Because Python has a convenience function for this? Come on son.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Programming |OT| C is better than C++! No, C++ is better than C
- Thread starter Kikarian
- Start date
Because Python has a convenience function for this? Come on son.
close to the edge
Member
I've been messing around wiith the new Java Nashorn Javascript engine and I can't figure out how to call a method on a Java object in Javascript code. I've been looking around but I can't really find anything. Here's what I want to do: I have a global variable that is set by an init function. The init function sets the global variable to an object that is passed from Java code. Then I want to call a method from that global object. This is where my code fails, I get the following error:
https://gist.github.com/GyrosOfWar/e41f7b5adab0b186dddd
Here's a bit of code that illustrates this problem.
edit: Oh man, I fixed it. The mistake was that the Foo class (in this case) has to be public, not just package-visible so it can be called from JS.
Code:
javax.script.ScriptException: TypeError: Foo{a='hello', b=1245} has no such function "getA" in <eval> at line number 1https://gist.github.com/GyrosOfWar/e41f7b5adab0b186dddd
Here's a bit of code that illustrates this problem.
edit: Oh man, I fixed it. The mistake was that the Foo class (in this case) has to be public, not just package-visible so it can be called from JS.
Treechopper
Member
Hey, I'm far from good. As Der Flatulator showed, it's not so hard in C++ either.
EDIT: Not that I'm fine being where I'm at, mind you. It's aggravating so I've taken the initiative to really dig in and get better at a much quicker pace than I have before.
You should try C#:
Code:
using System.IO;
using System;
class Program
{
static void Main()
{
Console.WriteLine("SomeString".Reverse());
}
}
//As reusable Extension method
public static class ExtensionMethods{
public static string Reverse(this string s )
{
char[] reversalArray = s.ToCharArray();
Array.Reverse(reversalArray);
return new string(reversalArray);
}
}You could just dump Reverse(string) in a regular method, but that's what Extension methods are for.
I'm using it for work right now, it has been smooth so far outside of some bugginess with the XCode Beta.
Well... Python slice operators are not really a convenience function for strings.
Treechopper
Member
Real talk, I want to learn C#. For some reason, I find Microsoft's materials unhelpful in actually learning the language though. I'm probably not looking at the right stuff, so if you or anybody can point me in a good direction that would be a tremendous help.
I'm a big fan of C# in a Nutshell. If you're not into step-by-step tutorials and already have experience in other languages, it's a very good choice for both the basics as well as advanced stuff like delegates, LINQ, threading, serialization etc. It's rather expensive though, $50 I think. Google books has a preview version IIRC, but it's from one of the first editions which doesn't cover C# 4 and 5.
Purple Cheeto
Member
It is okay. Its easy enough to pick up as someone who already knows Obj-C because it uses a reduced version of the same API. I don't really like that it defaults floating precision to double but I get it. It still feels incredibly immature though. I wonder if it really will be ready on release.
You don't even have to know that there's a ready-made implementation of reverse in <algorithm> if you do:
Code:
string str = "some string";
auto reversed = string(str.crbegin(), str.crend());close to the edge
Member
So I recently found this awesome tutorial on building a toy browser engine in Rust (systems programming language developed by Mozilla). I've spent the last two evenings re-implementing the thing in Java 8 and I have to say, Java 8 is so much nicer to use than Java 7. There's still a couple of warts (why can I only have one public class/interface/whatever per file?) but it's kinda fun to use now.
As for the engine, it only parses a small subset of HTML and CSS, but it's still pretty cool. I've managed to hook up the Nashorn JS engine to execute scripts that are contained in the HTML file or linked in the HTML file and it now also has a DOM object that at some point will make it possible to manipulate the DOM from Javascript. Still have to figure out how to feed the manipulated DOM from the JS back to the browser engine.
Here's the source.
As for the engine, it only parses a small subset of HTML and CSS, but it's still pretty cool. I've managed to hook up the Nashorn JS engine to execute scripts that are contained in the HTML file or linked in the HTML file and it now also has a DOM object that at some point will make it possible to manipulate the DOM from Javascript. Still have to figure out how to feed the manipulated DOM from the JS back to the browser engine.
Here's the source.
Has anyone tried Shake? I have to admit, I like the idea of learning a Haskell-based build system.
oxrock
Gravity is a myth, the Earth SUCKS!
From my experience, python is one of the easiest languages to understand for beginners and hit the ground running right out of the gate. Most other languages take more time to understand exactly how things work first. So regardless of python's built in ability to read strings backwards, the time investment you posted makes perfect sense for a beginner to me.
That doesn't mean forsake the C languages! It just means it'll likely take more time until you're as comfortable writing in them. It'll come through time and experience and before long you can look down upon all us "noob" scripting language programmers.
oxrock
Gravity is a myth, the Earth SUCKS!
