Related
I've got a method that I'm calling ... well, very often. And now the GC is becoming a problem, because apparently this method is creating garbage, but I can't seem to figure out why.
Item item;
foreach (Point3D p in Point3D.AllDirections)
{
item = grid[gridPosition + p];
if (item != null && item.State != State.Initialized)
else
return;
}
OtherMethod(this);
This method is inside the Item class and figures out the state of neighbouring items. The gridPosition variable is obviously the position of this item inside the grid and the Point3D is a struct that just contains 3 ints, one for each axis. It also contains a few predefined static arrays such as AllDirections which itself only contains Point3Ds.
Now, the state of items can change at any time and once all neighbours have a specific state (or rather aren't in the Initialized state) I do some other stuff (OtherMethod), which also changes the state of the item so this method is no longer called. As such the produced garbage is only a problem when neighbours don't change their state (for example when there is no user input). I guess you could think of this as a substitution for an event driven system.
I admit that I have a rather limited understanding of memory management and I tend to only really understand things by doing them, so reading up on it didn't really help me much with this problem. However, after reading about it I got the impression that every value type, that is defined within a method and doesn't leave it, will be collected upon leaving the method. So I have a little trouble figureing out what's actually left after the return.
Any ideas on what's actually generating the garbage here ? Of course the best case would be to produce no garbage at all, if at all possible.
(Just in case this is relevant, I'm working in Unity and the method in question is the Monobehaviour Update)
Edit for all the comments:
Holy cow you guys are quick. So...
# Jeppe Stig Nielsen: There is nothing between if and else, there was once but I didn't need that anymore. It's just a quick way to leave if no neighbour was found. Also, Indexer ! Completely forgot about that, here it is:
public Item this[Point3D gridPosition]
{
get
{
Item item;
items.TryGetValue(gridPosition, out item);
return item;
}
}
Could the TryGetValue be the reason ?
# Luaan: I already tried a normal for loop, didn't change anything. That looked something like this:
for (int i = 0; i < 6; i++)
{
item = grid[gridPosition + Point3D.AllDirections[i]];
}
If it was the enumerator, that should've fixed it, no ?
# Rufus: Yes that's a better way to write the if, will fix that. The whole thing is a work in progress right now, so that's just a remnant of past experimentation. There is currently is nothing else in the if statement.
# Tom Jonckheere: Well there isn't really much else. It's just Unitys Update() followed by two if statements for the State. And that's really just because I currently only work with two states but have already setup more, so someday there will probably be a switch instead of ifs.
# Adam Houldsworth: Well yes, the problem is certainly that I call it so often. The odd thing is, the amount of garbage produced varies wildly. As far as I can tell it's in the range of 28 bytes to 1.8KB. As for the whole red herring thing, I don't think that's the case. The profiler points to the Update method, which only contains two ifs, one which is the code from above, the other one isn't being used when no state changes occur. And when testing, there are no state changes, except for intial setup.
Also, sorry for the wrong tag, first time I've visited this site. Looks like I'll have some reading to do :)
foreach calls GetEnumerator, which, in turn, creates a new instance of Enumerator every time it's called.
So after Jeppe Stig Nielsen and Luaan made me remember that I'm infact using an indexer I did some googleing and found out that this is not a new issue. Apparently Dictonary.TryGetValue works by calling Equals(object o) on the keys. Now these have to be unboxed which, as far as I understand it, creates the garbage. I now followed the advice I found on Google and implemented the IEquatable interface in my Point3D struct, which does not requiere any boxing. And as it turns out, a big chunk of the garbage is gone.
So the generated garbage per call went from ~28B-1.8KB down to ~28B-168B. Which is pretty good, no more GC spikes. Though I'm still curious where the rest of the garbage is coming from.
Anyway, this was already a big help, so thanks for your answers guys.
Edit: Figured out where the rest of the garbage was comeing from. I just called base.GetHashCode, because I didn't know any better. I did some research into pretty much everything related to hash codes and then implemented it correctly. And now I'm down to exactly 0 bytes of garbage. Pretty neat.
