|
Thread Rules
1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution.
2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20)
3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible.
4. Use [code] tags to format code blocks. |
On December 28 2014 04:11 CatNzHat wrote:Show nested quote +On December 26 2014 19:04 Manit0u wrote:
JS really needs some coherent way of writing code. Like you can use the semicolon but you can just as well not use it at all. And then you have to deal with code written by various people that mixes the styles and it's a mess even bigger than in the other languages that adhere to a much more strict way of doing things. And because in JS everything is an object you can chain everything and you end up working with lots of code violating the Single Responsibility Principle etc.
Seriously, JS scripts suffer from excessive code bloat more often than anything I've seen and debugging it is painful as hell. JSLINT is your friend. If you use the module design pattern you can more strictly adhere to SRP. The language is not at fault here, it's the programmers.
I've taken up JS very recently (although I had intention to for a long time) and I found it pretty easy. Of course, there is a big lack of well-established best practices in comparison to C/C++/Java.
In terms of debugging though. GOD BLESS JAVA AND JVM.
|
|
|
On December 30 2014 10:41 Animzor wrote:
Quick noob Java question:
I'm supposed to write a method that prints every phone number that belongs to two or more people. I have a person class, phone number class and each person has an arraylist with phone numbers. I have tried everything that I can think but I just can't figure this one out. I've gotten very close though by using nested for loops where I check if one person has the same number as another person, but it never works out. I feel like I'm overlooking some easy solution, if anyone can point me in the right direction I'd be grateful.
It might be easier to make the phone number have an arraylist of people and maybe a counter for how many people have that number. At least for that specific problem.
|
Inelegant solution would be to make two new lists. Iterate through each person and add numbers to the first list. If that number is already on the first list then add it to the second list (you'll have to check that it isn't already on the second list as well so you don't print duplicates). Then you can print the second list which will have all the numbers.
|
|
|
On December 30 2014 10:41 Animzor wrote:
Quick noob Java question:
I'm supposed to write a method that prints every phone number that belongs to two or more people. I have a person class, phone number class and each person has an arraylist with phone numbers. I have tried everything that I can think but I just can't figure this one out. I've gotten very close though by using nested for loops where I check if one person has the same number as another person, but it never works out. I feel like I'm overlooking some easy solution, if anyone can point me in the right direction I'd be grateful.
For each person, for each phone number you could add the phone number to a hashset, it returns a boolean according to the existence in the hashset. If the add to the hashset isn't new, you can print the phone number.
If you want to use a hashset on a custom class you'll need to implement hashcode, just make sure to keep it consistent with equals.
On December 30 2014 11:10 Animzor wrote:I forgot to add. I also need to list the people that the number belongs to like this: number1, name, name, name number2, name, name number3, name, name, name, name and I have to have an arraylist of numbers in the person class for other functions. This is just a small part of the program, I've finished everything else, but I can't figure this one out 
If you also need the people , you could use HashMap<PhoneNumber, List<Person>>. If you loop once through all the people you can populate the map with the info, then loop over the hashmap looking for where the list.size() > 2. You'll still need to look at hashcode and equals if you want to use your custom phone number class as the key.
|
Make pairs of people and numbers, then group by numbers, then print all groups that have more than 1 element.
In C# using Linq:
void printNumbers(IEnumerable<Person> persons)
{
// for each person, makes as many pairs as the person has number
// then puts the pairs for all persons in a single list
var personNumberPairs = persons.SelectMany(p => p.Numbers.Select(n => new { Person = p, Number = n}));
var pairsGroupedByNumber = personNumberPairs.GroupBy(p => p.Number);
var pairsWithMoreThanOne = pairsGroupedByNumber.Where(p => p.Count() > 1);
foreach(var groupOfPairs in pairsWithMoreThanOne)
{
// print contents of groupOfPairs
// where number = groupOfPairs.Key
// and names = groupOfPairs.Select(g => g.Person.Name)
}
}
No idea how that translates into Java, does it have a proper Linq equivalent? In any case you need a remapping of the persons->numbers mapping so there can not be a trivial solution. In terms of steps required I think the above is as simple as it gets, with the HashMap solution above being pretty much the same thing in a more low level form.
