I've been planning the architecture and design goals of the first to-be-public release of SubLinux for about half a year.
SubLinux is the personal XFCE i586+ distro I built from source, and from which I am writing this. Generally, only me, family and friends are ever given copies of SubLinux. Source for SubLinux 1 has been lost.
Some background
SubLinux 1 was released in mid-2012, and was also an i586 distro. However, it had many severe flaws, and while functional, bugs were often encountered. It had a busybox/GNU hybrid userland, and while efficient, this frequently caused problems with software compilation. It was a small system, less than 2GB big once installed. It came with the Midori browser.
It was mostly used as a Fedora substitute on machines with no i686/CMOV support, and on thumbdrives as a portable system.
It used the syslinux bootloader instead of GRUB, which is used by most distributions. It did not support UEFI.
SubLinux 2 was released in early 2014, and has replaced Fedora on all of my desktops, excluding my server.
It's also i586-based, but has a true GNU userland. It shipped with XFCE 4.10, but has since been upgraded to 4.12, and has had numerous kernel upgrades. It has no package manager, was built from source, and provides both a PAE and non-PAE kernel. Unlike SubLinux 1, it has debugging symbols stripped from all binaries, yet provides both static and shared libraries of all system libraries, and all headers are present at installation, and thanks to the stability of the GNU userland, SubLinux 2 has proven to be an outstanding system for software development. It has a GCC 4.8.2 based toolchain. Like SubLinux 1, it also uses the syslinux bootloader. SubLinux 2 does not support UEFI. Systems wishing to use it must boot in "legacy" mode.
Moving forward
There are a handful of reasons I want to make SubLinux 3 a public, supported distribution.
* I am spoiled now that I've had a distribution that suits my tastes precisely, and I'd like to share it with the like-minded.
* It will showcase my Epoch Init System, my anti-systemd vaccine.
* There is a lack of distributions with the desired "vibe" of freedom, catering to power users, and enabling development.
Major design goals:
* Support for both i586 and x86_64, possibly in one single build, with the userland being primarily i586, but containing all libraries necessary to execute x86_64 code, and providing an optional x86_64 kernel. This would guarantee that no matter which system you boot it on, it will indeed boot. And, if the system supports x86_64, such applications will be able to run.
The compiler toolchain should also support both i586 and x86_64.
* Great for development, making it easy to deal with a large variety of programming languages.
* Vanilla software, only adding patches to software that are required to compile it. This includes the kernel.
* A complete graphical desktop, probably XFCE.
* Includes other power user tools, like terminator, isomaster, etc.
* No unnecessary duplicate applications in the default installation. One that runs in the CLI, one that runs in GUI. For example, mousepad will be installed, but leafpad will not.
* Efficient, with a focus on keeping system resource usage down to a minimum while still providing the aforementioned complete GUI and desktop.
* Annual release, with updates provided over the internet periodically.
SubLinux 3 will be a desktop-oriented distribution, with a full GUI. It will, again, include all development related files (excluding complete source) in each default package, so you'll never need to install some "-devel" package. And like SubLinux 2, it will include both static and shared libraries for every package. It will likely ship with XFCE as the default desktop, and will rely on (me-endorsed) community builds for other desktops. And that's where our next piece of the puzzle comes in.
packrat, a new package manager
I've been writing a package manager called packrat. It has one major design goal: Make it really easy to create packages.
As a consequence of this, it does not include dependency resolution support, making it similar to Slackware's package management scheme. You can, at present, in the unstable builds of packrat, create a package with one command. It uses tarballs for storage, and has metadata and sha1 checksum file verification. Packages will likely not support signing.
It will support downloading "action lists" from the internet for updates and deprecations, and will include a *built in* GTK frontend, eliminating the need for 3rd party tools for graphical package management.
That's about it for now guys. I'll let you know if I come up with any other brilliant ideas. :^)
-Subsentient
Laptops
Storage
Consoles
Categories
Articles
I've recently done some touchups to my SubStrings library, and have reminded myself of my own undying glory, er, I mean, come to the conclusion that it's useful enough to warrant me advertising it a little bit. I use it in a large percentage of my projects nowadays. I thought I'd showcase part of what makes it awesome to me.