EDIT: I changed the Title to a more appropriate name, base on Comments from David
I am experienced with c# programming and have never really run into this before, though I am surprised I haven't
I am using Long name variables, because I plan to come back to my program in a few months from now, but I am annoyed at how long it is taking me to write code. I considered using shorter names then just replacing them all later. But I was wondering if there is syntax sugar for something like this
Using (var b as TheButtonThatMakesThingsHappen)
{
b.Name="TheButtonThatMakesThingsHappen";
b.Location=...
b.etc...
}
Instead of
TheButtongThatMakesThingsHappen.Name="TheButtonThatMakesThingsHappen";
TheButtongThatMakesThingsHappen.Location=...
TheButtongThatMakesThingsHappen.etc...
I did a google search and tried here on stackoverflow, but I don't know what this is called,
Any Ideas?
As David posted in the comments section on your question, I think that the method that you described for renaming is a needless shortcut to get around the issue of long variable names. Try to name the variables to something more constructive, like SubmitButton or UserSubmitButton instead of TheButtongThatMakesThingsHappen.There is something very wrong with your variable naming scheme... if you don't actually use it.
Another alternative to long variable names is comments with shorter, more descriptive variable names- for example:
button SubmitButton; //this button is used in the second form field of the website
I don't feel like there is an answer here yet. While I am starting to think the concept itself is flawed, at least for the reasons I want to do it, I think it is unfair not to have a proper solution for someone looking later. I have taking the suggestions From David and others and compiled this as my answer.
#region TheButtonThatMakesThingsHappen
{
var b = TheButtonThatMakesThingsHappen;
b.Name = "TheButtonThatMakesThingsHappen";
b.Location=...;
b.etc...;
}
#endregion
It allows a local scope alias and the region tags allow for collapsing code, which keeps things neater.
I recently had a coding bug where under certain conditions a variable wasn't being initialized and I was getting a NullReferenceException . This took a while to debug as I had to find the bits of data that would generate this to recreate it the error as the exception doesn't give the variable name.
Obviously I could check every variable before use and throw an informative exception but is there a better (read less coding) way of doing this? Another thought I had was shipping with the pdb files so that the error information would contain the code line that caused the error. How do other people avoid / handle this problem?
Thanks
Firstly: don't do too much in a single statement. If you have huge numbers of dereferencing operations in one line, it's going to be much harder to find the culprit. The Law of Demeter helps with this too - if you've got something like order.SalesClerk.Manager.Address.Street.Length then you've got a lot of options to wade through when you get an exception. (I'm not dogmatic about the Law of Demeter, but everything in moderation...)
Secondly: prefer casting over using as, unless it's valid for the object to be a different type, which normally involves a null check immediately afterwards. So here:
// What if foo is actually a Control, but we expect it to be String?
string text = foo as string;
// Several lines later
int length = text.Length; // Bang!
Here we'd get a NullReferenceException and eventually trace it back to text being null - but then you wouldn't know whether that's because foo was null, or because it was an unexpected type. If it should really, really be a string, then cast instead:
string text = (string) foo;
Now you'll be able to tell the difference between the two scenarios.
Thirdly: as others have said, validate your data - typically arguments to public and potentially internal APIs. I do this in enough places in Noda Time that I've got a utility class to help me declutter the check. So for example (from Period):
internal LocalInstant AddTo(LocalInstant localInstant,
CalendarSystem calendar, int scalar)
{
Preconditions.CheckNotNull(calendar, "calendar");
...
}
You should document what can and can't be null, too.
In a lot of cases it's near impossible to plan and account for every type of exception that might happen at any given point in the execution flow of your application. Defensive coding is effective only to a certain point. The trick is to have a solid diagnostics stack incorporated into your application that can give you meaningful information about unhandled errors and crashes. Having a good top-level (last ditch) handler at the app-domain level will help a lot with that.
Yes, shipping the PDBs (even with a release build) is a good way to obtain a complete stack trace that can pinpoint the exact location and causes of errors. But whatever diagnostics approach you pick, it needs to be baked into the design of the application to begin with (ideally). Retrofitting an existing app can be tedious and time/money-intensive.
Sorry to say that I will always make a check to verify that any object I am using in a particular method is not null.
