profiling in libraries
Hi all, The problem I had with profiling turned out to be a problem with a profiling library. In other words, the module I want to profile was in a shared library. I compiled all of the object files with -p (-gp doesn't work, gave a warning about not supported), then created the shared lib as usual. I then compiled the application with -p and linked with the shared library. This doesn't work. Do I need to do something special to create a profiling shared lib? Thanks in advance for any assistance. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
GCC profiling
Hi, I can't seem to get profiling to work on hamm. Do I need to do something special to get profiling to work? Thanks in advance. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: GCC profiling
On 1 Oct 1998, Ole J. Tetlie wrote: *-Jimen Ching [EMAIL PROTECTED] | I can't seem to get profiling to work on hamm. Do I need to do | something special to get profiling to work? Thanks in advance. It should be just a matter of compiling and linking with -pg (and running gprof afterwards). What exactly have you done? I used -p. I believe I also tried -pg and got the same result. I'll try again with -pg, but I know -p doesn't work. I also installed the profiling libs. But it shouldn't matter, as profiling shouldn't depend on the libraries. This is with gcc package. I haven't tried with the egcs package yet. Did you try -pg and generate the gmon.out file? --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
PageUp/Dn doesn't work
Hi all, The page up and down keys don't work in netscape and nedit. I'm using the stock Debian 2.0 release. Anyone else have this problem? Note, these keys work fine in xterm. So I know X is working ok. How do I get netscape and nedit to work? --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: PPP connects every second attempt
On Sat, 22 Aug 1998, Jor-el wrote: # This file was generated by pppconfig. #You can edit the following lines # but please do not delete lines or the change the comments or you will # confuse pppconfig. noauth #pppconfig_noauth connect /usr/sbin/chat -e -v -f /etc/chatscripts/IBM #pppconfig_connect debug #pppconfig_debug ^^^ /dev/ttyS0 #pppconfig_dev 115200 #pppconfig_speed defaultroute #pppconfig_route noipdefault #pppconfig_ipdefault user 'XXX' #pppconfig_user # End of pppconfig controlled lines. You can add lines below here without # confusing pppconfig. I do not remember being asked if debugging should be enabled. In the future, I don't think this should be default. The debug info was taking up all of my disk space. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: [Help] resc1440.bin (kernel) crash
On Fri, 21 Aug 1998 [EMAIL PROTECTED] wrote: On Thu, 20 Aug 1998, Jimen Ching wrote: Instead, try to unpack base2_0.tgz directly onto your hosed system. IIRC it contains all the necessary devices and essential libs. Just boot the rescue floppy, switch to VC2 and mount your root filesystem on /mnt and Thanks for pointing this out. As I said before, I couldn't boot off the rescue floppy (of Debian 2.0, cheapbytes), so I used the boot/root floppies of Debian 1.1. But it worked. I repartition and started from scratch. It was easier this way. cpio -i -dbm /path/to/base2_0.tgz The Debian 1.1 boot/root disk had gzip and star (as another user suggested). I don't know how to use cpio either, so I use the star instead. dpkg --configure --pending I wouldn't be surprized if you got a lot of errors from that, but it Actually, this didn't do anything. It thought about it for a few minutes, and took me back to the prompt. ;-) dpkg -iBOR /cdrom/debian/dists/stable/main/binary-i386 That should just reinstall all packages that are registered to be installed on your system. Yes. Well, as an update. The kernel crash problem was not reproducible (easily) with a recompiled 2.0.34. But I did once experience the crash. I reset the computer and it didn't happen the second time. So I don't know what's wrong with kernel in the rescue disk. I also enabled initrd in the kernels I built. This did not fix the problem I had with the boot sequence where it tries to mount the root fs. The error message is something like: RAMDISK: Compressed file system found on block 0. Kernel panic: VFS: unable to mount root fs on device 01:00 I can ctrl-alt-del my way out of this. But enabling ramdisk and initrd didn't help. What is odd is that install.bat says the root device is /dev/ram, and the rescue disk says the root device is /dev/ramdisk0. Neither of these exist in my /dev directory. The Debian 1.1 boot/root does have /dev/ramdisk0, though. I'll try this and let you know if it is the problem. Now my system is back where it was, before all of this started. I do have a few questions about the new Debian version. 