Before we begin, there is one thing I must mention: SubStrings is only designed to work well with null terminated strings only. It's almost certainly unsafe to use it for anything else.
And, one last thing: This is indeed a C library, written in C89 and works in C. The OOP appearance, such as SubStrings.Length(), is function pointer trickery for cleanliness' sake. SubStrings is also a non-hosted library, meaning it has no dependencies, and can thus even be used in your bootloader's source code.
String copy and concatenation, truly safe
SubStrings copy and concatenation functions have precise bounds checking, and always result in a null terminated string, and, if the size parameter is given accurately, SubStrings never has buffer overflows. Let's illustrate.
#include "substrings/substrings.h"
void MyFunc(void)
{
char Array[1024];//So, the maximum string data copied will be sizeof Array - 1, so there is room for the '\0'.
SubStrings.Copy(Array, "My string is awesome!", sizeof Array);/*SubStrings.Cat's size parameter needs to be the *maximum capacity* of the destination. You don't need to subtract from
the max size each iteration of a loop. SubStrings knows how to do it.*/
SubStrings.Cat(Array, " And it merges really nice too!", sizeof Array);
}
strcpy and strcat() are in general, unsafe, and strncat and strncpy have differing and confusing behavior. SubStrings eliminates these problems with one consistent approach for concatenation and copy operations.
All the necessities provided
SubStrings also provides all the other functions you might want for basic string operations, including Length(), Compare(), NCompare() [analogous to strncmp()], Find(), and CFind() [Find a single character], and FindAnyOf() [Analogous to strchr()].
Find() and CFind() accept another, new argument, allowing you to directly request the N-th occurrence of the matching string/characters. There is also IsLowerS, IsLowerC, IsUpperC, and conversion functions like LowerS, LowerC, UpperS, and UpperC.
In general, functions ending in S deal with strings, and functions ending in C deal with single characters.
High level stuff is here too
Some of the cooler stuff is stuff you might expect to find in Python.
Stuff like StartsWith, EndsWith, Replace, Strip, Reverse, and StripLeadingChars and StripTrailingChars.
And some original ideas too
There are other functions, like Extract(), which pulls the string content that's in between two sequences.
Let's have an example.
#include "substrings/substrings.h"
int main(int argc, char **argv)
{
int Inc = 1;
char Buf[256];for (; Inc < argc; ++Inc)
{
if (SubStrings.StartsWith("--config=", argv[Inc]))
{
SubStrings.Extract(Buf, sizeof Buf, "=", NULL, argv[Inc]);
DoSomething(Buf);
}
}
}
What's happening here is that Extract is pulling the data that starts after the =, and since the next parameter is NULL, it reads on until the end of the string. This makes handling command line arguments marginally simpler.
There's other goodies, like SubStrings.CopyUntil(). Let's take a look.
#include "substrings/substrings.h"
void MyFunc(void)
{
const char *const String = "Wibble[END]Nurble[END]Aburble[END]Farts";
const char *Iter = String;
char Buf[256];
while (SubStrings.CopyUntil(Buf, sizeof Buf, &Iter, "[END]", true))
{
puts(Buf);
}
}
This produces the output:
Wibble
Nurble
Aburble
Farts
I actually find myself using CopyUntil and its sister function CopyUntilC quite often. Then there's SubStrings.Line.GetLine(), which is a specialized CopyUntil that helps with processing multi-line C strings.
Lastly, there's another useful function, Split().
#include "substrings/substrings.h"
void MyFunc(void)
{
const char *String = "Gerbil|Wibble";
char One[256], Two[256];SubStrings.Split(One, Two, "|", String, SPLIT_HALFONE);
}
Now, One[] contains "Gerbil|" and Two[] contains "Wibble". You can specify to discard the split tokens, or put them in half one or two. The options are SPLIT_NOKEEP, SPLIT_HALFONE, and SPLIT_HALFTWO. Because Split() doesn't ask for buffer sizes for the sake of convenience, the way to do it safely is to make sure that both One and Two will be able to hold the entire length of String, if needed.
