, have been having problems getting emacs to > recognise the Alt key under X. (The system is a Toshiba laptop.) I'm running a Toshiba Tecra 730CDT, and the Alt key in emacs under X works fine in combination with other keys without any special setup. What model is your laptop and what window manager are you using? One problem with fvwm2 it is the line in /etc/X11/fvwm2/system.fvwm2rc which breaks Meta-mouse: Mouse 0 W M move-and-raise-or-raiselower I fix it by adding this line to my ~/.fvwm2/post.hook: Mouse 0 W M - Meta-mouse is used in emacs for secondary selection. I will bring this up with the fvwm2 maintainer if this line is still in the hamm version. Kirk Hilliard P.S. I would rather use "emacs -nw" in an xterm than use emacs in its own window if it were not for secondary selection and mouse-yank-at-point. Put (setq mouse-yank-at-point t) in your ~/.emacs and give it a try. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Mirroring hamm question
I don't have an answer to Caleb's questioning about mirroring hamm, but I do have a related question. There are symbolic links in the hamm tree which point to the corresponding bo file for files which have not yet been upgraded to hamm. I would like to mirror all of hamm, and only that part of bo which is pointed to by links from hamm. From reading the mirror man page, it looks to me like I could do this by "flattening out" the remote site symlinks using ``flags_recursive+L'', but I would rather have those bo files pointed to by hamm links actually mirrored in a (partial) bo tree, and keep the links as links in my local hamm mirror. Is it possible to tell mirror to do that? I don't have enough disk space to have a complete mirror of bo in addition to hamm, and I like to be able to see the links as they indicate that the file has not yet been uploaded for hamm. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: ThinkPad 560
Hi Paul! Paul Rightley <[EMAIL PROTECTED]> writes: > I know of no Thinkpads that can boot bzImages (though I do not know > many Thinkpads) - which is why the 'tecra' disks work (i.e. they are > zImages). Is this true? I thought that the Tecra disks were still bzImages, but were compiled with the kernel patch from Jen's Maurer's site, "Linux on the Toshiba Tecra series": http://www.cck.uni-kl.de/misc/tecra710/ If this is still the case, is the patch provided in one of the source packages? Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Cron and find
[EMAIL PROTECTED] writes: > I often have some nfs filesystems mounted on my machine and whenever cron > runs updatedb I end up with a huge mail listing each file on the nfs > filesystems saying permission denied. Is there a way to stop this message > without unmounting the filesystems? Hi G'razel! Have you tried adding the NFS mount point to PRUNEPATHS in /etc/updatedb.conf? Check out updatedb(1). HTH, Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: ALT key in X for Emacs
Mark Phillips <[EMAIL PROTECTED]> wrote: > The ALT key works under emacs when in console mode. But it doesn't work > under X. Is there any reason for this? How can I fix it? > > Perhaps I need to use xmodmap? I heard something about a userfriendly > version of this. Does anyone know what it's called? Paul Seelig already mentioned xkeycaps. Perhaps this is what you were looking for, but if your problem is that Meta-Mouse buttons (used for secondary selection -- a very important emacs feature allowing selection of text without moving point) no longer work, it may be due to this insidious line which crept into bo's /etc/X11/fvwm2/system.fvwm2rc: Mouse 0 W M move-and-raise-or-raiselower I override it with these lines in my ~/.fvwm2/post.hook: # Allow Meta-mouse events to pass on to the clients. This is # especially important for emacs with secondary selection. Mouse 0 W M - I have been meaning to check and see if a bug report has been filed on this yet. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: emacs and .Xdefaults
Last week I issued a plea for help: > I just made a new installation of bo, and emacs no longer seems to > read my ~/.Xdefaults file like it used to in rex. And Wintermute <[EMAIL PROTECTED]> came to the rescue: > Try this.. > > xrdb -merge ~/.Xdefaults That does the job! Since it only works for the current X session, I added it to my ~/.xsession file and all is fine. Still, I did not have to do this in rex. Is this change a bug or a feature? Thanks, Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