It's as simple as
if( this.SubObject == null )
{
throw new Exception("Could not perform METHOD - SubObject is null.");
}
else
{
...
}
Otherwise I can't think of any way to be thorough. Wouldn't make much sense to me not to make these checks anyway; I feel it's just good practice.
First of all you should always validate your inputs. If null is not allowed, throw an ArgumentNullException.
Now, I know how that can be painful, so you could look into assembly rewriting tools that do it for you. The idea is that you'd have a kind of attribute that would mark those arguments that can't be null:
public void Method([NotNull] string name) { ...
And the rewriter would fill in the blanks...
Or a simple extension method could make it easier
name.CheckNotNull();
If you are just looking for a more compact way to code against having null references, don't overlook the null-coalescing operator ?? MSDN
Obviously, it depends what you are doing but it can be used to avoid extra if statements.
I have a simple question for you (i hope) :)
I have pretty much always used void as a "return" type when doing CRUD operations on data.
Eg. Consider this code:
public void Insert(IAuctionItem item) {
if (item == null) {
AuctionLogger.LogException(new ArgumentNullException("item is null"));
}
_dataStore.DataContext.AuctionItems.InsertOnSubmit((AuctionItem)item);
_dataStore.DataContext.SubmitChanges();
}
and then considen this code:
public bool Insert(IAuctionItem item) {
if (item == null) {
AuctionLogger.LogException(new ArgumentNullException("item is null"));
}
_dataStore.DataContext.AuctionItems.InsertOnSubmit((AuctionItem)item);
_dataStore.DataContext.SubmitChanges();
return true;
}
It actually just comes down to whether you should notify that something was inserted (and went well) or not ?
I typically go with the first option there.
Given your code, if something goes wrong with the insert there will be an Exception thrown.
Since you have no try/catch block around the Data Access code, the calling code will have to handle that Exception...thus it will know both if and why it failed. If you just returned true/false, the calling code will have no idea why there was a failure (it may or may not care).
I think it would make more sense if in the case where "item == null" that you returned "false". That would indicate that it was a case that you expect to happen not infrequently, and that therefore you don't want it to raise an exception but the calling code could handle the "false" return value.
As it standards, you'll return "true" or there'll be an exception - that doesn't really help you much.
Don't fight the framework you happen to be in. If you are writing C code, where return values are the most common mechanism for communicating errors (for lack of a better built in construct), then use that.
.NET base class libraries use Exceptions to communicate errors and their absence means everything is okay. Because almost all code uses the BCL, much of it will be written to expect exceptions, except when it gets to a library written as if C# was C with no support for Exceptions, each invocation will need to be wrapped in a if(!myObject.DoSomething){ System.Writeline("Damn");} block.
For the next developer to use your code (which could be you after a few years when you've forgotten how you originally did it), it will be a pain to start writing all the calling code to take advantage of having error conditions passed as return values, as changes to values in an output parameter, as custom events, as callbacks, as messages to queue or any of the other imaginable ways to communicate failure or lack thereof.
I think it depends. Imaging that your user want to add a new post onto a forum. And the adding fail by some reason, then if you don't tell the user, they will never know that something wrong. The best way is to throw another exception with a nice message for them
And if it does not relate to the user, and you already logged it out to database log, you shouldn't care about return or not any more
I think it is a good idea to notify the user if the operation went well or not. Regardless how much you test your code and try to think out of the box, it is most likely that during its existence the software will encounter a problem you did not cater for, thus making it behave incorrectly. The use of notifications, to my opinion, allow the user to take action, a sort of Plan B if you like when the program fails. This action can either be a simple work around or else, inform people from the IT department so that they can fix it.
I'd rather click that extra "OK" button than learn that something went wrong when it is too late.
You should stick with void, if you need more data - use variables for it, as either you'll need specific data (And it can be more than one number/string) and an excpetion mechanism is a good solution for handling errors.
so.. if you want to know how many rows affected, if a sp returned something ect... - a return type will limit you..
In C# I use the #warning and #error directives,
#warning This is dirty code...
#error Fix this before everything explodes!
