Welcome! This *was* my blog at some point.
This blog is a wiki. Feel free to edit any page.
This website is amazing. If you make an edit, then you are amazing as well.
A few people think that some of the stuff I write is offensive. This is not true. If you find it offensive, then there is a problem with your cognitive process.
Non-instructional posts are dedicated to people who can read more than 2 sentences in a row.
I decided to post about interesting things I notice.
Here's one of the most interesting comments on a bug report I've ever seen:
https://rt.perl.org/Ticket/Display.html?id=128462#txn-1494756
This is old news, but I haven't updated this blog in a while. In any case, here's a good read:
Please check out the whole article, but here is a TL;DR anyway:
“If Telegram wants to prove that their protocol is better than this absolute garbage protocol, then I challenge them to publish the plaintext of Alice’s message. […] Here we have a messaging protocol that employs the NSA backdoored random number generator (Dual_EC_DRBG), weak public key cryptography (896bit RSA, no padding, no signatures, no authenticity), the worst cryptographic hash function possible as a KDF (MD2), and XOR as a cipher.”
Just check this out: https://felixmilea.com/
These +
signs start spinning when the mouse goes over them.
Want to make it even cooler? Here is an extra tip! Use some unicode character, like 💩
. How cool is that?!
… Ehh…
Not only it is bloat, it also lags a lot on my laptop.
In case the author does the right thing and removes this crap, here is a screenshot (for historical purposes):
It reminds me of snowflakes from the days when stupid javascript effects were considered cool.
Sometimes the background on your website tells more about you than the content. At least, it may give a bad impression.
cc1
, cc1plus
and lto1
binaries so that it accepts any xclm
binary.These instructions are mostly based on this page: http://www.jubatian.com/articles/turning-on-optimizations-in-microchips-xc32/
First you have to download the source of the compiler. Go to http://www.microchip.com/mplab/compilers, click on “Downloads Archive”, scroll down and find “Source Archives”. There you can choose which version to download. At the time of writing the latest link is for xc32-v1.40.
# ⚠ change these variables according to your setup.
installpath="$HOME/microshit" # this is where the compiler will be installed
pic32mxpath='/opt/microchip/xc32/v1.40/pic32mx/' # assuming that you already have installed xc32 from microchip
mkdir -- "$installpath"
export CPPFLAGS="-I$installpath/usr/include"
export LDFLAGS="-L$installpath/usr/lib"
# gmp
mkdir -p gmp_build && cd gmp_build
../gmp/configure "--prefix=$installpath/usr" --disable-shared --enable-static --enable-cxx
make
make install
cd ..
# ppl
mkdir -p ppl_build && cd ppl_build
../ppl/configure "--prefix=$installpath/usr" --disable-shared --enable-static
make
make install
cd ..
Unfortunately I get this error during make:
../../ppl/src/Generator_System.defs.hh:253:7: error: ‘ptrdiff_t’ was not declared in this scope
It seems like it does not compile under newer gcc versions.
This leads me to this page: https://gcc.gnu.org/gcc-4.6/porting_to.html
It says that you should #include <cstddef>
. You have to do that with these files:
Just add #include <cstddef>
and save. Now run *make* again.
Now you'll probably see this error:
../../ppl/src/Interval.defs.hh:451:77: error: ‘f_info’ was not declared in this scope, and no declarations were found by argument-dependent lookup at the point of instantiation [-fpermissive]
Result rl = Boundary_NS::assign(UPPER, upper(), info(), UPPER, x, f_info(x, open));
It looks like it can be solved by using -fpermissive flag.
Try rerunning make like this:
make CXXFLAGS='-g -O2 -frounding-math -W -Wall -fpermissive'
This should probably work.
# cloog
mkdir -p cloog_build && cd cloog_build
../cloog/configure "--prefix=$installpath/usr" --disable-shared --enable-static "--with-ppl=$installpath/usr"
make
make install
cd ..
# libelf
mkdir -p libelf_build && cd libelf_build
../libelf/configure "--prefix=$installpath/usr" --disable-shared
make
make install
cd ..
# zlib
cd zlib
./configure "--prefix=$installpath/usr"
make
make install
cd ..
