We Keep Coding

PUN 2 Getting Custom Properties

I've recently taken on the task of custom properties in Photon. I have been able to figure out how to set the custom properties, but not get the custom properties. My hashtable is in my player controller script, while the place where I set (and where I want to get) properties is in a round loop script.
From RoundSystem:
private IEnumerator TeamBalance()
{
angelCount = Mathf.Floor(PhotonNetwork.PlayerList.Length * angelPercent);
currentAngels = angelCount;
currentPlayers = PhotonNetwork.PlayerList.Length;
foreach (var item in PhotonNetwork.PlayerList)
{
var itemPhotonView = (PhotonView)item.TagObject;
itemPhotonView.RPC("SetPlayerTeam", item, citiString);
}
for (int i = 0; i < angelCount;)
{
var item = PhotonNetwork.PlayerList[Random.Range(0, PhotonNetwork.PlayerList.Length)];
var itemPhotonView = (PhotonView)item.TagObject;
if (/* random player selected's, AKA, item's team == citiString */)
{
itemPhotonView.RPC("SetPlayerTeam", item, angelString);
i++;
}
}
yield return null;
//the reason this is in an IEnumerator with 'yield return null'
//is because I plan to add a waiting period once I figure this out
//it's for the game loop
}
From PlayerController:
[PunRPC]
public void SetPlayerTeam(string teamString)
{
//in the class: private ExitGames.Client.Photon.Hashtable playerProperties;
if (!playerProperties.ContainsKey("team"))
{
playerProperties.Add("team", teamString);
}
playerProperties["team"] = teamString;
PhotonNetwork.LocalPlayer.SetCustomProperties(playerProperties);
}
At the beginning of the round, a percentage (in this case 1/3) of players are chosen to be an "angel". The check here is needed because in cases of multiple angels, you don't want an already existing angel to count as a new change. (Also, it's probably important to known generally how to get custom properties if I'm going to be using them.) If I don't include the check in RoundSystem, the outcome is 2 citizens and 1 angel (in a test with 3 players). Also, if you see any spaghetti code that could be improved on, please don't hesitate to tell me. :)

Use Player.CustomProperties dictionary to access player's custom properties.
foreach (var item in PhotonNetwork.PlayerList)
{
if (item.CustomProperties.ContainsKey("team"))
{
Debug.Log(item.CustomProperties["team"]);
}
}
Also, the RoundSystem can implement IInRoomCallbacks interface and listen to OnPlayerPropertiesUpdate to catch the exact moment when the team gets updated. https://doc-api.photonengine.com/en/pun/v2/interface_photon_1_1_realtime_1_1_i_in_room_callbacks.html

Question about PhotonNetwork.CurrentRoom.CustomProperties

When the game starts in multiplayer, the master client sends a PunRPC to have all clients run a function. This function tries to get the room properties to see if the game is active, if so it does something. For some reason a client gets a null reference error, but the master client does not. The strange thing is, a debug of the hash table for room properties is visible but I cannot get a specific item in it.
Tried Debugging the hash table to make sure that the key was set when the code was run. It was. "(System.String)ag=(System.Boolean)True ag=activeGame" This shows in Debug.Log(hash); But (bool)hash[rpk.activeGame] gets a null reference error. But only on the client side not the master client. So the key also works.
// Call all the clients to set up the room settings in the sub menu.
[PunRPC]
private void GameRoomSetup (string pOne, string pTwo, int pOneColor, int pTwoColor)
{
GameObject gameMenu = GameObject.Find ("GameMenu");
gameMenu.GetComponent<SubMenu> ().UpdatePlayers (pOne, pTwo, pOneColor, pTwoColor);
gameMenu.GetComponent<SubMenu> ().StartGameSetup ();
// If you are a player, change the active buttons that are visible.
if (PhotonNetwork.NickName == pOne || PhotonNetwork.NickName == pTwo)
{
gameMenu.GetComponent<GameButtonManager> ().GameStart ();
}
hash = PhotonNetwork.CurrentRoom.CustomProperties;
Debug.Log(hash);
if ((bool)hash[rpk.activeGame]) // Error on this line on client but not on master client. Says null reference.
{
GameObject.Find ("SoundManager").GetComponent<SoundManagerScript> ().PlayBackgroundTwo ();
GameObject.Find ("GameMenu").GetComponent<SubMenu> ().ChangeSubMenuActive (false);
}
}
I'm trying to run my if statement as a client but I get an error.