emacs and .Xdefaults
I just made a new installation of bo, and emacs no longer seems to read my ~/.Xdefaults file like it used to in rex. ~ $ cat ~/.Xdefaults emacs.font: -misc-fixed-medium-r-semicondensed--13-*-*-*-c-60-iso8859-1 ~ $ cat ~/.Xdefaults-globe emacs.font: -misc-fixed-medium-r-semicondensed--13-*-*-*-c-60-iso8859-1 ~ $ hostname globe ~ $ emacs ~ $ echo "That one came up with the big ugly default font." That one came up with the big ugly default font. ~ $ emacs -fn "-misc-fixed-medium-r-semicondensed--13-*-*-*-c-60-iso8859-1" ~ $ echo "That one came up with the nice small font that I want." That one came up with the nice small font that I want. ~ $ How can I get this to work again? Kirk "alias emacs='emacs -fn fixed'" Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Laptop computers
On Nov 26, 10:10, "Kevin Traas" wrote: > Well, I've got a Tecra 510, but I haven't found an XServer other than VGA > that works with the video in this thing So, X isn't much of an > option > > Unless anyone else can help - I'd be much appreciative! (Chipset is > CL65550.) Hi Kevin! Check out: Jens Maurer's Linux on the Toshiba Tecra series http://www.cck.uni-kl.de/misc/tecra710/ While Jens owns a 710, the page discusses all of the Tecras. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Yamaha CDrecorder 400tx
On Nov 24, 11:31, Timothy Phan wrote: > I'd like to know has anyone here used the Yamaha CD Recorder 400tx > with Adaptec 2940 scsi control under Debian/Linux to record files > succesfully? > > I've just purchased the recorder and try out the command: > $ cdwrite --eject > and the kernel/Adaptec driver kept printing out of a lot of message > concerning the timeout on the CDrecorder device. Hi Timothy! Use cdrecord, not cdwrite (which I understand is very out of date). I have only done simple things, such as recording images made with mkisofs, but cdrecord has worked with my Yamaha CDR400tx. Unfortunately, last I looked, the only Debian cdrecord package was in hamm (unstable), so I grabbed the sources from http://www.fokus.gmd.de/nthp/employees/schilling/cdrecord.html and built it myself without any problems. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
loadlin -- A20 gating failed (tecra/bo)
Hi! I just installed Debian 1.3 on my Toshiba Tecra 730CDT (this is a "fresh" installation, separate from my 1.2 installation) and I can't seem to get it to boot using loadlin. I used the 1997-06-20/tecra rescue and modules disks, and both the rescue disk and the resulting boot disk work fine, but when I try to boot from loadlin I get an "A20 gating failed" message and the computer locks up. The vmlinuz that loadlin is using is the same as both /boot/vmlinuz-2.0.29 and the boot floppy's linux. I know that there is a problem with Tecras and bzimges, but that's what the "tecra" disks are for. What is going on here? Loadlin works fine with rex's vmlinuz-2.0.27 tecra kernel. (I am using loadlin-1.6.) Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: cdwrite trouble: SENSE_ERROR
Thanks to all who offered help on getting my Yamaha CDR400tx (1.0d) CD recorder working, especially to Nathan Norman who pointed me to http://www.shop.de/cgi-bin/winni/lsc.pl where I discovered that cdwrite-2.0 does not support the CDR400tx but that cdrecord-1.5 does. There does not seem to be a Debian package for cdrecord (is someone working on this?) so I built it from the source at ftp://ftp.fokus.gmd.de/pub/unix/cdrecord/cdrecord.tar.gz and it works great! I do seem to have a reliability issue with some fairly cheap green on gold CDRs. (US$4 ea @ qty 100 with a black label thermally printed on them by a local company -- supposed to be Sony CDRs, but they were totally blank before being printed). The cdrom drives on a couple of our UNIX boxes can't seem to reliably mount them if they are written at 4X, but (so far, at least) they seem to be OK if written at 1X or 2X. Does write speed affect the quality of the CDR? Also, are gold on gold CDRs supposed to be better? I found what I think is a pretty good deal on some HP C4432A gold on gold CDRs at CompUSA -- qty 25 for US$70 minus US$20 rebate (limit two on the rebate). Kirk Hilliard [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