This way, the compiler will let me know that I still have work to do. What technique do you use to mark code so you won't forget about it?
Mark them with // TODO, // HACK or other comment tokens that will show up in the task pane in Visual Studio.
See Using the Task List.
Todo comment as well.
We've also added a special keyword NOCHECKIN, we've added a commit-hook to our source control system (very easy to do with at least cvs or svn) where it scans all files and refuses to check in the file if it finds the text NOCHECKIN anywhere.
This is very useful if you just want to test something out and be certain that it doesn't accidentaly gets checked in (passed the watchful eyes during the diff of everything thats commited to source control).
I use a combination of //TODO: //HACK: and throw new NotImplementedException(); on my methods to denote work that was not done. Also, I add bookmarks in Visual Studio on lines that are incomplete.
//TODO: Person's name - please fix this.
This is in Java, you can then look at tasks in Eclipse which will locate all references to this tag, and can group them by person so that you can assign a TODO to someone else, or only look at your own.
If I've got to drop everything in the middle of a change, then
#error finish this
If it's something I should do later, it goes into my bug tracker (which is used for all tasks).
'To do' comments are great in theory, but not so good in practice, at least in my experience. If you are going to be pulled away for long enough to need them, then they tend to get forgotten.
I favor Jon T's general strategy, but I usually do it by just plain breaking the code temporarily - I often insert a deliberately undefined method reference and let the compiler remind me about what I need to get back to:
PutTheUpdateCodeHere();
An approach that I've really liked is "Hack Bombing", as demonstrated by Oren Eini here.
try
{
//do stuff
return true;
}
catch // no idea how to prevent an exception here at the moment, this make it work for now...
{
if (DateTime.Today > new DateTime(2007, 2, 7))
throw new InvalidOperationException("fix me already!! no catching exceptions like this!");
return false;
}
Add a test in a disabled state. They show up in all the build reports.
If that doesn't work, I file a bug.
In particular, I haven't seen TODO comments ever decrease in quantity in any meaningful way. If I didn't have time to do it when I wrote the comment, I don't know why I'd have time later.
//TODO: Finish this
If you use VS you can setup your own Task Tags under Tools>Options>Environment>Task List
gvim highlights both "// XXX" and "// TODO" in yellow, which amazed me the first time I marked some code that way to remind myself to come back to it.
I'm a C++ programmer, but I imagine my technique could be easily implemented in C# or any other language for that matter:
I have a ToDo(msg) macro that expands into constructing a static object at local scope whose constructor outputs a log message. That way, the first time I execute unfinished code, I get a reminder in my log output that tells me that I can defer the task no longer.
It looks like this:
class ToDo_helper
{
public:
ToDo_helper(const std::string& msg, const char* file, int line)
{
std::string header(79, '*');
Log(LOG_WARNING) << header << '\n'
<< " TO DO:\n"
<< " Task: " << msg << '\n'
<< " File: " << file << '\n'
<< " Line: " << line << '\n'
<< header;
}
};
#define TODO_HELPER_2(X, file, line) \
static Error::ToDo_helper tdh##line(X, file, line)
#define TODO_HELPER_1(X, file, line) TODO_HELPER_2(X, file, line)
#define ToDo(X) TODO_HELPER_1(X, __FILE__, __LINE__)
... and you use it like this:
void some_unfinished_business() {
ToDo("Take care of unfinished business");
}
It's not a perfect world, and we don't always have infinite time to refactor or ponder the code.
I sometimes put //REVIEW in the code if it's something I want to come back to later. i.e. code is working, but perhaps not convinced it's the best way.
// REVIEW - RP - Is this the best way to achieve x? Could we use algorithm y?
Same goes for //REFACTOR
// REFACTOR - should pull this method up and remove near-dupe code in XYZ.cs
I use // TODO: or // HACK: as a reminder that something is unfinished with a note explaining why.
I often (read 'rarely') go back and finish those things due to time constraints.
However, when I'm looking over the code I have a record of what was left uncompleted and more importantly WHY.
One more comment I use often at the end of the day or week:
// START HERE CHRIS
^^^^^^^^^^^^^^^^^^^^
Tells me where I left off so I can minimize my bootstrap time on Monday morning.