# binutils
mkdir -p binutils_build && cd binutils_build
../binutils/configure "--prefix=$installpath/usr" --disable-shared --enable-static --target=pic32mx --with-dwarf2
make
make install
cd ..
You will get this error:
../../binutils/bfd/cpu-pic32.c:32:55: fatal error: ../../c30_resource/src/xc32/resource_info.h: No such file or directory
OK this is a real issue. It seems like they forgot to include the file. See this link: http://www.microchip.com/forums/m897811.aspx
# Microchip libraries
cp -a -- "$pic32mxpath" "$installpath/usr/pic32mx"
# gcc
mkdir -p gcc_build && cd gcc_build
../gcc/configure "--prefix=$installpath/usr" --disable-shared --enable-static --target=pic32mx --enable-languages=c,c++,lto --disable-objc-gc --enable-lto --with-host-libstdcxx=-lstdc++ --enable-multilib --with-dwarf2 --disable-sjlj-exceptions "--with-sysroot=$installpath/usr/pic32mx"
make
make install
After all these efforts you end up with a half-broken build.
Just edit the SHA sum in cc1
, cc1plus
and lto1
binaries and you're done. No need to compile anything.
There is a chance that the build process will become easier once Microchip updates their compiler to the latest GCC version. Latest XC compiler version is 1.40 (2015-06-17), which is pretty old. The reason for such delay is probably that they are unable to compile the thing with the newer GCC versions. Ha-ha. Good luck, Microchip.
Lately there are only negative posts on this blog. Maybe there is a way to fix that?
Look, I've found something positive!
…
Just kidding. Dumb people are attacking again.
Often you can see people screaming out terms like “big data”. It is usually accompanied with enormous amounts of saliva coming out of their mouths. Like they believe that marketing terms like “big data” are good enough reasons to become brain-dead epileptics (and like that is going to help someone).
Interestingly, it is also associated with another problem. Programmers tend to think that problems they solve are very difficult, when in reality most of these tasks are dead simple. We are just not smart enough yet to come up with an elegant solution.
Anyway, see this amazing article about command line tools being 235 times faster than hadoop cluster.
What's the lesson to learn here? A few gigabytes is NOT big data. It is not big data if it fits into your RAM or if you can store it on your HDD. Trying to use “big data” tools to process such amounts is just plain stupid.
And yes, most of these times you don't even need a database management system. Plain files will do the trick.
Be conscious and learn to use command line tools.
I don't know who is the author of this priceless piece of art but I want to share it anyway (if you know the source then please let me know).
Anyway, so why would I mention it?
Actually, I expressed this opinion more than one year ago as a comment here. After a couple of additional encounters with this nonsense I decided to extract it to this blog.
Here are some of the issues:
This just proves that there is no point to have authorities on such websites. Just let the peer review to do its thing.
StackExchange is rather new, all we have to do is wait a few years until it is filled with useless bullshit. Then the problem will fix itself (stackoverflow will die and we will get back to wikis, FAQs, docs, IRC and other stuff that actually works in a long run).
You can also read my other comment.
Be conscious and try not to support Stackoverflow.
Oh these moments: http://superuser.com/a/787763/521612 (read the comment!)
UPDATE: I think it was removed…
-- AlexDaniel 2016-01-24 02:10 UTC
Agree. I've never liked SO and even go so far as to include
-site:stackoverflow.com in my Google searches so I don't have to wade through the massive amounts of garbage and misinformation that is so prevalent there.
-- CoderDude 2016-07-19 18:15 UTC
The decline of Stack Overflow - How trolls have taken over your favorite programming Q&A site :
https://hackernoon.com/the-decline-of-stack-overflow-7cb69faa575d
-- johnslegers 2016-12-06 22:02 UTC
As you might know, changing the name from Internet Explorer to Edge does not solve the problem.
Internet Explorer sucked and it still sucks. Edge is just a continuation of that, it is still far behind other browsers. If you are a web developer, you probably know this, but if not, then check out http://caniuse.com.
The funny thing is that there is a platform where you can make suggestions about further development of various Microsoft… tools (which reminds me of this, by the way). Can you guess what was the best suggestion out there?