cdwrite trouble: SENSE_ERROR
Greetings, I am having difficulty writing a CD using a Yamaha CDR400tx on a Gateway2000 P5-90 (using a BusLogic B-958 SCSI card) running Debian 1.3.1 (cdwrite 2.0, mkisofs v1.11). This same CD writer seems to work on an NT machine. The CD-Writing mini-Howto talked about using scsi devices such as /dev/sg[a-h], but my box has /dev/sg[0-7]. The CD writer is SCSI ID 1, but is the only device on the chain, so I guessed /dev/sg0, and that seems to be right since the following successfully ejects the CD: # cdwrite -v --eject --device /dev/sg0 cdwrite 2.0 Manufacturer: YAMAHA Model: CDR400t Revision: 1.0d Using mode:Yamaha Using speed: 2 But when I try to write a CD I get the following errors: # cdwrite -v --device /dev/sg0 UnixCD.iso cdwrite 2.0 Track 01: data 26 Mb Manufacturer: YAMAHA Model: CDR400t Revision: 1.0d Using mode:Yamaha Using speed: 2 mode_select6#2 result 0, pack_id 12 sense 70 00 05 00 00 00 00 0A 00 00 00 00 26 02 00 00 18 of 18 mode_select6#2 reply bytes: 00 00 00 00 31 02 00 10 59 41 4D 41 48 41 20 20 43 44 write_data_track result 0, pack_id 14 sense 70 00 05 00 00 00 00 0A 00 00 00 00 20 00 00 00 18 of 18 write_data_track reply bytes: 00 00 00 08 00 00 00 00 00 00 08 00 48 41 20 20 43 44 SENSE_ERROR iter 1: pipe_to_cd result 0, pack_id 15 sense 70 00 05 00 00 00 00 0A 00 00 00 00 20 00 00 00 18 of 18 pipe_to_cd reply bytes: 00 00 00 08 00 00 00 00 00 00 08 00 48 41 20 20 43 44 SENSE_ERROR iter 2: pipe_to_cd result 0, pack_id 15 sense 70 00 05 00 00 00 00 0A 00 00 00 00 20 00 00 00 ... and so on until I kill it. Has anyone here seen this before? TIA, Kirk Hilliard [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Maarten Boekhold et. al.
Oops. Kirk "can't differentiate between Subject: and Cc:" Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Maarten Boekhold , Manoj Srivastava , sanvila@ctv.es
Hi, [Concerning a possible bug in flex-2.5.4] For the example I gave, Manoj and Maarten suggest adding ``return 0'' to the action after the input() receives an EOF, and in this case it would solve the problem, but I don't think that this should be necessary and there are other situations where this would not work so well. I still think that there is a problem with flex-2.5.4, but this discussion doesn't really belong on debian-user, so I will email my objections to Manoj and Maarten (and anyone else who sends a request), and then ask some questions on comp.lang.c before taking it to gnu.utils.bug. I'll post a note back here vindicating either flex-2.5.4 or my suspicions when I figure it all out. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Possible flex-2.5.4 bug
Hi,
[In response to my post regarding a possible flex-2.5.4 bug]
Manoj Srivastava <[EMAIL PROTECTED]> wrote:
>Each argument is scanned separately, with yy_scan_string. An
> end of file is generated, then, at the end of each string (each
> argument). There is no yywrap function defined.
Right. My intention was not to wrap but to call yylex() thrice, thus
my use of "%option noyywrap". I fear that the multiple calls in my
example may have obscured the problem.
> So, you call yylex thrice, on three separate, non-wrapped files. The
> second file (string, argument, whatever) does end inside a comment,
> and since there is no wrapping, the eof is final. The error is as
> you asked it to be reported.
Right, but the program (when build with flex-2.5.4) then dies inside
the yylex() call and the third argument is never processed. To
clarify this, I have modified the example to print a message when
yylex() returns. For variety, I have also added an explicit yywrap()
function (which always returns 1). This time I provide single
arguments to the program to avoid the red herring of the multiple
yylex() calls.
In the example below, the program is run twice when built with
flex-2.5.2 and twice when built with flex-2.5.4, first with a
terminated comment and then with an unterminated comment. Notice that
when built with flex-2.5.4, it dies immediately following the "input
in flex scanner failed" message (generated internally by the flex
skeleton) when given the unterminated comment.
