switch to room list switch to menu My folders
Go to page: [1] 2 3 4 5 ... Last
↑↑↑ Old messages ↑↑↑            ↓↓↓ New messages ↓↓↓
[#] Mon Mar 26 2012 11:09:50 EDT from wizard of aahz @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Build monkeys and python...

[#] Mon Mar 26 2012 12:20:35 EDT from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Exception handling is pretty cool. Even in my little program I've
already written pieces that are self-healing by simply catching any
exception and returning an error to the function's caller.

This was tongue in cheek, right? ;)

[#] Mon Mar 26 2012 14:13:02 EDT from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I'm still waiting on Ford's response.

[#] Mon Mar 26 2012 14:42:40 EDT from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Yes it's 2012 and I'm exploring the world of exception handling for the first time. Remember the part where I said I'm not a professional developer? :)

"Something went wrong over there; you should check it out" is a big change from the world I normally play in, where the response is "We're going to crash now; you'd better hope your users are smart enough to send in decent bug reports so you can fix it in the next release"

[#] Tue Mar 27 2012 07:08:15 EDT from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Exceptions yank control from the normal flow of a program, which on the one hand prevents everything from getting completely out of control, but on the other hand, make it hard to determine where the fault occured sometimes. It's very, very annoying to mysteriously have an exception get thrown somewhere that kills your program, yet not have a freaking clue who threw it, and perhaps a vague notion of why.

I suppose you can have the same problem without exceptions, but as a general rule, I prefer to pay attention to all values retrieved from the functions I call, and have code in place to react to them appropriately when something is out of place (e.g. the returned file pointer is not valid, so don't try to use it, react appropriately).

[#] Tue Mar 27 2012 14:37:41 EDT from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

In C++-land it's a little different than on the Java Ranch and in the Python Pit: in C++, every exception doesn't print out the stack trace of the faulting line... unless I'm quite mistaken.

[#] Tue Mar 27 2012 15:46:23 EDT from Spell Binder @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

TCL also dumps a stack trace for an uncaught exception. Though I can't generalize, since I've not done much programming in other scripting languages, it seems like exception handling is somewhat easier in an interpreted language like TCL, Perl, or Python, than it is in a compiled language like C or C++. I know that TCL, for example, gives the programmer the ability to create their own control structures, which can make exception handling very easy.
TCL Binder

[#] Tue Mar 27 2012 16:24:48 EDT from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Well, that's true. In C++, the only stack you'll get is if you're in a debugger... otherwise, you'll need the luck of the gods to help you track down an exception.

[#] Wed Mar 28 2012 09:28:15 EDT from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Python seems to *want* the programmer to use exceptions. For example, my program uses urllib2 to fetch data via HTTPS from a remote server. If that operation fails, it throws an exception. There is no return value to check.
If you want the program to not crash you have to catch the exception. That's just how it's done.

What I found convenient was that I could put the exception handler anywhere in the stack. I was able to tell the program, "if any of that stuff up there had a problem, take this action instead of that action." Since it's a garbage collected language, I also don't have to worry about freeing this buffer or that buffer depending on which failure mode occurred.

[#] Wed Mar 28 2012 12:32:53 EDT from Spell Binder @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Since I'm a computer engineer and not a computer scientist, I could be off on this idea, but it seems to me that exception handling is not too far off from event-oriented programming. Your program might encounter an exception or receive an event at pretty much any time. Either your code is set up to handle it, or it gets passed up to the next level.

One other thought I had about exceptions, especially for languages that don't automatically capture a stack trace, is to use correlators. I've seen these in some library APIs from time-to-time, but didn't really understand what they were until somewhat recently.