Thank you for choosing Photon!
To get a custom property, I recommend you use TryGetValue method as follows:
hash = PhotonNetwork.CurrentRoom.CustomProperties;
object temp;
string key = rpk.activeGame;
if (hash.TryGetValue(key, out temp))
{
if (temp is bool)
{
bool activeGame = (bool)temp;
}
else
{
// unexpected custom property value type
}
}
else
{
// custom property not found
}
If the custom property is not available yet, wait for the callback IInRoomCallbacks.OnRoomPropertiesUpdate(Hashtable propertiesThatChanged) (reference API).
Other notes and recommendations:
private void GameRoomSetup (string pOne, string pTwo, int pOneColor, int pTwoColor)
Not sure if it's supported or if it's a good idea to pass multiple parameters to a PUN RPC method.
To debug log a Dictionary or a Hashtable you could make use of SupportClass.DictionaryToString() method.
so instead of
Debug.Log(hash);
use
Debug.Log(SupportClass.DictionaryToString(hash));
Avoid calling expensive methods like GameObject.Find:
GameObject gameMenu = GameObject.Find ("GameMenu");
Also here you have duplicate calls to gameMenu.GetComponent<SubMenu>(), at least call it once and cache the component result found if any.
gameMenu.GetComponent<SubMenu> ().UpdatePlayers (pOne, pTwo, pOneColor, pTwoColor);
gameMenu.GetComponent<SubMenu> ().StartGameSetup ();
Comparing strings should not be done using == operator. At least use Equal method and proper StringComparison type. Read "How to compare strings in C#".
// If you are a player, change the active buttons that are visible.
if (PhotonNetwork.NickName == pOne || PhotonNetwork.NickName == pTwo)
{
gameMenu.GetComponent<GameButtonManager> ().GameStart ();
}
Besides, why do you use the Nickname to check if it's player one or two? maybe use ActorNumber or a custom player index. Or use the player count if the room is for 2 players only.

CRCase_RowSelecting fires with next case on subsequent save

I created an extension for CRCaseMaint, and added the event CRCase_RowSelecting. Here is the code I am currently using:
protected virtual void CRCase_RowSelecting(PXCache sender, PXRowSelectingEventArgs e)
{
CRCase row = e.Row as CRCase;
if (row == null) return;
PXDatabase.ResetSlot<List<CRCase>>("OriginalCase");
List<CRCase> originalCaseSlot = PXDatabase.GetSlot<List<CRCase>>("OriginalCase");
if (originalCaseSlot.Count == 0)
{
originalCaseSlot.Add(sender.CreateCopy(row) as CRCase);
}
else
{
originalCaseSlot[0] = sender.CreateCopy(row) as CRCase;
}
}
When I first open a case, this event will fire a couple times, and the last time it fires, the current case is correctly stored in e.Row, so this code works great. When I click Save, I have a RowPersisting event that compares the case stored in the originalCaseSlot with the updated case. At the end, it sets the original case slot to the updated case. This also works well.
However, when I make another change without leaving the case, and click save, e.Row on the RowSelecting event now has the next case stored on it rather than the current case. Since I am not touching the next case in any way, I am surprised that this is happening.
My question is, should I be using a different event instead of RowSelecting, or is there something else I am missing?
Thank you all for your help.