Manoj, thanks for the quick response. I receive a little grief at
work for my support of free software, especially from people who think
that an 800 phone number is necessary for support, but I feel that
commonly used free software typically has better support than most
commercial software. This is not an emergency since I was able to
install flex-2.5.2 on our RS/6000 from the sources on my rex disk. My
goal now is simply to report this bug. I am still a little skeptical
of my own code since it is not doing anything obscure and flex-2.5.4
has been out for over a year now. One thing which is clear is that my
program does what I expect it to with flex-2.5.2 but not with
flex-2.5.4.
Kirk Hilliard
--
~/flexbug $ flex --version
flex version 2.5.2
~/flexbug $ ./flex-2.5.4 --version
./flex-2.5.4 version 2.5.4
~/flexbug $ cat foo2.l
%array
%%
"<" { int c;
while ( (c = input()) != '>' && c != EOF )
;/* eat up text of comment */
if ( c == EOF ) printf( "\n" "error: EOF in comment\n" );
}
foo printf("bar");
%%
void main(int argc, char **argv) {
YY_BUFFER_STATE ybsInputBuffer;
while(--argc) {
ybsInputBuffer = yy_scan_string(*++argv);
yylex();
printf( "\n" "yylex() done\n" );
yy_delete_buffer(ybsInputBuffer);
}
} /* main */
int yywrap(void) {
return 1;
} /* yywrap */
~/flexbug $ flex foo2.l
~/flexbug $ gcc lex.yy.c
~/flexbug $ ./a.out "foobell"
barbell
yylex() done
~/flexbug $ ./a.out "foobell"
barbell
yylex() done
~/flexbug $ echo "*** Here comes the bug. ***" > /dev/null
~/flexbug $ ./a.out "foo /dev/null
--
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Possible flex-2.5.4 bug
Hi all!
Here is a simple lexer which works with flex-2.5.2 but not with
flex-2.5.4. Is this a bug in flex-2.5.4 or is 2.5.4 stricter,
revealing a bug in my code?
This simple lexer replaces "foo" with "bar" (just to do something) and
strips out comments delimited by '<' and '>'. (In real life I am
looking for C-style comments, but single character delimiters yield a
simpler example.) It works with flex-2.5.2 but with flex-2.5.4 it
blows up following the input() call which returns EOF.
There was a similar bug in flex-2.5.3, but the NEWS file that comes
with 2.5.4 says:
Changes between release 2.5.4 (11Sep96) and release 2.5.3:
- Fixed a bug introduced in 2.5.3 that blew it when a call
to input() occurred at the end of an input file.
[snip to end]
I have Debian 1.2 installed, giving me flex-2.5.2. I built flex-2.5.4
directly from the GNU package. Could someone running bo or hamm
verify that this lexer breaks with the Debain flex-2.5.4 package, and
could someone who knows a little more than I do about flex tell me if
I am screwing something up. If my code looks OK I will submit a bug
report.
(I discovered this problem on an RS/6000 where I have an immediate need
for a working flex. I could not locate flex-2.5.2 on the net, but was
happy to find the source on the rex CD in my laptop.)