1. How do I set my hostname? Everywhere the hostname should appear, the string (none) is placed there instead, no qoutes. 2. Two files, fstab and resolv.conf were not created. I had to manually create these myself. What did I do wrong that they weren't created? 3. When I first got my system up, before installing a lot of optional package, the nfsd and mountd were started. I do not remember it asking me whether I want these services started. After I installed everything I wanted, it didn't start these services anymore. nfsiod is still a process though. Is this normal? 4. I get this error in my /var/log/xdm-errors System: `/usr/X11R6/lib/X11/xkb/xkbcomp -w 1 -R/usr/X11R6/lib/X11/xkb -xkm -m us -em1 The XKEYBOARD keymap compiler (xkbcomp) reports: -emp -eml Errors from xkbcomp are not fatal to the X server keymap/xfree86 compiled/xfree86.xkm' What is causing this? How can I stop it from writing this to xdm-errors? 5. I'm using fvwm2 as my window manager. I added the following to my .fvwm2/post.hook. Key Left C R Scroll -100 0 Key UpC R Scroll +0 -100 Key Right C R Scroll +100 +0 Key Down C R Scroll +0 +100 I tried a bunch of other letters besides 'R', and it still doesn't allow me to flip to another virtual screen. I have setup the FvwmPager to be 3 by 3 screens. So I should be able to switch to another screen. I can do it with the mouse, but I positioned the pager at the upper left hand corner and it is sometimes a long distance to move the cursor. How do I setup fvwm2 so I can use ctrl-[arrow] to move between the virtual screens? I should mention that I'm using the XKB extension for my keymap handling. I was having problems with my delete and backspace key with the older version of Debian. I thought maybe this new extension in XFree 3.2.2 would help. It didn't. Or maybe I did something wrong? I have the 101key keyboards. 6. One more thing with fvwm2. I added the following to .fvwm2/post.hook. PixmapPath /usr/X11R6/include/X11/pixmaps But I still do not see any of the pixmaps on the popup menu. How do I setup the configuration files so the pixmaps show up in the popup menus? 7. I installed the scalable fonts for X. Does anyone know how to access these fonts from nedit? For some reason, the font looks really junk in nedit. 8. I installed Xaw3dg. I also enable the scrollbar in xterm. For some reason, the scrollbar foreground and background is black. If it wasn't for the outline of the scrollbar, I wouldn't be able to tell where the scrollbar was located. How do I change the scrollbar colors of xterm? Thanks in advance for any help. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: [Help] resc1440.bin (kernel) crash
On Thu, 20 Aug 1998 [EMAIL PROTECTED] wrote: On Thu, 20 Aug 1998, Jimen Ching wrote: Please help me. I accidentally hosed my linux upgrade and need to reinstall from scratch. Hmm, how far did your system fall beyond the edge? Chances are that you may very well be able to fix things and complete the upgrade. Really bad. Here's the story if you're interested. After I upgraded to hamm (no problems encountered, good job), I noticed some obsolete packages, which I progressed to remove. I did this in steps so I don't harm my system. I did all of this in X, which may be my first mistake. Everything went fine until the last obsolete package, 'base'. I got the warning about _knowing what I was doing_ to proceed. I made the assumption that since the new version is installed, there's no problem with uninstalling this package -- my first big mistake. The system is still usable, but all of my devices in /dev just disappeared. I don't know what else is missing. I was able to retrieve most of the devices by running '/dev/MAKEDEV update'. Unfortunately, this did not recreate the pty's or the tty's. I couldn't find any manpage or document on how to create pty's (didn't look that hard yet, due to the next thing I did). My next big mistake was trying to login from the console. As you guessed it, no tty == no login. What got my goose was that, when I tried to return to my X session (ctrl-alt-f7), the X session disappeared. The only shell I had access to just disappeared from under me. I didn't think rebooting would help since I can't login anyway (no tty). Here's my third big mistake. I got the idea that if all I wanted was to get the