Re: Bug#60891: xdm: Installs default tty7 line in Xservers -- break s system
Sven LUTHER [EMAIL PROTECTED] writes: Just a stupid clueless idea, but what if we start the xserver from the inittab, or at least have some line in the inittab saying that the Xserver/gdm/whatever will use the vtxx, in the same way that init starts getty's on his specific vt's ? This is what HPUX does, for what that's worth. It seems to work OK. BTW, should we have a runlevel with X and a runlevel without X ? or do we already have such a thing ? Again on HPUX, this is runlevel 4; runlevels 2,3 are multiuser but don't start vue (HP's xdm/kde/... equiv). It doesn't look like Debian distinguishes between 2,3,4,5, at least on my woody-with-filerc-i386 system. I mostly agree with Branden that having xdm and friends fiddle with important system config files is not so good, I just thought I'd point out that the init/runlevel thing does work on some unixen. - M -- Mark Monty Montague | [EMAIL PROTECTED] | I don't do Windows(tm) I'm dubious about any company whose assets can be destroyed by rm -rf URL:http://www.gg.caltech.edu/~monty/monty.shtml X-PGP-Fingerprint: E4 EA 6D B1 82 46 DB A1 B0 FF 60 B9 F9 5D 5C F7
Re: Netscape helpers, or replacing /bin/sh
jim [EMAIL PROTECTED] writes: I recently switched to debian 1.31 (from a homebrew system based loosely on slackware), and with it came bash 2.x, which I had been avoiding for irrational reasons - or maybe not. I installed Netscape Communicator 4.04 and the Real Audio player (v5.02beta, the latest), and followed the directions in the readme for adding rvplayer as a helper app. But now, when I actually try to play something, I get an error dialog from Netscape that says: sh: -c line 1: missing closing `)' for arthmetic expression sh: -1 line 1: syntax error near unexpected token `;' sh: -c line 1: `((/usr/local/bin/rvplayer /tmp/MO34B2F4B209B0136.ram); rm /tmp/MO34B2F4B209B0136.ram )' I've established that Communicator always fires off helper apps using similar syntax by invoking /bin/sh. My question is, if I were to replace /bin/sh (which is a symlink to /bin/bash) with something else (ash, or an older version of bash, for example), will it break anything debian-wise? I realize that it might break other things that expect /bin/sh to really be bash, but frankly, such things probably deserve it for making that assumption. I had the same problem and ended up replacing sh with bash 1.14.5(1), which I had lying around from an old slackware backup or something. I don't think this is a good solution, though. I saw some old discussion claiming that this is a netscape bug, but after further research, I think it's really a bash bug... run from tcsh, the command bash -c '( ( echo foo ) ; echo bar )' produces the expected foo bar output, while bash -c '(( echo foo ) ; echo bar )' says bash: -c: line 1: missing closing `)' for arithmetic expression bash: -c: line 1: syntax error near unexpected token `;' bash: -c: line 1: `(( echo foo ) ; echo bar )' Under HPUX 9.05 and 10.20, OSF1 v3.0, IRIX 6.3, SunOS 5.5.1, SunOS 4.1.4, and AIX 2.3 (? AIX uname is weird), the command sh -c '(( echo foo ) ; echo bar )' produces foo bar so I think this is a bash bug, and a Debian bug if sh is to be a link to bash. Someone claimed earlier on debian-user that bash's behavior is correct, but it seems like a goal of bash is to be compatible with sh. I don't know what the sh specs say about such things, but this seems to be non-standard with respect to sh on every system I've tried. I'm cc'ing this to gnu.bash.bugs, so I'll mention that Debian GNU linux 1.3.1 exhibits this bug on i386 systems, which is bash v2.00.0(1), according to bash -version. bash v.1.14.5(1) does not have the problem. Of course, it's probably also a bug that this sh code is hard-compiled into netscape, but I think that's secondary. my $0.02 - M -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- Mark Monty Montague | [EMAIL PROTECTED] | I don't do Windows(tm) -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- DON'T PANIC! I'm a trained professional, and far more | *Why* question qualified to panic in this situation than you are. | authority? -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Matrox Millenium II vs. #9 Revolution 3D