Sometimes when the primary record gets updated or the user clicks on a form toolbar button, the framework selects 2 records from database: the current primary record and the next one. This is why RowSelecting is invoked 2nd time for the next CRCase record.
Honestly, using PXDatabase Slots to store user session-specific records is not a good idea. PXDatabase Slots are shared among all user sessions and should only be used to cache frequently used data from database, which is not prone to frequent updates. This makes the main purpose of PXDatabase Slots to reduce number of database queries to widely and very often used configurable data, like Segment Key or Attribute configurations.
With that said, using the RowSelecting handler is definitely a step in the right direction. Besides, the RowSelecting handler, you should additionally define a separate PrevVersionCase data view to store the original CRCase record(s) and also override the Persist method to report about changes. The Locate method used on PXCache objects searches the cache for a data record that has the same key fields as the provided data record. This approach allows to compare changes between the originally cached and modified CRCase records having identical key field values.
public class CRCaseMaintExt : PXGraphExtension<CRCaseMaint>
{
[Serializable]
public class CRPrevVersionCase : CRCase
{ }
public PXSelect<CRPrevVersionCase> PrevVersionCase;
protected virtual void CRCase_RowSelecting(PXCache sender, PXRowSelectingEventArgs e)
{
CRCase row = e.Row as CRCase;
if (row == null || e.IsReadOnly) return;
var versionCase = new CRPrevVersionCase();
var versionCache = PrevVersionCase.Cache;
sender.RestoreCopy(versionCase, row);
if (versionCache.Locate(versionCase) == null)
{
versionCache.SetStatus(versionCase, PXEntryStatus.Held);
}
}
[PXOverride]
public void Persist(Action del)
{
var origCase = Base.Case.Current;
var origCache = Base.Case.Cache;
CRPrevVersionCase versionCase;
if (origCache.GetStatus(origCase) == PXEntryStatus.Updated)
{
versionCase = new CRPrevVersionCase();
origCache.RestoreCopy(versionCase, origCase);
versionCase = PrevVersionCase.Cache.Locate(versionCase) as CRPrevVersionCase;
if (versionCase != null)
{
foreach (var field in Base.Case.Cache.Fields)
{
if (!Base.Case.Cache.FieldValueEqual(origCase, versionCase, field))
{
PXTrace.WriteInformation(string.Format(
"Field {0} was updated", field));
}
}
}
}
del();
if (origCase != null)
{
PrevVersionCase.Cache.Clear();
versionCase = new CRPrevVersionCase();
Base.Case.Cache.RestoreCopy(versionCase, origCase);
PrevVersionCase.Cache.SetStatus(versionCase, PXEntryStatus.Held);
}
}
}
public static class PXCacheExtMethods
{
public static bool FieldValueEqual(this PXCache cache,
object a, object b, string fieldName)
{
return Equals(cache.GetValue(a, fieldName), cache.GetValue(b, fieldName));
}
}

Determine whether there are changes on a form when you click a submit button

I have a form with two textboxes. I am retrieving data from the
database to populate the boxes. When my user clicks on submit button
and the content of the 2 textboxes does not change, I dont want to go through
the code.
How do I determine when the content of the boxes changes and when it does not change?
Do I need to make some kind of comparison to what I have in memory?
public ActionResult Edit(profile objprofiler)
{
if (ModelState.IsValid)
{
//Go fetch the existing profile from the database
var currentProfile = db.Profiles.FirstOrDefault(p => p.ProfileId == objprofiler.ProfileId);
//Update the database record with the values from your model
currentProfile.City = objprofiler.City;
currentProfile.State = objprofiler.State;
//Commit to the database!
db.SaveChanges();
ViewBag.success = "Your changes have been saved";
return View(profiler);
}
}