|
You need a 3-step process here:
1. Create a hash map. Populate map keys with all the numbers.
2. Iterate over persons and their numbers, add person as map value associated with number as key.
3. Iterate over your map populated with keys and values now to create a new map which only meets the criteria (ie: value has at least 2 names in it).
|
On December 30 2014 17:52 spinesheath wrote:Make pairs of people and numbers, then group by numbers, then print all groups that have more than 1 element. In C# using Linq: void printNumbers(IEnumerable<Person> persons)
{
// for each person, makes as many pairs as the person has number
// then puts the pairs for all persons in a single list
var personNumberPairs = persons.SelectMany(p => p.Numbers.Select(n => new { Person = p, Number = n}));
var pairsGroupedByNumber = personNumberPairs.GroupBy(p => p.Number);
var pairsWithMoreThanOne = pairsGroupedByNumber.Where(p => p.Count() > 1);
foreach(var groupOfPairs in pairsWithMoreThanOne)
{
// print contents of groupOfPairs
// where number = groupOfPairs.Key
// and names = groupOfPairs.Select(g => g.Person.Name)
}
} No idea how that translates into Java, does it have a proper Linq equivalent? In any case you need a remapping of the persons->numbers mapping so there can not be a trivial solution. In terms of steps required I think the above is as simple as it gets, with the HashMap solution above being pretty much the same thing in a more low level form.
nice one
instead of that last foreach tho
var result = pairsWithMoreThanOne.Select(t=> new { Number = t.Key, NamesBelonging = string.Join(",", t.Select(g => g.Person.Name)) });
i hate foreaches ,they're old and ugly
|
You still have to print the result though:
var rows = pairsWithMoreThanOne.Select(p => p.Key + ", " + string.Join(", ", p.Select(g => g.Person.Name)));
var result = string.Join(Environment.NewLine, rows);
Console.WriteLine(result);
I usually try to pull as much code out of a foreach as I can reasonably do. Especially control flow statements. I still prefer foreach over anything that involves the keyword let in a Linq block though.
I'd also wrap the longer lambdas into Methods like such:
var rows = pairsWithMoreThanOne.Select(CreateRow);
string CreateRow(IGrouping<Number, Person> p)
{
return p.Key + ", " + string.Join(", ", p.Select(g => g.Person.Name));
}
Which separates the details of row layout from the other code, which is quite nice.
|
yeah separation is cleaner more readable, and more unit testable (if this was something more complex, worthy of unit tests)
|
@amnesty
i dl'ed visual15 preview today, and tried to get seq and abstract to compile.
am having some problems with constexpr when you expand a parameter pack through a constexpr function into an initializer list and return it, along the lines of:
template<class... T>
struct
F
{
static constexpr std::array<return_type_of_constexpr_func,sizeof(T...)>
apply()
{ return { constexpr_function<T>()... }; }
};
in addition it seems like the direct product (or coordinate function) is instantiating the 'wrong' specilalizations (different compared to gcc).
i'll have to take another stab at it later.
but it prompted me to rearrange the metastructure of the seq library in a saner fashion:
namespace seq{namespace guts{
template<class>
struct
metafun_impl;
template<class... arguments>
struct
metafun_impl<use arguments...>
{
...
};
}
template<class... arguments>
using metafun=guts::metafun_impl<...>;
}
trying to employ a naming convention along the lines of metafunction = indirect object, and metafunction parameter = direct object was rough, my vernacular was too feeble and the convention lead me into the borderlands of absurd identifiers.
see how he (the programmer) uses a spanner to tighten that nut
template<class>
struct
spanner;
template<class... descriptors>
struct
spanner<nut<descriptors...>>
{
...
};
template<class nut>
using tighten=typename spanner<nut>::output;
...
using too_tight_of_a_nut=tighten<already_tight_nut>;
disclaimer: i'm not very grammar, but the example shows what i mean.
|
Happy New Year guys! Less software bugs, more profit in the New Year! 
|
On December 30 2014 19:10 Manit0u wrote:
You need a 3-step process here:
1. Create a hash map. Populate map keys with all the numbers.
2. Iterate over persons and their numbers, add person as map value associated with number as key.
3. Iterate over your map populated with keys and values now to create a new map which only meets the criteria (ie: value has at least 2 names in it).