This got me wondering, as someone who primarily codes in python I often looks for ways to speed things up a bit. Does using iron python or jython lead to faster runtimes than IDLE with identical programs? If so, would it be considerable?
Spent this weekend writing a PopWords solver in python. PopWords is a game on the AppStore with a daily challenge.
Rules:
8x8 grid of letters
Make words by chaining letters together (start at a letter and continue to any unused letters left/right/up/down/diagonal).
Words of length 3 or higher. Points per word length are as follows: 1, 2, 4, 7, 11, 17, 25, 35?, 50, 70?, 100?, 140, 200, 270, unknown higher....
When a word is made, letters fall down to fill empty spots, then empty columns are removed.
Removing all letters doubles the user's score.
Current situation:
I've been able to write a solver (see: https://github.com/Lathentar/PopWordSolver), but it is far too slow for the 8x8 grid. So far it's taken well over 5 hours for the daily puzzle.
Approach:
Take the word bank and make a n-tree where n is the number of letters available at each step of the tree. When finding words, find the letter, see if the letter is available in the given spot of the tree. Then use the tree at that letter for the next letter. Search for a '*' in the tree to see if you've made a completed word as well.
I use a depth first traversal after discovering all words that can be found on the current board sorted by length. For every board I encounter, I generate a string and put it into a dictionary with the result of that depth first search. Before traversing the board, I check to see if I've already seen that board and already know what the best solution is (memoization). I also check to see if it's possible to beat the best current score given the number of letters remaining on the board.
What other approaches do you guys think I could take. A greedy algorithm gives a good solution, but certainly not the best solution. Feel free to look at my code, though it's awful. First real program in python.
Rules:
8x8 grid of letters
Make words by chaining letters together (start at a letter and continue to any unused letters left/right/up/down/diagonal).
Words of length 3 or higher. Points per word length are as follows: 1, 2, 4, 7, 11, 17, 25, 35?, 50, 70?, 100?, 140, 200, 270, unknown higher....
When a word is made, letters fall down to fill empty spots, then empty columns are removed.
Removing all letters doubles the user's score.
Current situation:
I've been able to write a solver (see: https://github.com/Lathentar/PopWordSolver), but it is far too slow for the 8x8 grid. So far it's taken well over 5 hours for the daily puzzle.
Approach:
Take the word bank and make a n-tree where n is the number of letters available at each step of the tree. When finding words, find the letter, see if the letter is available in the given spot of the tree. Then use the tree at that letter for the next letter. Search for a '*' in the tree to see if you've made a completed word as well.
I use a depth first traversal after discovering all words that can be found on the current board sorted by length. For every board I encounter, I generate a string and put it into a dictionary with the result of that depth first search. Before traversing the board, I check to see if I've already seen that board and already know what the best solution is (memoization). I also check to see if it's possible to beat the best current score given the number of letters remaining on the board.
What other approaches do you guys think I could take. A greedy algorithm gives a good solution, but certainly not the best solution. Feel free to look at my code, though it's awful. First real program in python.
Elfforkusu
Member
cmake is good. It's what I use for all my C++ projects.
Its documentation is garbage (or at least it was last I checked), but once you surmount that unnecessary and annoying obstacle you're in the clear. I have very few complaints about it.
lorebringer
Member
Der Flatulator
Member
Code:
$ python -m timeit '"".join(reversed("the quick brown fox spent all day lurking neogaf."))'
1000000 loops, best of 3: 1.49 usec per loop
$ python -m timeit '"the quick brown fox spent all day lurking neogaf."[::-1]'
1000000 loops, best of 3: 0.188 usec per loopBam.
Der Flatulator
Member
Also, the slice version is much prettier
How's this for pretty?
Perl
Code:
my $x = "the quick brown fox spent all day lurking neogaf.";
print +($x =~ /./g)[-$_] for (1 .. $#{[$x =~ /./g]} + 1);
# --> .fagoen gnikrul yad lla tneps xof nworb kciuq ehtWhammy.
lorebringer
Member
I'm starting a class on Assembly next week, and I was wondering if anyone had any tips for approaching the language on my own?
This class along with me putting in the time to get a better understanding of java, algorithms, and data structures should make for a fun semester lol.
This class along with me putting in the time to get a better understanding of java, algorithms, and data structures should make for a fun semester lol.
wat
Well if you know what processor you'll be working with you can look up the spec sheet and manual for it to get a jumpstart.
In general, get really good at converting between hex, binary, and decimal. Also remember each instruction does only one single simple thing, nothing more.
That's good advice. Also read up on binary operators (definitely) and caching (if you have a chance).
Purple Cheeto
Member
Pay more attention than I did in my ASM class back in college!