Kirk Hilliard
$ flex --version
flex version 2.5.2
$ ./flex-2.5.4 --version
./flex-2.5.4 version 2.5.4
$ cat foo.l
%array
%option noyywrap
%%
"<" { int c;
while ( (c = input()) != '>' && c != EOF )
;/* eat up text of comment */
if ( c == EOF ) printf( "\n" "error: EOF in comment\n" );
}
foo printf("bar");
%%
void main(int argc, char **argv) {
YY_BUFFER_STATE ybsInputBuffer;
while(--argc) {
ybsInputBuffer = yy_scan_string(*++argv);
yylex();
yy_delete_buffer(ybsInputBuffer);
putchar('\n');
}
} /* main */
$ flex foo.l
$ gcc lex.yy.c
$ ./a.out "foobar" "foobell"
barbar
bar
error: EOF in comment
barbell
$ ./flex-2.5.4 foo.l
$ gcc lex.yy.c
$ ./a.out "foobar" "foobell"
barbar
bar
error: EOF in comment
input in flex scanner failed
$
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: Netscape uses SOCKS host for local servers
> Subject: Re: Netscape uses SOCKS host for local servers > Kirk Hilliard wrote: > > How do I get Netscape to not use a SOCKS host for machines in my local > domain? > > I am running "Mozilla/3.01Gold (X11; I; Linux 2.0.27 i586)" on my > Debian 1.2 box, which is on a network behind a (SOCKS 4) firewall. > After I set "Options/Network Preferences/Proxies" to "Manual Proxy > Configuration" and filled in the "SOCKS Host" field it had no problems > getting through to the outside world. However, it goes through the > SOCKS host even to contact local http servers. (This is a problem > because the SOCKS host is in a different building to which I am > connected with only 10Kbps pipe.) I set the "No Proxy for" field to > the local domain name but this had no effect. > > "Mozilla/3.0Gold (WinNT; 1)", similarly set up to use the SOCKS host, > connects to local http servers directly, even without an entry in the > "No Proxy for" field. Jens B. Jorgensen replied: > If the net you want to be local is, say, 192.168.2.0, (class-C) then > in the "No Proxy for" field put '192.168.2.'. Get the idea? Jens, Thanks, for the help, but I have tried both this (with and without the third dot) and putting the full four byte dotted IP address for the server in the field, and it connects but still goes through the SOCKS host. Does this actually work for you? (One way to find out is to use a CGI script which prints out $REMOTE_HOST.) If so, what version of netscape are you using? I have also tried the symbolic address both for the local net and for the server, and all of the above followed by :80 (the port number), all to no avail. Any hints? Kirk -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Netscape uses SOCKS host for local servers
How do I get Netscape to not use a SOCKS host for machines in my local domain? I am running "Mozilla/3.01Gold (X11; I; Linux 2.0.27 i586)" on my Debian 1.2 box, which is on a network behind a (SOCKS 4) firewall. After I set "Options/Network Preferences/Proxies" to "Manual Proxy Configuration" and filled in the "SOCKS Host" field it had no problems getting through to the outside world. However, it goes through the SOCKS host even to contact local http servers. (This is a problem because the SOCKS host is in a different building to which I am connected with only 10Kbps pipe.) I set the "No Proxy for" field to the local domain name but this had no effect. "Mozilla/3.0Gold (WinNT; 1)", similarly set up to use the SOCKS host, connects to local http servers directly, even without an entry in the "No Proxy for" field. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Latest rescue disk locks up Thinkpad
Dean Carpenter <[EMAIL PROTECTED]> wrote: > I tried installing frozen on an IBM Thinkpad 760E last night - no > success. The rescue disk would boot, load root.bin then linix.bin? > and lock up after a bunch of "loading dots". Locked up tight - cold > boot time. The 1997-05-16/resc1440.bin rescue disk also fails to boot Toshiba Tecras. I recall some discussion recently that the Tecra patch was no longer going to be applied because it was thought to interfere with other computers. This confused me because Sven Rudolph forwarded a message to me in which he wrote (concerning the Tecra patch): > > (I deleted the patch because I thought it became part of > > 2.0.29. Supposedly I were wrong.) and Herbert Xu replied: > It is part of 2.0.30-1. And 2.0.29 is now obsolete. This raises the questions: - Has the Tecra patch become an integral official part of the newer kernels, and what kernel is being used on the rescue disk? - Does the Tecra patch also help the Thinkpad problem or is it unrelated? - Is there going to be a Tecra version of the rescue disk generally available? Dean, I could provide you a pair of (old) rescue disks, one with, and one without the Tecra patch, if you want to experiment. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: pointing device on Toshiba Satellite 110CS
On Mar 19, 8:38, Karsten Bolding wrote: > On a Toshiba Satellite 110CS I would like to be able to use the > pointing deviec situated on the keyboard instead of an external > mouse attached to the serial port, which I want that to use for a > modem. The info I'm after is the device and protocol to enter in > XF86Setup. I guess I have to recompile the kernel with the proper > support? Karsten, The AccuPoint "eraser tip pointer thingy" (which I expected to hate, but learned to love) on my Toshiba 730 is just a regular PS/2 mouse. You can use Protocol "PS/2" Device "/dev/psaux" in your "Pointer" Section. I used to use gpm a lot back when I had a machine on which X was slow, but now I rarely leave X. Still, I want gpm available on those rare occasions that I use the console, and the 2.0 kernels only allow one user of /dev/psaux at a time, so I run gpm with gpm -m /dev/psaux -t ps2 -R (`-R' setting up the repeater). I then put Protocol "MouseSystems" Device "/dev/gpmdata" in my XF86Config, and the mouse works great both in X and on the console. (I understand that gpm's `-M' switch will even allow you to use two pointers simultaneously.) You will, of course, want to add Emulate3Buttons Emulate3Timeout 50 to your XF86Config. I find that I use mouse-2 in X much more than mouse-3, so I run xmodmap -e "pointer = 1 3 2 to swap the two. The one exception is that I use mouse-3 (WindowList) against the fvwm2 background much more than I do mouse-2, so I added Mouse 2 R A WindowList Mouse 3 R A Menu Window-Ops to my ~/.fvwm2/post.hook to swap those two. Email me if you need more details. Kirk Hilliard
Re: Is there a dictionary for abbreviations like "WTF"?