This worked amazingly well btw. Thank you!
|
finally got clang working on my computer, need to recompile gcc 4.10 as well.
variable templates, guuu.
![[image loading]](http://2.bp.blogspot.com/-VTfewMOceeg/VKUroPuvMMI/AAAAAAAACRQ/Cp6mFkyTPew/s1600/cpp14conf.png)
src
instant disappointment, it doesn't seem very sfinae-friendly...
hope this is a bug and not the intended functionality!
#include<type_traits>
template<class argument>
constexpr bool condition=argument::output;
template<class argument,class=void>
struct
outer;
template<class argument>
struct
outer<argument,std::enable_if_t<condition<argument>>>
{ using output=int; };
template<class argument>
struct
outer<argument,std::enable_if_t<!condition<argument>>>
{ using output=float; };
template<class argument>
using inner=typename outer<argument>::output;
struct
argument
{ static constexpr bool output{false}; };
int
main()
{ inner<argument> value; }
clang++ -c -std=c++1y -I. -I../ -Wpedantic -Wfatal-errors -fdiagnostics-color -fmessage-length=64 -O2 work.cpp
work.cpp:55:1: fatal error: implicit instantiation of undefined
template 'outer<argument, void>'
using inner=typename outer<argument>::output;
^
work.cpp:64:2: note: in instantiation of template type alias
'inner' requested here
inner<argument> value;
^
work.cpp:42:2: note: template is declared here
outer;
^
surely outer<argument,std::enable_if_t<!argument::output>> ought to be instantiated, and not the primary template (if i define it that's the one that gets instantiated).
|
Haha, nunez you're so set on C++11/14. It's kind of pointless though because most (?) companies haven't caught up yet.
|
o ye of little faith! that is irrelevant before god.
be not anxious for your life, what ye may eat, and what ye may drink, nor for your body, what ye may put on.
seek ye first the reign of The Most Conforming Compiler and Its Righteousness, and all these shall be added to you.
i'm compiling fresh gcc now, it think it has support for variable templates.
if it agrees with me, and disagrees with the verdict of the Clang, i will cast the Clang back into the abyss.
|
I just spent several hours tinkering with GC options on the JVM to get my shit to work. One piece of advice for those who like me are pretty sure they don't have any memory leak in their code (mine was in Scala and it's pretty easy to code so that you don't leak in that language ;D) and whose program ends up running full GCs for nothing like 100% of the computing time: start thinking about tuning the GC. I mean, think about it last, but don't forget to think about it either, because that would be a mistake 
Oh and happy new year guys!
|
On January 01 2015 23:09 nunez wrote:
instant disappointment, it doesn't seem very sfinae-friendly...
hope this is a bug and not the intended functionality!
<snip>
This works on whatever version of gcc Coliru uses, if you get rid of the bool template variable: http://coliru.stacked-crooked.com/a/dd147352e035c760
You could just do this, though: http://coliru.stacked-crooked.com/a/e17282b0b18b948d
I find the direction C++ is going in a little odd. vector<bool> doesn't even work properly, but sure lets have constexpr and template variables, and enable_if all over the place.
|
sorry i misread.
you had removed the variable template: the raison d'etre of my prayer.
coliru uses 4.9.2 (__VERSION__ macro), so i didn't bother checking it earlier (since i thought it was the same i had on my puter).
i will not forsake clang yet.
my wish was exactly not to do any of those.
actually my real wish is to use variable templates instead of constexpr functions to call value metafunctions in my little seq library, so that value metafunction calls and type metafunction calls have uniform syntax.
i agree with your critique, it's odd how there are things, and also other things at the same time.
edit:
same result in gcc-5.0, however there's already a bugreport: src. i believe the author has his shit together, so i will conclude it is indeed a bug, and not the intended semantic, and will have to wait a bit longer... actually it's not quite the same. clang accepts his code, but not mine.
|
|
|
|
|
|
EPT