This library is getting more touchups. You can find the github here, and the SubStrings homepage here.
Thoughts, ideas or suggestions? Let me know.
Expounding on what I feel ban-worthy as linked in a recent story I subbed...
These are my opinions not site policy. That said, they are also my minimum requirements to remain on staff.
Over-the-top spam: This means dozens of spam comments to a single story, automated or not. Anything less can and should be dealt with by simply modding the comments as Spam.
Gross/repeated illegal activity: Linking to copyrighted works and other illegality of the minor variety that we notice should be resolved by editing the comment in the database and letting the user know why their comment was edited in a reply. Bans should be reserved for things such as illegal and credible threats or multiple instances of minor illegal activity that was not ceased when notified that it should be.
Opinions, truly held or of a trollish nature: Absolutely never should this be criteria for a ban.
The meaning of "site bans": Banning an account or IP address from posting to the website. I don't have as much issue with IRC bans, it's a secondary means of communication for us not our primary one.
Any further clarification or other questions, feel free to ask.
Didn't watch this time. Ben Carson sure is popular isn't he?
Wikipedia
Fireworks Fly In Unruly Third GOP Presidential Debate
Four Republicans Battle For Attention At Undercard Debate
PolitiFact
Bizarre, campy song explains China's 13th 5-year plan
Song (low information content)
Uses of Force Abroad 1798-2015, and More from CRS
A newly updated tabulation of U.S. military actions has been prepared by the Congressional Research Service, up to and including the October 14, 2015 deployment of 90 U.S. troops to Cameroon. The CRS listing does not include covert actions, disaster relief operations or training exercises. See Instances of Use of United States Armed Forces Abroad, 1798-2015, October 15, 2015.
[other CRS reports listed at source]
The PDF is 38 pages.
I've just installed Windows 10 on an i3 laptop and am looking for suggestions on how to do an in depth traffic analysis of Windows 10 after the "stop spying" program has been used to see if its possible to "de-fang" Windows 10. The requirements will be 1.- It has to be software based as I no longer have any hardware firewalls capable of logging IP addresses, 2.- If its Linux based I'm gonna need a link to a how-to as I've never used Linux for networking, only web servers, and 3.- If it works off a stick or live CD so much the better, I really don't want to have to reinstall the OS on the test box afterwards.
Anyway I figure I'll disable the wireless and run the ethernet into a second box with 2 ethernet ports as a pass-through to do the logging. I figure it will have some sort of net connection test so merely plugging into another box won't work, but by using the second box as a bridge we should be able to capture exactly what is going out and see if its truly possible to kill Win 10 spying. The pass through box is a Phenom I quad with 4GB of RAM so it should have no issues running whatever we need to run, and all results will be published here.
So other than Wireshark any suggestions?
Previously:
The candidates participating in the debate are Lincoln Chafee, Hillary Clinton, Martin O'Malley, Bernie Sanders, and Jim Webb. Larry Lessig did not make the cut. CNN coverage begins at 8:30 PM EDT, the debate begins at 9 PM EDT and ends at 11 PM EDT.
Links:
The first debate is scheduled for October 13, 2015, at the Wynn hotel in Las Vegas, beginning at 9 p.m. Eastern time. It will air on CNN, and will also be broadcast on radio by Westwood One [AUDIO-only should be available online via CBS Radio]. Anderson Cooper will be the moderator of the debate, with Dana Bash and Juan Carlos Lopez asking additional questions and Don Lemon presenting questions submitted by voters via Facebook.
To be invited to the debate, a candidate must have achieved an average of at least 1% in three recognized national polls released between August 1 and October 10. In addition, a candidate must either have filed a statement of candidacy with the Federal Election Commission or declare that one will be filed by October 14, the day after the debate. The latter criterion would accommodate Vice President Joe Biden if he decided to enter the presidential race as late as the day of the debate.
Should Vice President Biden decide to enter the race and take part in the debate, there would be a podium placed on the stage for him as well.