On Feb 18, Dr. Andreas Wehler wrote > I would like to resolve these many abbreviations today, as > So, is there any appropriate dictionary? Thanks. J. H. M. Dassen gave several URL's which should answer your questions, but the start of your subject line reminded me of something else that I am interested in -- a ``webster'' like dictionary that I could install on my linux box. I think that ``webster'' might come with NeXT machines, but it has been available at every university at which I have had a UNIX account, and I find myself telnetting into my university account several times per day, specifically to make use of it. (Yes, I have a real paper and ink dictionary, and I usually choose to use it when I am reading a book, but I want an online dictionary when I am typing or reading mail.) Several linux dictionary options come to mind. WWW -- I believe that there are www dictionary sites available, but I want a dictionary installed on my machine that I can use whether or not I am connected to the net. Commercial UNIX dictionaries -- I recently saw a c.o.l.a article announcing the availability of a ``webster''-like American Heritage dictionary for linux, but the price was a bit steep. (I think it was $89 US for a single user license.) DOS or Windows dictionary, converted for linux -- I would spend $20 - $30 US for such a dictionary on CDROM if I could break the format of the data file and write my own linux viewer for it. (This would probably be a violation of the letter of the license, but not the spirit, as I would only use the "linux version" installed one single user machine.) GNU Free Dictionary project -- What ever happened to this? Gutenberg Project dictionary -- I just noticed that the Gutenberg Project released a 1919 Webster dictionary last year, both as an ASCII flat file, and in html format. That is a little old (necessitated by copyright concerns) but it would serve nearly all my purposes (and add some historical flavor to boot). I believe that it is 40 MB uncompressed, and I plan on downloading it shortly. Has anyone else had a look at it? While an html version would be usable, I would like to see a dictionary specific browser/viewer built for it. It is not appropriate to wrap up all the works of the Gutenberg Project as debian packages, but it might be appropriate to do so with a dictionary and thesaurus. (They have also released a version of Roget's Thesaurus.) At the very least, a contrib package could be built containing viewers (and possible even ``webster''-like client and server) and including instructions for downloading the data files. Is any one else interested in this? Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: loadlin (was: CD-ROM -- how to mount, etc.)