You can compare the values with a simple if condition. Something like this:
if ((currentProfile.City != objprofiler.City) || (currentProfile.State != objprofiler.State))
{
currentProfile.City = objprofiler.City;
currentProfile.State = objprofiler.State;
db.SaveChanges();
}
Or use whatever logic you're trying to achieve, really. Whether you want to compare for each field individually, use a && instead of an ||, etc. The logic you want to implement is up to you. But you'd perform the comparison in an if statement.
Note also that you can use string.Equals() instead of just the == operator to compare strings with some more options, such as case sensitivity options and other useful things.
If the comparison gets more complex, you might also encapsulate it in the profile object itself. Perhaps by overriding .Equals(), though that has other implications when testing for equality. Maybe just a simple helper function:
public bool IsEqualTo(profile obj)
{
return this.City == obj.City
&& this.State == obj.State;
}
Then in the controller you can just use that method:
if (!currentProfile.IsEqualTo(objprofiler))
db.SaveChanges();

The way I typically handle this is by setting a 'dirty' flag any time a data change event occurs on any of the form's controls.
When the user comes to submit the form, I just check the state of the flag to see whether any changes need to be saved. This avoids having to compare all data to their previous states, which can be a nuisance if there are a lot of input controls on the form.
For example:
bool isDirty;
private void textBox_TextChanged(object sender, EventArgs e)
{
// Possible validation here
SetDirty(true);
}
private void SetDirty(bool dirty)
{
// Possible global validation here
isDirty = dirty;
}
private void Submit()
{
if(isDirty)
{
// Save logic
}
}
This approach allows you to run any global validation logic whenever any data is changed.
Caveat: If a user makes a change then reverts it, the form will still submit the data.

On the client side you can check if the value has changed by running some js to compare the elements value to its initial value. Something like this.
function hasFormChanged() {
//textboxes, textareas
var els = document.querySelectorAll('input[type="text"], textarea, input[type="number"]');
for (i = 0; i < els.length; i++) {
var el = els[i];
if (el.value !== el.defaultValue) {
return true;
}
}
//checkboxes and radios
els = document.querySelectorAll('input[type="radio"], input[type="checkbox"]');
for (i = 0; i < els.length; i++) {
var el = els[i];
if (el.checked !== el.defaultChecked) {
return true;
}
}
//select
els = document.querySelectorAll('select');
for (i = 0; i < els.length; i++) {
var el = els[i];
if (el.options[el.selectedIndex].value != '') {
if (!el.options[el.selectedIndex].defaultSelected) {
return true;
}
}
}
//if we get here then nothing must have changed
return false;
}
and it that function return true indicating that something has changed you can set a hidden form value like this
<input type="hidden" value="false" id="AnyUpdates" name="AnyUpdates"/>
to true.
Then in your controller update read that field to determine if you need to do your db stuff.