As I understand it, a correlator is nothing more than a unique identifier that gets passed into a stack chain from a higher-level caller. For example:

int corr = 13;
int rc = myFunc(arg1, arg2, corr);
int myFunc(int a, int b, int corr) {
int c;

c = a + b;
return someOtherFunc(c, a, b, corr);
int someOtherFunc(int a, int b, int c, int corr) {
if (c < 0) {
printf("In function someOtherFunc, correlator=%d: %d is less than zero.\n", corr, c);
return -1;
return a + b + c;

(Pardon any coding errors, I'm a little rusty with C)

The idea is to have a unique correlator for every high-level "operation" that occurs. That way, if a lower-level function encounters an exceptional condition, the correlator can be used to trace back to the original operation.

Of course, I think it's somewhat clumsy to have to pass a correlator value into every function call. I wonder if that's one reason I haven't seen them used more often.
Correlated Binder

[#] Wed Mar 28 2012 13:10:10 EDT from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

For my part, I program to avoid using exceptions most of the time, and prefer to pass error values instead. I do most of my programming in C++.

When I have to use something that forces me to catch exceptions, I catch them, and react accordingly.

I also try to do the whole RIIA thing (or whatever that acronym is) where you use constructors to create resources and destructors to remove them in a clean fashion, so an exception will automatically clean things up for you.

Still, Python will allow you to walk an exception stack, as I recall, so at least you can work out why your program stopped, and take action accordingly.

[#] Wed Mar 28 2012 17:21:41 EDT from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

There's no doubt about it, Python is a nice language. If I had to choose between it, Perl and Ruby, Python usually comes out on top. Clean, crisp, readable syntax. The indenting thing is not the way I would have designed it, but I can live with it.

If your job requires you to occasionally wade through academic papers describing complex algorithms, you'll find that an increasing number of authors are expressing their pseudocode in a suspiciously Pythonesque syntax.

[#] Thu Mar 29 2012 10:19:46 EDT from ax25 @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I am a fan of Python for quick what if proggies and have made some use of the ctypes module for extra added fun poking around in C libs:

You don't have to re-write C libraries, just use them from Python.

[#] Fri Mar 30 2012 17:13:12 EDT from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I just finished writing a screen scraper that reformats the received data, reformats it so that it is spoofing another system's output, and feeds it into another screen scraper. I feel so dirty. (But hey, it's my first Python program, so ... #winning?)

[#] Sat Mar 31 2012 04:03:46 EDT from shadowjester @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Personally, I love using Python. I can code the same program/script/data structure/spam/eggs about half the time than it would take me to dev in another language. Actually, I just finished doing some proof-of-concept code for semantic web development with Python. It was very procedural, but it demonstrated the OWL ontology in a representational form quite nicely.
I do tend to use Python where I used to use BASIC back in the day. Haven't made the leap to using it as my shell, though. Maybe my kids will?

[#] Sun Apr 01 2012 13:35:34 EDT from dothebart @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]


Tue Mar 27 2012 16:24:48 EDT from fleeb @ Uncensored

Well, that's true. In C++, the only stack you'll get is if you're in a debugger... otherwise, you'll need the luck of the gods to help you track down an exception.

at least in linux, its not like that.;a=blob;f=citadel/citserver.c;h=088b584e1107a751cc5b935f4e08c8c858ea5f58;hb=HEAD#l90

does a simple backtrace...

[#] Sun Apr 01 2012 16:44:01 EDT from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Yeah, I'm not aware of anything like that in Windows.

[#] Sat Apr 14 2012 11:31:23 EDT from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

git should be called GNU/git because even though Linus Torvalds created it, Richard Stallman is capable of louder and more prominent whining. (Hey, what goes for operating systems goes for source code management tools too, right?)

[#] Fri Apr 20 2012 02:59:13 EDT from kc5tja @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

You might as well call it Nougat then. :)

[#] Fri Apr 20 2012 17:22:04 EDT from kc5tja @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Anyone here familiar with acceptance testing using Capybara and Chromium browsers, particularly in an EC2 Ubuntu instance?

Go to page: [1] 2 3 4 5 ... Last