So I spent some time with your code and was able to get about 2x speedup. The trick is to only do deepcopy when necessary.
Instead of
Code:
newBoard = copy.deepcopy(board)
newBoard[pos[0]][pos[1]] = '*'
if (len(word) >= 3 and '*' in wordlist[charAtPos]):
foundWords.append((word, newBoard))
# keep searching the rest of the board with this letter removed
...I did
Code:
saveOne = board[pos[0]][pos[1]]
board[pos[0]][pos[1]] = '*'
if (len(word) >= 3 and '*' in wordlist[charAtPos]):
newBoard = copy.deepcopy(board)
foundWords.append((word, newBoard))
# keep searching the rest of the board with this letter removed
...
board[pos[0]][pos[1]] = saveOne
return foundWordsBTW, there was a minor bug at line "if ( len(wordlist) >= 1 ):", it was "if ( len(wordlist) > 1 ):", which missed the last word in the dictionary.
It was kind of fun to practice reading someone else's code and a little bit python at the same time. Thanks for an interesting post.
PS: Turns out you can get rid of another deepcopy in minimize_board (do shallow copy only) yes, the result printout takes a bit to figure out ( need to read them backwards), but now it's about 1/3 of the original run time.
Building a vagrant box from scratch is a little tricky. It's quite nice for sharing environments among team members, though.
So I have to do a lab in an hour in a language I have 0 experience with, Python. I've brushed up for 15 minutes now and I've basically solved the problem with it already except for one little damn thing, and it's making me feel silly.
I have a this:
Now I have an else: statement right after this, but I just cannot get the damn thing to just input a line break? What am I missing? Evidently my googling sucks because nothings working.
I have a this:
Code:
for x in range(0, 10):
if x < 10:
print("| %d |" % locations[x]),Now I have an else: statement right after this, but I just cannot get the damn thing to just input a line break? What am I missing? Evidently my googling sucks because nothings working.
FallingEdge
Member
\n is for a new line
Was an error in my range()
Ops can be hard sometimes. If you need help, post about it. I do DevOps, and I use Vagrant nearly every day for testing different deployment scenarios. I might be able to help with the usual hangups with getting Vagrant to do what you want. And I'm pretty sure I'm not the only one here either.
That being said, it's all about Docker now. Oh we're back to bash now? That's cool.
This is for a PHP box that I will use to manage a Laravel site. However, I used http://puphpet.com/ to generate a configuration for me.
The only issue I have now is that Xdebug is a huge fuck up. I've followed several guides on configuring it and making it work with PHPStorm, but it's flaky. I managed to make it work the first time, without issue. Then I tried to run it again later, and it wouldn't work. Then I tried following some other guides and nothing happened. I really, really hate that PHP doesn't have its own debugging feature. Makes me miss Visual Studio and ASP.NET. -_-
I had followed this, and it worked for a moment.
http://www.sitepoint.com/install-xdebug-phpstorm-vagrant/
Then I followed this, and it worked for a few hours yesterday.
http://www.mailbeyond.com/phpstorm-vagrant-install-xdebug-php
And today, I started a new box because I somehow accidentally erased the previous and, and it doesn't want to work. -__-
Note, that it actually works on the default vhost, localhost:80 or localhost:888 on the host machine. But the issue is when I try to reach my other vhost (which houses the Laravel site), nothing happens. Xdebug doesn't trigger for it at a consistent level. Even now, it's not reacting to anything I do.
I added this to the hosts file on the host machine, to reach it through a browser, and that works fine. I can see the site at http//app.dev. But to actually get it to debug has been crap.
I hate you, Xdebug or Apache or whatever you are. Damn you, PHP!
Edit: Seems like Windows itself is an issue this case.
I modified the firewall to ignore the VM's adapter and it seems to work now. I imagine tomorrow everything will break again.
https://stackoverflow.com/questions/20914636/open-windows-firewall-for-xdebug
The only issue I have now is that Xdebug is a huge fuck up. I've followed several guides on configuring it and making it work with PHPStorm, but it's flaky. I managed to make it work the first time, without issue. Then I tried to run it again later, and it wouldn't work. Then I tried following some other guides and nothing happened. I really, really hate that PHP doesn't have its own debugging feature. Makes me miss Visual Studio and ASP.NET. -_-
I had followed this, and it worked for a moment.
http://www.sitepoint.com/install-xdebug-phpstorm-vagrant/
Then I followed this, and it worked for a few hours yesterday.
http://www.mailbeyond.com/phpstorm-vagrant-install-xdebug-php
And today, I started a new box because I somehow accidentally erased the previous and, and it doesn't want to work. -__-
Note, that it actually works on the default vhost, localhost:80 or localhost:888 on the host machine. But the issue is when I try to reach my other vhost (which houses the Laravel site), nothing happens. Xdebug doesn't trigger for it at a consistent level. Even now, it's not reacting to anything I do.