pty/tty's back, just reinstall 'base' from my previous distribution. I did this by going back to my previous Debian version, 1.1 (I upgraded to 1.2 but that was from a cdrom I borrowed from a friend, which I no longer have). I proceeded to install the base system, which was stupid, because it installed a whole bunch of other stuff (libs, fs admin tools, etc). This hosed fsck, which means the kernel can't do the fsck at boot time, which causes me to dump into single usermode(?). At this point, not much works. Instead of reinstalling the 1.1 release, which I know would work. I thought I might has well install the 2.0, so I don't have any libc5 stuff on my system (which was what I wanted to remove in the first place). That's when I tried the rescue/root disk. From the above, you would think I'm no newbie. But somehow, I've must of executed the F0 0F instruction and my brain halted. ;-) I've never thought I could hosed my system this bad. :( Peculiar. Can you reproduce this? If so, you might want to consider I was able to run mount, which means I got access to the 2.0 cdrom (from cheapbytes). I tried all of the disk images, tecra, fast, etc (even the dos install method, yes I still have a FAT partition). All of them had the same problem. I finally tried the aic7xxx=no_reset(0), this did not produce an 'Aiee! Killing interrupt handler', but it died later with some 'scsi0:' message (froze hard). I did check LinuxHQ for the 2.0.35 release. There was a fix to the adaptec scsi cards. But no mention of AHA-2842. I'll compile 2.0.34 and try it just to make sure. But I think the problem is fixed. an empty filesystem after the kernel boots. With initrd support, it loads the rootfs image into ram before actually booting the kernel. On the image you should then set the rootdevice to /dev/ram0 (IIRC) because the kernel makefile makes it default to the rootdevice of the machine your building on. I'll compile with initrd and try again. Just a note, the install.bat file uses /dev/ram, not /dev/ram0. I did not see a /dev/ram in /dev. I do have some questions about the upgrade procedure, if you don't mind. 1. How does one remove all libc5 stuff off the system? I didn't see any option to remove all of the libs and stuff. 2. Why was the package 'base' from Debian 1.1 still listed as an installed package? Doesn't base-files or some other package replaced it? If there's a new version of the 'base' package, why is it listed under the obsolete section? Shouldn't it be under required/current? 3. What is the difference between the libc6 and libc6-dev? What is considered a development library? Is it needed to develop libc6 only, or any development at all, i.e. C++/X programming? Thanks for the quick response. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
[Help] resc1440.bin (kernel) crash
Hi all, Please help me. I accidentally hosed my linux upgrade and need to reinstall from scratch. But when I tried to boot from the rescue disk, the kernel crashed on me. I have an AHA-2842 SCSI controller. The kernel detected it and tries to reset something. The next thing I see is a message about dereferencing a null pointer and the standard register dump from the kernel. The last line says: Aiee! Killing interrupt handler. It then sits there and I have to press the reset button. I tried building a 2.0.35 kernel and copy it into the rescue disk (which is just a dos formated disk). The kernel now boots without problem, but when it got to the point of mounting the root fs, I got an error message about 'unable to open/access the root device 01:00'. I assume this is the ramdisk? I did compile the ramdisk support into the kernel, as well as dos fat and dosfs support, adaptec scsi driver, ppp, multi-device, and some others that I wanted. Is there something I need to make the kernel boot with a ramdisk? Or did I do something wrong when I copied the kernel to the floppy? I did run rdev.sh from the mounted floppy. Maybe I did that wrong? I should note, the kernel I built was compressed. I don't know if the 'linux' file on the rescue disk is a compressed kernel. It is 700+k, while my kernel is 400+k. I'll try build a 2.0.34 kernel and see if it is actually causing the problem. Any help in this matter is greatly appreciated. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: The real Debian.org
