Forgive me if I'm incohernet - I ended up hacking all night on a VAX and doing micro arch and other home work 8^( (Hey, it runs Unix, not VMS so it's a real computer) In article <kvn3k1INN6mv@almaak.usc.edu> ajayshah@almaak.usc.edu (Ajay Shah) writes:The binaries are almost ALL gnu utilities. File utils, shell utils, binutils, the C compiler, debugger, standard make, last I checked awk, sed, tar, etc. We don't have dbx, we do have gdb. And you use gdb in text mode. And if you are extremely sick, there is also an EMACS interface to gdb. Try it - you'll like it . GDB has more feeping creatures than DBX.... POSIX, leaning towards SYSV, but with BSD features like the setre{g,u}id syscalls. No. The separation of buffercache / usermemory is done by the "classical" method. No. But VFS is implimented, and you can add something like FFS or even beter LFS. Also, more important than the file system speed is a defficiency in the buffercache / disk driver code. Basically, if your disk is not track buffered, you will get no performance at all. Write calls are passed to the device driver in 1 block chunks - ie 1K. If subsequent blocks are requested, you get to wait another revolution until the next one comes along. People are working on this. Fine, if you set the km option in the termcap. Full emacs is several megs. Definately not. Life without sendmail is a breeze.... =8^) Linux has users, and you're likely to get yourself in trouble if you don't have a separate "you" and root account. Linux doesn't do anything about specific vendors motherboards. Basically, it definately won't run on microchannel, and some flakey motherboards do not work. You can mmap(2) the frame buffer into user space, and program it like you would in real mode. X is being ported, and there has been some talk (I don't know how serious) about getting MGR / a textronic / other graphics terminal emulation in the kernel. TeX yes, previewer no. /dev/ttys - you talk to it like any other character device, and can do non-blocking IO if you fcntl it or use select(2). Linux DOES NOT support hard flow control, and in my experience drops characters at 19.2K (386-33). People are rectifing the hard flow problem, and adding 16550 support. The company that brings you win3.1, a $50 bug fix to win 3.0. Go figure.
| Eric Paris | TALPA - a threat model? well sorta. |
| Vladislav Bolkhovitin | Re: Integration of SCST in the mainstream Linux kernel |
| Andrew Morton | Re: [BUILD-FAILURE] 2.6.26-rc8-mm1 - x86 - __ptep_modify_prot_start() missing |
| Bodo Eggert | Re: [PATCH] x86: provide a DMI based port 0x80 I/O delay override. |
git: | |
| Pavel Roskin | Implementing branch attributes in git config |
| Jon Smirl | Importing Mozilla CVS into git |
| Jon Smirl | Figured out how to get Mozilla into git |
| Jakub Narebski | Re: VCS comparison table |
| Richard Stallman | Real men don't attack straw men |
| Adam Getchell | Re: About Xen: maybe a reiterative question but .. |
| carlopmart | About Xen: maybe a reiterative question but .. |
| Bertram Scharpf | First install: Grub doesn't find partitions |
| Waskiewicz Jr, Peter P | RE: [PATCH 2/3][NET_BATCH] net core use batching |
| Jeff Kirsher | [PATCH 1/3] e1000e: add support for the 82567LM-4 device |
| Ayaz Abdulla | [PATCH] forcedeth: msi interrupts |
| Corey Hickey | [PATCH 10/10] Use nested compat attributes to pass parameters. |
