pthread_exit() in C++

Today I ran into an interesting problem that I would like to share. I am working on multi-threaded code in C++. Here’s what happened.

I started a thread that looks like this:

try {
} catch (...) {
    std::cout << "Got unknown exception" << std::endl;

The do_something() routine eventually called pthread_exit().

Once I ran this piece of code, I instantly got an unknown exception notification. Long story short, here’s what I found out.

When calling pthread_exit() in C++, it has to destruct all objects that has been created on stack. This process called stack unwinding and this is exactly what happens when you throw an exception. pthread_exit() utilizes this feature of C++ to cleanup before shutting down the thread for good.

To do that pthread_exit() throws some obscure exception and catches it right before ditching the thread. This way it cleans up all objects nicely. On the other hand, catching … becomes impossible.

My next programming language

This week-end I’ve been playing with various version control systems. Until now, I’ve been doing all my home codings with subversion. I’ve written about bazaar in the past, but it seems to me that bazaar isn’t going anywhere and it busts any piece of motivation that I have to continue writing about it.

Version control that I did try is git. This is a very popular version control system and for a good reason. Comparing git and subversion brought me to a conclusion that git is really a very consequence of how things work in the world. Thing about git is that it is a distributed version control while subversion is not. You can make a distributed version of subversion, but it won’t be subversion anymore.

Distributivity is one thing, but there are more. Take the standard trunk/branches/tags layout that you have to create in subversion – version control could do it for you, as git/bazaar/mercurial do. At first, after working with CVS for some time, having an option to have non-standard layout seemed cool for some time. But then it appeared completely useless.

This brings a notion of evolution into version control systems and I bet there’s similar process with programming languages. I think we can already starting drawing a programming language that will replace Python.

Yep, Python isn’t perfect. I thought I’d compile a list of things that shall be different in ought to be programming language that comes after Python.

Read the rest of this entry »

Call a constructor or allocate an object in-place

Since I joined Dell, my main field of research and work has somewhat changed. Now I am mostly working with C++ and file-systems. This world is not entirely new to me, but apparently I have a lot of stuff to learn.

Today I’d like to talk about one nice trick that I learned few days ago.

When working with large software systems, memory management becomes an imperative. In C, you can easily allocate a large chunk of memory and allocate structure right on that buffer. This is by far more difficult in C++, because compiler has to call consturctor.

Apparently, you can, in a way, directly call object’s constructor . I.e. you can allocate an object, on specified memory region, without actually allocating this region.

This is how you do it.

char* s = new char[1024];

SomeClass* p = new (s) SomeClass;

First new operator just allocates 1024 bytes. This is good old allocation as we know it. Note the special syntax of the second new operator. It allocates the new object on memory specified in brackets. Basically, this calls SomeClass’s constructor using s as storage.

One thing that I don’t know how to do is how to call destructor on the object – i.e. how to delete an object in place.

This is me, couple of weeks ago.

I think this is going to become my new icon. What do you think?

How less processes its input

Here’s an interesting bit I ran into few days ago. I got curious how is that less (or more) can read file contents from standard input and yet it is able to process input that comes from user. Both of them come from standard input, yet these are quiet heterogeneous streams of information. So, how can it be?

Read the rest of this entry »

I got a new job, take 3

I got a new job, again. Dell decided to obtain what have left of Exanet LTD, a company I worked for in the past. They kindly decided to offer a job to majority of ex-Exanet engineers, me included.

So, I decided to leave Fabrix and as of last week I am software engineer and senior consultant at Dell IDC (Israel Developer Center).

What is direct I/O anyway?

Few days ago I’ve written a post explaining how to do a direct I/O in Python. But then I thought that it might be a good idea to explain what direct I/O is. So, here we go.

As surprising as it is, when you write some information to the disk, it doesn’t get there immediately. In Linux especially, the kernel tries to cache write requests in memory as much as it can. This means that in addition to writing the data to the disk, kernel keeps it in memory. Consecutive read request to the same place on the disk will be much faster because there’s no need to read the information from slow disk – it is already in memory. On the other hand, the information goes to the hard-disk only after some period of time (short though) or when the system runs out of memory. In the meantime, Linux reports that data has been written, despite it is not yet on the disk. Read the rest of this entry »

Recursively deleting symbolic link and what it points to

Recently I looked for a solution to this little problem. how do you, programmatically, delete a symbolic link and a file that it points to?

One problem that you should take care of when tackling this problem, is that symbolic link can point to a symbolic link. Then symbolic link should also point to symbolic link. And once again, and again and again…

Read the rest of this entry »

Direct IO in Python

Doing file I/O of any kind in Python is really easy. You can start with plain open() and friends, working with Python’s file objects. by the way, Python’s open() resembles C‘s fopen() so closely that I can’t stop thinking that open() may be based on fopen().

When its not enough, you can always upgrade to open() and close() from os module. Opening man page on open() (the system call – open(2)) reveals all those O_something options that you can pass to But not all of them can be used in Python. For example, if you open a file with O_DIRECT and then try to write to it, you will end up with some strange error message.

Read the rest of this entry »

I got a new job, take 2

I got a new job, again :-) Today I signed a contract with a company named Fabrix.TV. Fabrix is developing a new generation of video content delivery platform.

I am joining as a chief video content consumer senior software engineer.