First off, I am not a developer of Debian packages and thus do not know how Bruce Perens manages the distribution. But basically I have two comments. 1. To DC, your first mistake was to post your request to all of the debian lists. The problem with this is that this is a developers' issue, not a users' issue. Thus all those responses sound like they were from users of debian, not the developers. In this case, user's opinions are irrelavent. Who cares if a user likes Bruce Perens as the project leader? They are not the one who must work with him. Developers like you and others are the ones who matter. What do other developers think? Are they having the same problem as you? 2. To the users, it's so pitiful that users of a distribution is so quick to kick out a DEVELOPER just because he disagrees with the leader. As a user, I must laugh at some of the criticisms. Spelling mistakes? Use of initials instead of the full name? These are reasons for ignoring a complaint? It is scary to find out that this type of group can decide who becomes a developer for debian and who doesn't. I should mention that this is not the first time there was disagreement in the Debian project. I remember in the beginning that Bruce and RMS also had a disagreement. Fortunately, that was resolved. I hope this one gets resolved as well. Unlike others, I do see some problems with the distribution and improvements can be made. Dave, I don't know what kind of contributions you make to debian. But if it is in the dpkg program, then I hope you can continue to improve it. If you do branch off into another distribution (great, yet another distribution), then dpkg is the first program you should improve. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
GCC/G++ with frepo
Is there a debian package of gcc with the repository mechanism patch from cygnus? I had a older version of gcc with the patch, 2.7.1. But I've upgraded to debian 1.2 and this version no longer works. So I need a new version. I would prefer a binary distribution, but source is fine too. --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
GCC frepo patch don't work.
Hi all, I've just installed debian linux 1.1 from Iconnect. I am working on a C++ project which depends heavily on the use of templates. Thus, I would like to use the -frepo patch from Cygnus. The installed gcc/g++ doesn't seem to have been patched correctly. At least the 'ld' is not working as described by Mumit Khan. I.e. the patched 'ld' was supposed to do a recompile/relink process. This was never done and I get lots of unresolved symbols. Mumit Khan sent me some binaries about a year ago for the Slackware distribution I used back then. I tried moving the files over to my current system, and things seem to work ok. But I am getting another error message from the patched 'ld'. This time, it says it can't find one of my header files. I.e. the header file which contained my class definition could not be found. This error is produced while in the recompile/relink phase (using Mumit's binaries). Has anyone ever tried the -frepo feature? It is actually patched in? Using the switch to g++ doesn't tell you anything since g++ ignores arguments it doesn't understand. This feature is very important. I can't continue my project without it. Please help, thanks in advance... --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]
GCC frepo (continue)
Oh, one more thing. I wanted to check whether I'm still on the debian-user list. I sent a help message to the request robot, and got back intro instructuions (subscribe/unsubscribe). I wanted to know what are the available commands to list the subscribed users. How do I get this info from the -request robot? Most other mailing lists (using listserv) has this info in the intro document. Also, since I couldn't get the help page, I don't really know if I'm subscribed. I guess this message will decide which. And one more thing, I have a wish list. It would be nice if the install procedure does not waste it's time testing each an every package in a directory and displaying it. I.e. dpkg -iGROEB goes through the entire directory and displaying the selecton status. This is slow and painful. I don't see why the dpkg program can't scan the entire directory and create a list of all installed packages and display it in sorted order. Then it would go through the list of selected packages remaining which are marked 'not installed' and just install those (in sorted order, at this point, the section really doesn't matter does it?). There are 450+ packages. I've only installed about 100 or so. I don't need dpkg to display the 300+ packages that I've marked as 'deselected'. Was this supposed to be a feature? --jc -- Jimen Ching (WH6BRR) [EMAIL PROTECTED] [EMAIL PROTECTED]