Lawrence [EMAIL PROTECTED] writes: Alex Yukhimets wrote: I am about to get a new P II system and had to decide on many alterantives available. My current concern is what video card would suit me the best. I have no doubt that I would go with AGP one (even not for the sake of performance, but to save PCI slot :) and my current choice is between Matrox Millenium II AGP and Number Nine Revolution 3D AGP. Matrox seems to be the fastest under X (and free driver is already available from S.u.S.E.), Revolution 3D is faster under Windows. I do not intend to use Windows a lot, so the best bet would be Matrox, but I heard the opinion that in spite of the fact that it is the fastest, it's image quality is substantially worse than that of Number Nine cards. Could anyone confirm this? As luck would have it, I just installed debian on one of each of these (neither was mine :( ), and I didn't notice a problem with either in terms of image quality. Both could drive their monitors up to 1800x1440 at some flicker-free rate (but I'm very tolerant), and both work fairly well with alpha versions of XFree86. If you wanted to use XFree86, the Matrox is a better choice, since there is an SuSE server for that. The alpha servers work well enough that it's a good bet both cards will be well supported in the next XFree86 release. The Matrox code is down to mostly performance tuning, while the #9 still has a few glitches. The Matrox does packed-24bpp, which is nicer on memory but causes a few pixmap weirdnesses, though. And another thing, assuming I would have to use Accelerated X server with my card, what are the cons of the fact that server is libc5 compiled and my system will be libc6-based (of course, I would have to install libc5 runtime libraries also). I don't know about AccelX, but I believe that the Xfree86 compiles under both. I think the SuSE server was compiled with ibc5, since it runs under vanilla debian 1.3.1(bo). Both cards use WRAM. Millenium II uses 250Mhz RAMDAC and Revolution 3D uses 220Mhz, if you have a high end monitor, eg. Viewsonic 815, I don't think Revolution 3D can display 1600x1200 16-bit color at 85Hz refresh rate. The XFree86 run a max pixel clock of 230MHz on the Mil2, which runs 1600x1200 at 85Hz and 1800x1440 at 64 Hz, both of which look flicker-free to everyone I asked to look on a Sony 21 monitor. The i128 server used for the #9 does max out at 220MHz, so I had to doctor the 1800x1440, but it still seems fine to me on a Nokia 445X... this modeline ModeLine 1800x1440m 2201800 1896 2088 2392 1440 1441 1444 1490 +HSync +VSync looks fine to me, but should only be around 61Hz. The following modeline is a standard(ish) one for XFree86, which should let you do 1600x1200 @85Hz on both cards: # 1600x1200 @ 85 Hz, 105.77 kHz hsync Modeline 1600x1200 2201600 1616 1808 2080 1200 1204 1207 1244 +HSync +VSync In any case, both cards can do their 1800x1440 at 8,16,24, or 32bpp, so I'm not sure where you got the 16bpp number. I also should check the 250MHz number, but I'm not near my Matrox documentation; certainly, XFree86 assumes it's 230MHz max. Anyway, they're both good cards, and they should both be XFree86able in the next release, but if you want XFree support now, there is an SuSE server for the Mil2AGP but not (last time I checked) for the #9 Rev3d. $0.02 - M -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- Mark Monty Montague | [EMAIL PROTECTED] | I don't do Windows(tm) -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- DON'T PANIC! I'm a trained professional, and far more | *Why* question qualified to panic in this situation than you are. | authority? -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Debian on Aptiva
[EMAIL PROTECTED] (Tim Sailer) writes: Folks, Has anyone installed Debian on an IBM Aptiva L5H? I have a user here that wants to get one, but only if Linux will install and run with little hassle. My girlfriend has Debian running on an Aptiva, but it's not an L5H. I think it's an S9C. Anyway, the only probs were that it needed XFree v3.3.1, which isn't in bo, and that the stupid modem is one of those windows only modems, so it doesn't work with linux. - M -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- Mark Monty Montague | [EMAIL PROTECTED] | I don't do Windows(tm) -=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=+=-=-=-=-=-=-=-=-=-=-=-=- DON'T PANIC! I'm a trained professional, and far more | *Why* question qualified to panic in this situation than you are. | authority? -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .