Reply With QuoteComtux do you have an iso of what is available so far? I could put some time into getting the Stabalizer started.
Wow! Sounds nice. I have a lot of really cool (at least I think so) ideas for my own little distro. Of course, as mentioned previously, I don't actually have time to create my own from scratch. This'll be the next best thing.
A lot of my ideas center around making the distro more user-friendly and, as such, deal heavily with the window manager. One of my primary ideas is about the use of the desktop.
I think the desktop should be context sensitive and super-smart. Users can configure the desktop to do certain things with certain files drag&dropped onto it. For instance:
URL links from a browser can cause it to download that file, store the url in an easy to access "hot link" bar, or some other action.
Images can cause it to open the image in a viewer, and editor, or place it in an image folder.
Various selections of text will cause the desktop to parse the selection and determine the appropriate action:
- If it's an address type selection, add it as an entry (or update an entry) in the address book.
- Basic text will cause it to open a text editor with the selected text.
Any item dragged to the desktop while the cntrl key is held down will be sent to the trash.
I have fleshed it out more than this, but the basis of it is that any file will be processed using a set of defined actions/parameters. Users can add/edit these parameters at will. Maybe also add in Bayesian statistics to have the desktop get "smarter".
I also have many other ideas but, again, they center around making the whole experience easier and faster for the user.
The only thing I will miss from making my private distro, I think, is the name. When this thing really takes off, Comtux will be in the lime light.
Comtux do you have an iso of what is available so far? I could put some time into getting the Stabalizer started.
Sure, but I would like to make a disclaimer, this is not my original work, I took someones brilliant Idea from the GEntoo forums and modified it to support ACPI battery reading and take out a few superfluous things that I didnt need.Originally Posted by jro";p="4733
Hey Zen, mind posting your PS1 env var? I am interested to see how you did it.
First I use the following script to create the prompt, This is done so you can actually use more complex bash scripting and have the ability to edit it all quickly without changing a million .bashrc files.
Then I add the following line to my bashrc filesCode:# SET THIS TO 0 if you are not using Gentoo Linux (http://www.gentoo.org) DIST_GENTOO=1 batt=`/usr/bin/battery | cut -d " " -f 2` hostnam=$(hostname) unset temp function prompt_command { local EXIT_CODE=$? # Getting last command exit code. TERMWIDTH=$(tput cols) # Getting Term width. #awful way to get current emerge(if you have sandbox documentation email me :) ) local C_EMERGE=`ps aux | grep sandbox | cut -d ']' -f 1 | cut -d '[' -f 2 | grep -v grep` # n the term wide calculation. (length of your prompt PS1) # Will be changed soon... usernam=$(whoami) local temp="$BLUEGREEN[ $usernam $BLUEGREEN]=[$GREY $hostnam $BLUEGREEN]$BLUEGREEN=[$GREY bg :$WHITE $JOBS $BLUEGREEN]=--$BLUEGREEN--=[$GREY $(date +%D) $(date +%T) $BLUEGREEN]" # Getting size of unused terminal space :) ==> for filling with $fill characters. local WHITE="\[\033[1;37m\]" local NO_COLOR="\[\033[0m\]" local RED="\[\033[1;31m\]" local BLUEGREEN="\[\033[0;36m\]" if [ ${#C_EMERGE} == 0 ] then C_DISK=$(df -h "$PWD" | grep / | awk '{print $4}') C_DISKPER=$(df -h "$PWD" | grep / | awk '{print $5}' | sed "s:%::") let fillsize=${TERMWIDTH}-${#temp}-${#C_DISK}-3 if [ $C_DISKPER != "-" ] then if (( $C_DISKPER <= 65 )) then C_DISK=$BLUE_GREEN$C_DISK$NO_COLOR elif (( $C_DISKPER <= 85 )) then C_DISK=$WHITE$C_DISK$NO_COLOR else C_DISK=$RED$C_DISK$NO_COLOR fi else C_DISK=$WHITE"None"$NO_COLOR let fillsize=${TERMWIDTH}-${#temp}-7 fi else let fillsize=${TERMWIDTH}-${#temp}-${#C_EMERGE}-3 fi if [ "$fillsize" -gt "0" ] then fill=" " # Just used for filling with this line when term wide calculation. Verify C_EMERGE !!! if [ $DIST_GENTOO != "1" ] then let fillsize=${TERMWIDTH}-${#temp} fill="${fill:0:${fillsize}}" # non Gentoo distribution. else if [ ${#C_EMERGE} == 0 ] then fill="${fill:0:${fillsize}/2} $C_DISK ${fill:0:${fillsize}/2}" # DS else fill="${fill:0:${fillsize}/2} $C_EMERGE ${fill:0:${fillsize}/2}" # Gentoo fi fi newPWD="`echo ${PWD} | sed s:${HOME}:~:`" fi # prompt path size adaptor ... need to be changed... # doesn't work as expected now ... if [ "$fillsize" -lt "0" ] then fill="" # truncate PATH with "..." let cut=3-${fillsize} newPWD="...${PWD:${cut}}" fi # Calling the personnal prompt fct. lycos } PROMPT_COMMAND=prompt_command function lycos { # dirty ansi escape color codes... change to terminfo soon... local WHITE="\[\033[1;37m\]" local NO_COLOR="\[\033[0m\]" local BLUEGREEN="\[\033[0;36m\]" local GREY="\[\033[0;37m\]" local RED="\[\033[1;31m\]" local PATH_COLOR=$NO_COLOR local APM_PATH="/bin/apm" local JOBS="$(jobs | wc -l | awk '{ print $1 }')" # Getting current nb of background jobs. # Detecting superuser with visible color :) if [ $usernam != "root" ] then usernam="$GREY\u" else usernam="$RED\u" fi # Add a bold red number for exitcode if its nonzero if [ $EXIT_CODE -ne 0 ] then EXIT_STR="$RED$EXIT_CODE" else EXIT_STR="$WHITE$EXIT_CODE" fi if [ -f $APM_PATH ] then local POWER_APM="$BLUEGREEN[$WHITE$(apm | sed -e "s/.*: \([1-9][0-9]*\)%/\1/" | tr -d " ")$BLUEGREEN]" # Getting APM power status. else local POWER_APM="" fi # Directory color path. if [ -O "$PWD" ] then PATH_COLOR=$GREY #if owner. elif [ -w "$PWD" ] then PATH_COLOR=$BLUEGREEN #if write permission. else PATH_COLOR=$RED #if denied for write. fi # Prompt line... make your prompt here ... (Need to be serialized.) PS1="$BLUEGREEN[ $usernam $BLUEGREEN]=[$GREY $hostnam $BLUEGREEN]$BLUEGREEN=[$GREY bg :$WHITE $JOBS $BLUEGREEN]=-- ${fill}$BLUEGREEN--=[$GREY $(date +%D) $(date +%T) $BLUEGREEN]\n\ $BLUEGREEN[$GREY $PATH_COLOR$newPWD $BLUEGREEN]-[ $RED$batt $BLUEGREEN]:$WHITE " PS2="$NO_COLOR>" PS3="$NO_COLOR>" PS4="$NO_COLOR>" }
and there you have it.Code:[ -f /path/to/prompt.bash ] && source /path/to/prompt.bash
Sure maccorin let me finash the Lfs 6.0 build & the doc's
for the installer and ill up load a copy of the iso some where right now the iso is like 25 megs and doesn't do much but sit there and look pretty ill try to have the base done in a day or 2 so that way it will install a base system and you can tear it a part.
Buy the way question should i keep the base as 1 binary package or break them up into there own seperate packages i mean with the base you have no other option you have to install it so what's the point of haveing the base broke up into seperate packages?
Wenn Sie Spaß meines Englisch mich Willensfuckingtötung Sie bilden.
i would break it up so updates are easier later on.
I've also been thinking about the implementation of a system stabalizer, and have a few thoughts.
So far the cleanest implementation that I have thought of would be a small kernel patch that hooks into do_coredump and logs fatal signals and some information to a special device, we could eventually provide ioctl's to change/setup what info get's logged. Then as things should be keep the real work in userspace, have a daemon that watches the device file and embeds a scripting engine of some sort, the user could provide callbacks for what to do upon certain signals. We would of course provide defaults.
Any thoughts on that?
Maccorin that sounds like what i had in mind but instead of a kernel patch i would have built a module that did the same thing.
But then again you sound like you really know what your talking about my kernel
experience is extremely limited.
Pugio those are some very fresh ideas one's that i have not even considered id like to hear some more on how you would go about such thing's.
the main kernel-space functionality could be easily put into a module, but we need to hook into somewhere to be notified of the signals that mattered, and do_coredump seems to be the perfect place, but then again, there is a new subsystem in 2.6.10 for userspace programs to listen for events, i wonder what it would take to add the events we need to that... hmm.... /me reads some code tonight.
Maccorin is tooooo coool let me know.
I ran in to a problem with LFS 6.0 Turns out you can't compile glibc kernel 2.6 optimized with a host 2.4 kernel.
Ok maccorin i am on chapter 6.11 compileing Glibc-2.3.4
the only problem as far as upgradeing packages i see is with GCC Glibc & Binutils everything should upgrade just fine when the time come's.
Upgradeing GCC Glibc & Binutils might have to be made into some special type of packages then again i am not sure.
Some have told me that i shouldn't have a problem as long
as i upgrade like this Glibc-2.3.4 --> to Glibc-2.3.5 now from what i have heard upgradeing like this.
Glibc-2.3.4 --> Glibc-2.4.0 will not work the same goes for GCC & Binutils.
Everything so far is turning out really good.
Oh i am useing kernel 2.6.8.1 as the default kernel
for now.
Wenn Sie Spaß meines Englisch mich Willensfuckingtötung Sie bilden.
Posting Permissions
Bookmarks