I added this to the hosts file on the host machine, to reach it through a browser, and that works fine. I can see the site at http//app.dev. But to actually get it to debug has been crap.
Code:
192.168.56.101 app.devI hate you, Xdebug or Apache or whatever you are. Damn you, PHP!
Edit: Seems like Windows itself is an issue this case.
I modified the firewall to ignore the VM's adapter and it seems to work now. I imagine tomorrow everything will break again.
https://stackoverflow.com/questions/20914636/open-windows-firewall-for-xdebug
Huh, I've never seen puphpet before. Neat idea. It seems to generate pretty clean Vagrant files too.
That being said, what you're doing should be pretty straightforward, so yeah as you already figured out, it doesn't seem like a Vagrant specific issue. Working with a Vagrant box, often the trickiest part is getting the networking bits to properly communicate between the VM, the host, and the outside world. Again though, it sounds like you were able to work it out.
I will do the snarky thing and say that you have reminded me of some of the reasons I'm glad I haven't done any PHP or Windows-specific work in a while. Mostly because I think I've blissfully forgotten a lot of the annoyances that were just par for the course for me before.
Of course, I suppose that now I'm just used to a different set of annoyances. Hey guys, how about that CoreOS?
That being said, what you're doing should be pretty straightforward, so yeah as you already figured out, it doesn't seem like a Vagrant specific issue. Working with a Vagrant box, often the trickiest part is getting the networking bits to properly communicate between the VM, the host, and the outside world. Again though, it sounds like you were able to work it out.
I will do the snarky thing and say that you have reminded me of some of the reasons I'm glad I haven't done any PHP or Windows-specific work in a while. Mostly because I think I've blissfully forgotten a lot of the annoyances that were just par for the course for me before.
Of course, I suppose that now I'm just used to a different set of annoyances. Hey guys, how about that CoreOS?
Any Android devs on here?
What's the best way to implement location auto complete?(in a AutoCompleteTextView). I'm using Google Places API right now, and it works well but seems very limiting at 1000 requests per day. Foursquare's API seems nice at 5000 per hour, but it seems limited to an area (so my user can't type any city in the world). I'm wondering if I should just use a local database at this point?
What's the best way to implement location auto complete?(in a AutoCompleteTextView). I'm using Google Places API right now, and it works well but seems very limiting at 1000 requests per day. Foursquare's API seems nice at 5000 per hour, but it seems limited to an area (so my user can't type any city in the world). I'm wondering if I should just use a local database at this point?
I had to use Google Places API for a 1-day hackathon project a few weeks ago. I believe you can increase your quota on the Google Places API for free up to 10,000. You have to set up an actual account with them with a credit card so they can keep track of you though.
I'm waiting on a (voluntary) flight delay so I'll try and find the relevant info for you to kill some time.
Edit - Even better, it bumps you to 100,000 per day:
Der Flatulator
Member
Local db seems a little unmaintainable. How about running Google Places API through a caching server to limit requests? That way you could possibly get more bang for your buck per API request, whilst sending minimal data between the server and the device.
Thanks for the great response. I'm not surprised there were some issues. Good job getting rid of a number of the deep copies, I'm sure there are ways to reduce it even more. However, to get the best solution for the problem the runtime is still too long. I'm pretty certain this is a problem that is exponential.
It was fun for my first crack at python. I'll probably rewrite it again to write more readable code, basically replacing all of the tuples with a class. It was getting difficult to remember where everything was supposed to go.
I'll probably them rewrite this again in D. I've been curious in that language for a while.
BTW. Nice catch on the len(wordlist) > 0 issue. It should probably be len(wordlist[charAtPos]) >= 1. Also, I think moving the deepcopy into the if statement causes another bug as you'll be writing the '*' into the board object, but not clearing it properly. The better change would be to only do a deep copy when appending to the foundWords and undoing the assignment to '*' for that character before the return statement. Thus you only make a deep copy of the board for completed words. This should dramatically reduce copies.
Thanks!
I actually wasn't planning on maintaining it since places don't really change
And a caching server would probably be a little overkill for this little project
There are some improvements can still be made, e.g., better estimation of the value of remaining board, and faster stopping criteria for word searching, but the overall structure is pretty much set since you want the best solution. Switching to C++ and using some rudimentary data types (strings and arrays) should speed things up quite a bit (10x?). Another thing is to try make it parallel. That could be fun.
CrudeDiatribe
Member
What's the script do with that parameter?
If you make a barebones foo.php and just have it echo $_GET['n'], do you still get a 403 for 'eyelasers'?