Since NIH syndrome is often a cause of brain death, they deleted it. Every next day there was a new one that basically said the same thing – Stop Internet Explorer Development, maybe just slightly reworded. And yes, they kept deleting these posts! This suggestion was so popular that it was getting to the top very quickly, sometimes it was reaching the top position before they were removing it.
Yea, Microsoft, just shove your shit into everyone's ass and shut everyone up.
But here is another good one: Use Blink (or other open rendering engine) and start contributing code. This totally makes sense, right? If they cannot do it right (and they definitely can't, not even in foreseeable future), then why bother with creating your own engine? Just take one that works and slap your surveillance features on top, you'll get exactly the same thing but at least it will be on par with useful browsers.
Look at the answer! Just imagine that, someone is getting paid for writing shitty responses like this one! “We also understand and value the importance about being more open with our engine.” – what a bullshit. Microsoft understands the importance of being open? Look, the whole Microsoft is just a shitball that keeps rolling, you can't say that it values something because there is no ideology behind it, and personal opinion of somebody does not play any important role.
You can even take that answer and put it into history books, just for our next generations to laugh.
Be conscious and stop supporting Internet Explorer/Edge.
“Wirth's law is a computing adage which states that software is getting slower more rapidly than hardware becomes faster.”
Just open your eyes and make your code less shitty. Maybe these 5 software frameworks that you smashed into your hello world project is a little bit too much?
Sometimes I stumble upon pages with stuff that looks like this:
These are actually links.
I'm not sure who did that (most likely the author of this website just used some library or engine which already had this), but this problem is very common nowadays.
So, let me help you to understand what these circles actually stand for:
I have no idea what were they thinking, perhaps they are not thinking at all (brain death is a common reason to become a UI designer). When you are creating an interface, there are several things you have to think about:
And much more, actually. But these ones are actually essential if you want your website not to suck. Color-blind people might not see your flat buttons in the way you are seeing it. They might confuse it with decorative elements or even not see it at all.
Blind people might have better luck, since some screen readers will read the link out loud.
And it does not matter if you style your circles differently or not. Text-based browsers don't understand this shit at all, because there is no text to represent these circles. These circles, however, are rendered… as circles! At least that's how it looks like in w3m:
You can actually follow these links! Either place your cursor into the right spot or use tab to get to it.
Although I prefer textual buttons, I'm OK with all kinds of pictograms or other visual elements.
But this is just complete nonsense, I have to hover over every button to find one that I actually need. I don't even know which links are there before I hover over all of the links.
This is just nuts.
The first question that pops up in your mind when you see this – what is going to happen if I hover over these buttons? Maybe the icon will appear?
This does not really help to find the right link you are looking for, since you still have to hover over these buttons. But it could make your life a little bit easier, right?
This is what happens when you put your mouse over one of the buttons:
Now guess what happens when you hover over other buttons.
…
Yes, they become yellow.
Please design your websites consciously.
It seems like there was a bug when the icons were not being loaded if you used https. That's good. It means that the author did not do that shit on purpose – it's just a bug! This, however, does not really make it better. See HTML Hell section “menus made entirely from image maps”. You see, today even text-based browsers can display images. BUT NO, NOW WE DO ICONS WITH CSS. And what this CSS does? Sure, it loads a custom font! Nothing has changed… In other words, if every icon on that website had a little text near to it, then there would be no such blog post.
“You really should research something like that before writing an entire article on bashing someone.” says the author. And he is right! He is so right that I've even deleted this post. But it is a wiki, so someone reverted it back. I find this post funny so I would not delete it again. :)
1.5 years later and the problem is still not fixed. That being said, if you have an ad blocker (and you're retarded if you don't), then you will see “AROUND THE WEB” without any circles. Nothing around the web.
Hahaha
-- AlexSchroeder 2015-09-01 12:00 UTC
I've always felt that we need to introduce more changes to spelling, but I have never imagined that there are so many systems that were never accepted.
First, I stumbled upon this: http://everything2.com/title/spelling+reform.
Then I felt that the ideas behind Wijk's Regularized English are very good, so I found this page: Notes on Wijk's Regularized English, and more generally, this page: List of Spelling Systems.
Just look at that list! Impressive.
Some of these are more than 100 years old...
Another interesting finding is Common Errors in English Usage.