|
Tony Wesley wrote: I use goto's
It's OK, so do I, in C#! 
xacc.ideIronScheme a R5RS-compliant Scheme on the DLR
The rule of three: "The first time you notice something that might repeat, don't generalize it. The second time the situation occurs, develop in a similar fashion -- possibly even copy/paste -- but don't generalize yet. On the third time, look to generalize the approach."
|
|
|
|
|
So what happens to your "goto" cleanup routine when an exception is thrown?
I'll give you a hint -- it doesn't get called.
|
|
|
|
|
CurtD wrote: So what happens to your "goto" cleanup routine when an exception is thrown?
I'll give you a hint -- it doesn't get called.
The goto doesn't figure into that. Just above it in the thread is another snippet of my code that doesn't use a goto.
Same issue. What happens to the code after the while loop when an exception is thrown? It doesn't get executed.
Once could argue that the problem is with exception. As I pointed out elsewhere, Joel Spolsky argues that exceptions [...] create an abrupt jump from one point of code to another. In fact they are significantly worse than goto's: (emphasis added)
I don't quite agree with Joel. But there was a time, when I was debugging an app that another programmer had written, that I understood that completely. It was spaghetti exception handling. I never knew where I had come from.
I'm not saying you shouldn't use exceptions. But they can be abused as badly as gotos.
|
|
|
|
|
Tony Wesley wrote: The goto doesn't figure into that. Just above it in the thread is another snippet of my code that doesn't use a goto.
Same issue. What happens to the code after the while loop when an exception is thrown? It doesn't get executed.
Once could argue that the problem is with exception. As I pointed out elsewhere, Joel Spolsky argues that exceptions [...] create an abrupt jump from one point of code to another. In fact they are significantly worse than goto's: (emphasis added)
I don't quite agree with Joel. But there was a time, when I was debugging an app that another programmer had written, that I understood that completely. It was spaghetti exception handling. I never knew where I had come from.
I'm not saying you shouldn't use exceptions. But they can be abused as badly as gotos.
I'm talking about the case where coders like to put a block of cleanup code at the bottom of a method and "goto" it when something goes wrong or fall into it in a non-error situation. If you are depending on blocks of cleanup code, an exception will bypass it completely. This worked back in C, but not in OO design.
That's why I use objects that clean themselves up when destroyed -- automatically. I am far too lazy to try to figure out every possible error situation, catch exceptions from anything that might throw them, and ensure that there is an appropriate goto.
I disagree completely with Joel Spolsky. He suggests using return codes instead of exceptions. Return codes and exceptions are completely different things. Exceptions are not meant to be return codes. They are "exceptional" situations where usually the app cannot continue. And exceptions have the major advantage of unwinding the stack, which he fails to mention. They are not "an abrupt jump from one point of code to another."
It seems like OO programming has really failed to catch on with a lot of developers.
modified on Sunday, December 16, 2007 12:10:46 AM
|
|
|
|
|
But then why not
do
{
...
} while ( false ) ;
|
|
|
|
|
|
heh.
of course that's just a coding-standards-compliant way of writing:
if (set (ifrFragment))
{
result = true;
ifrTerm = ifrFragment;
goto done;
}
if (errorCode)
goto done;
if (keyword (TokenSubtype::Not) && term (ifrFragment))
{
result = true;
ifrTerm = TokenSubtype::Not.asString() + " " + ifrFragment;
goto done;
}
if (errorCode)
goto done;
done:
|
|
|
|
|
|
You're sure you're not mistaking that "one" in lnIndice <= 1 for an "ell"?
Marc
|
|
|
|
|
Only a real programmer who codes in raw hex, or someone very daft, or someone with the editor font set to one with massive differences between 1 and l would ever call a variable "l".
i, on the other hand...
Ninja (the Nerd)
Confused? You will be...
|
|
|
|
|
That's what I call extendable coding. 
|
|
|
|
|
FINALLY!!!!
My first entry on a coding horror!
Is this a sign I am starting to learn, possibly, but even in goood old fashioned ms basic circa 1981 I would have spotted this load of old codswallop!
So pleased to say that I would never have done this!