On Jan 20, 9:23, "Seak, Teng-Fong" wrote: > One question: is there the loadlin package provided in Debian? Yes, Debian 1.2 contains a loadlin package with loadlin version 1.6. Beware that the tools directory on a Debian 1.2 CDROM (or the ftp site) contains loadlin 1.5 which cannot load bzImages. I could not find out who is responsible for this directory so I sent a note to Christoph Lameter, the loadlin package maintainer, which he forwarded on to Bruce. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Tecras and Debian 1.2
Concerning the Tecra bzImage boot patch, Hamish Moffatt wrote: > What does this fix exactly? I have a server machine located at my > service provider, which requires a keyboard; without one, it > says "Loading Linux..." then waits forever until a keyboard > is connected. Will this fix this? Nope, I don't think so. The Tecra problem is one of cache coherency when the A20 is first enabled, and it causes the machine to reboot after "Loading Linux...", returning you to "boot:". The kernel patch simply flushes the cache. More details can be found at: http://www.cck.uni-kl.de/misc/tecra710/ and a temporary replacement rescue disk can be found at: http://www.math.virginia.edu/~kdh5j/debian/tecra730 The next set of boot disks will contain this kernel patch. I think that your concern is a common one. Perhaps other list members will be able to help. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Tecras and Debian 1.2
On Jan 14, 3:20, Robert Nicholson wrote: > Ok, I'm hearing that somebody has applied the a20 gate patch to the > latest boot disk set is that right? > > Can we now expect to be able to install 1.2 on 700 series Tecras with > this new disk set? Dale Scheetz reported that the patch will be in the next set of boot disks. For those who can't wait, you can find a temporary replacement rescue disk at: http://www.math.virginia.edu/~kdh5j/debian/tecra730 This problem (machine reboots after "Loading linux ") is not just for Tecras anymore. It was also reported by an owner of a Compudyne Pentium 60 desktop, whose problems were solved with the patched rescue disk. > NOTE: for those who have been following this issue closely. Although > _somehow_ I was able to get Linux on my notebook with the rex rescue > disk for some reason _now_ it reboots like everybody else :-( So I'm > looking for an update on this issue. Welcome back into the fold of non-bzImage-booting Tecra 700 series owners. Kirk Hilliard Irreproducible events abound! -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: A proposal to improve dselect
Ian Jackson is listed as the maintainer of the dpkg package which contains dselect. I just sent him a short note letting him know that there is a group of users who wish to help improve dselect, and asking for his guidance. If he is very busy, he may prefer not to be a member of the dselect project mailing list, but instead keep in contact with a single representative of the group. I noticed from the Debian web page that he tops the list of maintainers with 141 outstanding bugs. I am certain that this is due to his central role in the project and not due to any lack of effort. Last year I read one of the mailing lists which he quite was active on, and for several months I thought that he was the Ian in Debian. I still consider him "the other Ian in Debian." It looks to me like he could use some help. I suggest that we await his response and guidance. While checking out the Debian web page, I notice that there is a debian-dpkg mailing list mentioned, but no archived messages are available for it. Does anyone know the story behind that. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
A proposal to improve dselect
in place, I propose that we form a working group to initially hash out how dselect should be improved, and then to actually implement those improvements. Kirk Hilliard Unix *is* user friendly -- it's just picky about who it makes friends with. Let's make dselect more sociable. Is this the best venue to continue this discussion, or should it be taken elsewhere? Does debian-talk still exist? -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Tecra Rescue Disk
With some help from Martin Stromberg I have built a temporary replacement rescue disk for Toshiba Tecra 700 series laptops. (They cannot boot a normal bzImage loaded by lilo.) It has the same generic 2.0.27 kernel as in the 1996-12-8 resq1440.bin rescue disk, but with Jens Maurer's kernel patch. This should help out until the new boot disks arrive. You can find my resq144t.bin at http://www.math.virginia.edu/~kdh5j/debian/tecra730 Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: rescue disks for a tecra