// TODO: <explanation>
if it's something that I haven't gotten around to implementing, and don't want to forget.
// FIXME: <explanation>
if it's something that I don't think works right, and want to come back later or have other eyes look at it.
Never thought of the #error/#warning options. Those could come in handy too.
I use //FIXME: xxx for broken code, and //CHGME: xxx for code that needs attention but works (perhaps only in a limited context).
Todo Comment.
These are the three different ways I have found helpful to flag something that needs to be addressed.
Place a comment flag next to the code that needs to be looked at. Most compilers can recognize common flags and display them in an organized fashion. Usually your IDE has a watch window specifically designed for these flags. The most common comment flag is: //TODO This how you would use it:
//TODO: Fix this before it is released. This causes an access violation because it is using memory that isn't created yet.
One way to flag something that needs to be addressed before release would be to create a useless variable. Most compilers will warn you if you have a variable that isn't used. Here is how you could use this technique:
int This_Is_An_Access_Violation = 0;
IDE Bookmarks. Most products will come with a way to place a bookmark in your code for future reference. This is a good idea, except that it can only be seen by you. When you share your code most IDE's won't share your bookmarks. You can check the help file system of your IDE to see how to use it's bookmarking features.
I also use TODO: comments. I understand the criticism that they rarely actually get fixed, and that they'd be better off reported as bugs. However, I think that misses a couple points:
I use them most during heavy development, when I'm constantly refactoring and redesigning things. So I'm looking at them all the time. In situations like that, most of them actually do get addressed. Plus it's easy to do a search for TODO: to make sure I didn't miss anything.
It can be very helpful for people reading your code, to know the spots that you think were poorly written or hacked together. If I'm reading unfamiliar code, I tend to look for organizational patterns, naming conventions, consistent logic, etc.. If that consistency had to be violated one or two times for expediency, I'd rather see a note to that effect. That way I don't waste time trying to find logic where there is none.
If it's some long term technical debt, you can comment like:
// TODO: This code loan causes an annual interest rate of 7.5% developer/hour. Upfront fee as stated by the current implementation. This contract is subject of prior authorization from the DCB (Developer's Code Bank), and tariff may change without warning.
... err. I guess a TODO will do it, as long as you don't simply ignore them.
This is my list of temporary comment tags I use:
//+TODO Usual meaning.
//+H Where I was working last time.
//+T Temporary/test code.
//+B Bug.
//+P Performance issue.
To indicate different priorities, e.g.: //+B vs //+B+++
Advantages:
Easy to search-in/remove-from the code (look for //+).
Easy to filter on a priority basis, e.g.: search for //+B to find all bugs, search for //+B+++ to only get high priority ones.
Can be used with C++, C#, Java, ...
Why the //+ notation? Because the + symbol looks like a little t, for temporary.
Note: this is not a Standard recommendation, just a personal one.
As most programmers seem to do here, I use TODO comments. Additionally, I use Eclipse's task interface Mylyn. When a task is active, Mylyn remembers all resources I have opened. This way I can track
where in a file I have to do something (and what),
in which files I have to do it, and
to what task they are related.
Besides keying off the "TODO:" comment, many IDE's also key off the "TASK:" comment. Some IDE's even let you configure your own special identifier.
It is probably not a good idea to sprinkle your code base with uninformative TODOs, especially if you have multiple contributors over time. This can be quite confusing to the newcomers. However, what seems to me to work well in practice is to state the author and when the TODO was written, with a header (50 characters max) and a longer body.
Whatever you pack into the TODO comments, I'd recommend to be systematic in how you track them. For example, there is a service that examines the TODO comments in your repository based on git blame (http://www.tickgit.com).
I developed my own command-line tool to enforce the consistent style of the TODO comments using ideas from the answers here (https://github.com/mristin/opinionated-csharp-todos). It was fairly easy to integrate it into the continuous integration so that the task list is re-generated on every push to the master.
It also makes sense to have the task list separate from your IDE for situations when you discuss the TODOs in a meeting with other people, when you want to share it by email etc.