------------------------------------
Happy Primes Lead to Happy Memories.
Don't Google FGI
|
|
|
|
|
heheheh . Cool thing....i myself as developer can understand that sort of things can happen when code passes through certain hands without proper comments.
|
|
|
|
|
Whether it was the basement of great code or the ruin of an unlucky project. Who knows... 
A good optimized cempiler would clean this up. 
Greetings from Germany
|
|
|
|
|
omg.. what else is there to say ??
Michael Davey
biproject.com rss and blog news in a more palatable format
mobile edition now available!
|
|
|
|
|
- Pascal - wrote: We found it very funny
I find it quite useful, actually. This is the sort of thing that sits in code that is prone to change. Why destroy the entire structure when a month or two from now, someone has a need to add in an index of 2 and 3?
Though, admittedly, I'd comment out everything but the 'do something' code when it got this low. But I'd still leave it THERE.
|
|
|
|
|
I was cleaning up the code in a parser. There is a switch/case statement that spans over 2,200 lines. I removed a bunch of exits from various case statements and moved it up as a global test on the errorCode.
It just didn't want to exit when the errorCode was set.
state=1;
do
{
result = token( tokstr, FALSE, NULL, FALSE);
switch(state)
{
if (errorCode != 0)
return FALSE;
case 1:
switch(result)
{
}
} while( result != Keyword::EOFTOK );
“Although distant in time, Homobonus does in fact figure as a saint for the Church and society of our time.”
-- Pope John Paul II
|
|
|
|
|
2200 lines of cases? Impressive, at least to me, humble employee of small companies.
|
|
|
|
|
Oshtri Deka wrote: 2200 lines of cases? Impressive, at least to me, humble employee of small companies
I'm not happy with 2200 lines of cases. I'm seeking to clean this up and reduce complexity. Things tend to just keeping growing.
I'm curious, do you see the problem with my posted code?
|
|
|
|
|
I was sarcastic, though I am impressed that someone wrote so many lines of code for one switch-case block. What does that application do?
checking before first case?
|
|
|
|
|
Oshtri Deka wrote: I was sarcastic, though I am impressed that someone wrote so many lines of code for one switch-case block. What does that application do?
That module has to do with parsing a parameter file. It probably was state-of-the-art in 1983. It has just grown and grown. We need an option at to handle FOOBAR? No problem, add a #DEFINE FOOBAR 1742 and one more case statement.
Continue the process for almost a quarter century.
The application classifies inpatient hospital records into disease categories and stages of the disease. The portion of the program that does that has been modernized.
checking before first case?
Yeah, that code was unreachable. I checked it into source control. It worked great if there were no errors. 
|
|
|
|
|
You should construct a data structure and keep the possible values in case statements in an ordered way in your data structure; then your incoming value should be compared with the values inside your data structure in binary search fashion...
this will be faster and save you from time consuming sequential searching
|
|
|
|
|
xenonysf wrote: You should construct a data structure and keep the possible values in case statements in an ordered way in your data structure; then your incoming value should be compared with the values inside your data structure in binary search fashion...
this will be faster and save you from time consuming sequential searchin
I agree with what you say but not for the reasons you give.
A switch/case statement has better performance than a sequence of if/else statements. The compiler may generate a jump table. For large tables, some use binary searches. See http://www.eventhelix.com/RealtimeMantra/Basics/CToAssemblyTranslation3.htm[^].
So performance isn't the driving factor here.
But I do agree that the cases -- where possible -- should be encapsulated in a data structure. We would like to have actions for the various cases in a data structure to simply the logic.
Complexity leads to increased probability of error. We have unneccessary complexity. However, it takes a while to untangle code that has grown for decades.
|
|
|
|
|
switch(state)
{
if (errorCode != 0)
return FALSE;
case 1:
switch(result)
{
My guess is that execution jumps to one of case labels and skip the if statement entirely?
[ My Blog] "Visual studio desperately needs some performance improvements. It is sometimes almost as slow as eclipse." - Rüdiger Klaehn "Real men use mspaint for writing code and notepad for designing graphics." - Anna-Jayne Metcalfe
|
|
|
|
|
My compiler woud throw exception
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