I must admit that only my second thought when I read Bruce's note was for his family's and his property's safety, my first being, "Oh no! There goes the new boot disks." The problem is that on Tecra 710/720/730 laptops (I don't know about the 500 series) lilo cannot load a bzImage. With the rescue disk, you get "Loading root.bin " and "Loading linux ", and then the laptop reboots and gives the "boot:" prompt again. Loadlin (version 1.6), however, does not have a problem with bzImages. It has been reported that the Tecra 710/720 BIOS upgrade (720V580.EXE of 12-06-96) fixes this. The Tecra 730 BIOS upgrade (730V530.EXE also of 12-06-96) does not. I contacted Toshiba about this, and the Technical representative that I spoke with was very friendly, but I have not heard back from them yet. The BIOS upgrades are available at http://www.toshiba.com/tais/csd/support/files/. At http://www.cck.uni-kl.de/misc/tecra710/, Jens Maurer discusses the problem and gives a kernel patch which works around it, a patch which Bruce was going to incorporate it into the next set of boot disks. I would try it myself as a service to other Tecra owners if I knew what went into the boot disks (possible just the rescue disk). Can any of the developers out there help? Debian 1.2 was going to be an initial installation on my machine, but due to the problem, I had to install 1.1 first and then upgrade. (PHT Linux Monthly CDs finally had a use!) I made a minimal installation from the 1.1 CD, but for the service of those who only have a 1.2 CD, I could test to see what the absolute minimum of 1.1 is necessary. Perhaps it is possible to use only the 1.1 floppies, and then use the 1.2 CD when dselect is first run. I doubt that it would be possible to use the 1.1 boot and root disks, and then substitute the 1.2 base disks, but I could try. Once you have a working system, you can make it bootable with the new bzImage via loadlin, but make sure that you use loadlin version 1.6 which is in the Debian 1.2 loadlin package, not loadlin version 1.5 which is in the /tools directory and can't handle bzImages at all (nothing to do with Tecras). (I have notified the loadlin maintainer about this, but he was not sure who maintained /tools, so he past the note on to Bruce. I notice that ftp.debian.org still has loadlin version 1.5.) The alternative to using loadlin, of course, is building a kernel with the patch. Richard G. Roberto <[EMAIL PROTECTED]> wrote: > The new kernel doesn't boot and the old kernel has been overwritten. > This was a 1.1 upgraded, so I don't even have an older working set > of 1.2 floppies. Richard, If you have a 710/720, I recommend the BIOS upgrade. If you have a 730, you need to boot your system somehow. If you could get a copy of the 1.1 boot and root disks, you could mount your system and then grab your new kernel for use with loadlin. Unfortunately, buzz seems to have disappeared from ftp.debian.org. Alternatively, I could send you a vmlinuz and loadlin.exe (657183 and 32208 bytes). Can you uudecode? If not, what other options are there for email binary transfer? I could try to get a minimal web site up tonight as an alternative. I don't have access to outgoing ftp, but a link to a binary should be downloadable over the web. Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Debian 1.2 rescue disk reboots Toshiba Tecra laptop
Greetings! History: Toshiba Tecra laptops (at least the 710, 720, and 730) can't boot the bzImage rescue disk. I mailed a patch (from http://www.cck.uni-kl.de/misc/tecra710/) to Bruce and this problem might be fixed on the next set of boot disks. I got Debian 1.2 up on my Tecra 730 by first installing Debian 1.1 and then upgrading to 1.2, setting the machine to boot from loadlin. Note that the loadlin in the tools directory is version 1.5 which does not load bzImages. The loadlin package contains version 1.6 which has no problems with bzImages. Is someone maintaining a list of 1.2 glitches and simple fixes a la mode de Scott Barker's "SimpleFixes for Slackware Linux"? I am very impressed with Debian 1.2 Congratulations! Thanks, Kirk Hilliard -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Debian 1.2 rescue disk reboots Toshiba Tecra laptop
Greetings! I have a Debian 1.2 CD from iConnect. When I boot my Toshiba Tecra 730CDT laptop from the rescue floppy (either resq1440.bin or the new new-resq1440.bin) I get "Loading root.bin " and "Loading linux ", and then the laptop reboots from the floopy, giving me the "boot:" prompt again. At http://www.cck.uni-kl.de/misc/tecra710/, Jens Maurer reports that his Tecra Laptop cannot load "bzImage" format kernels. (He has submitted a kernel patch for this.) Does the rescue floppy use a bzImage kernel? I do have a Debian 1.1.11 CD (PHT Nov 96 Monthly Archive), and I can boot from its installation boot and root disks. Will I be able to get a Debian 1.2 rescue floppy for this laptop? If not, I can install 1.1 and then upgrade to 1.2. Is it possible to shift to 1.2 part way through the installation, either by providing the 1.2 base disks or by substituting the 1.2 CD when I reboot, or do I need to run dselect on 1.1 before shift to the 1.2 upgrade? This problem does not affect loadlin. Thanks, Kirk Hilliard P.S. Is someone maintaining a list of 1.2 glitches and simple fixes? -- This message was delayed because the list mail delivery agent was down.