Linked list Advice

I've been given a task as follows:
I'd like to to build an implementation of a Linked List. Specifically I'd like it to be a doubly linked list.
My task:
Your program should use the linked list to model a train route using the linked list.
First the user will enter as many stops as they'd like the train to have, and the name of each stop.
The program should then print a map of the route.
Once finished they then enter the name of the stop they want to start at.
From there they can enter commands to move the train either forward to the next stop or backward to the previous one.
I've been told I'm not doing this task right but I don't really understand how not, I'd appreciate it if someone could explain what I'm not doing that I should be doing.
My Route class (it isn't finished but it would've been nearly finished if it was done correctly):
namespace TrainRoute
{
class Route
{
Stops root;
public LinkedList<Stops> linkedList = new LinkedList<Stops>();
public Stops MakeNewStop(string stopName)
{
Stops stopWithStopName = new Stops(stopName);
return stopWithStopName;
}
public void AddStops(Stops stopIWantToAdd)
{
if (linkedList.Count == 0)
{
linkedList.AddFirst(stopIWantToAdd);
}
else
{
//stopIWantToAdd.prevStop = linkedList.Last();
linkedList.AddLast(stopIWantToAdd);
}
}
public void StopRelationships()
{
for (int i = 0; i < linkedList.Count; i++)
{
if (linkedList.ElementAt<Stops>(i).nextStop == null && linkedList.ElementAt<Stops>((i + 1)) != null)
{
linkedList.ElementAt<Stops>(i).nextStop = linkedList.ElementAt<Stops>((i + 1));
}
if (linkedList.ElementAt<Stops>((i - 1)) != null)
{
linkedList.ElementAt<Stops>(i).prevStop = linkedList.ElementAt<Stops>(i - 1);
}
}
}
public void Print()
{
if (linkedList != null)
{
foreach (var item in linkedList)
{
Console.WriteLine("Stop name: " + item.stopName);
}
}
}
public int StopPosition(string usersInput)
{
int position = 0;
for (int i = 0; i < linkedList.Count; i++)
{
if (linkedList.ElementAt<Stops>(i).stopName == usersInput)
{
position = i;
break;
}
}
return position;
}
public int MoveForward(int indexPosition)
{
Console.WriteLine("The train is now at " +linkedList.ElementAt<Stops>(indexPosition).nextStop.stopName);
return (indexPosition + 1);
}
public int MoveBackwords(int indexPosition)
{
Console.WriteLine("The train is now at " + linkedList.ElementAt<Stops>(indexPosition).prevStop.stopName);
return (indexPosition - 1);
}
public bool VerifyRoute(int indexPosition, string prevOrForward)
{
if (prevOrForward.Contains("forward"))
{
if (linkedList.ElementAt<Stops>((indexPosition+1)) != null)
{
return true;
}
}
else
{
if (linkedList.ElementAt<Stops>((indexPosition-1)) != null)
{
return true;
}
}
return false;
}
}
}
I'm also not allowed to use the Linked list class but I'm to use a linked list (I'm not 100% sure what that means).
Any and all advice/help provided will be appreciated!

Let's piece together the breadcrumbs here:
I'd like to to build an implementation of a Linked List.
and this:
I'm also not allowed to use the Linked list class
Obviously the task here is for you to implement your own linked list (class), and not to use the existing one provided by .NET.
I'm assuming here the task is not to build the program handling the trains, but instead to learn how a linked list works, and how you would go about implementing one.
As such, your shortcut to simply grab the existing class is the wrong tool for the job. It would be perfect (probably) if your task was to build that program, but in this case the program is orthogonal to your task, it's there to create a context for what you're really asked to do:
Implement your own version of LinkedList<T> (though you probably don't need to make it generic).
Wikipedia has a very good article on linked lists if you're stumped on how such a data structure really works. There's undoubtedly other very good resources out on the net as well, and probably in your text book or other resources.
Additionally, I would urge you to find a classmate to peer with, from experience I can say that most of the really hard problems I've had in my programming career has (usually) been solved by having a sparring partner to work with.

Implementing a linked list isn't that difficult. I assume you have a textbook and it discusses linked list, read it, carefully. Also you want to clarify with your tutor exactly how much functionality your linked list needs to implement.
Basically, you'll start with a node class, if you don't need it to be generic, then you can create a StopNode class. The basics of your node class will be a reference to the next node in the list and, since this is a doubly linked list, a reference to the previous node:
public class StopNode
{
public StopNode Next { get; set; }
public StopNode Previous { get; set; }
// whatever other properties your stop class needs - such as name
}
Now your LinkedList class will manage the collection of stop nodes. It will need to keep a reference to the first or "head" node and probably the last node too (or "tail").
public class StopLinkedList
{
private StopNode Head { get; }
private StopNode Tail { get; }
}
And it will need to implement methods to add and remove nodes (at a minimum) and probably also insert.
Add is pretty easy - check if Head is null. If it is, just set Head and Tail both equal to your new node. If it's not, you will instead set the Next property of your Tail to your new node, then set the Previous of your new node to the Tail and then finally update your Tail to reference your new node.
To remove a node, if given the node to remove, you will need to check it's Previous and Next properties and (assuming one or both isn't null - you'll need to add logic for that), you set your nodes Previous.Next to your nodes Next and your nodes Next.Previous to your nodes Previous. This will cause your node to fall out of the list (you can set your nodes Next and Previous to null if required, but it's not strictly necessary unless your removed node is going to hang around).
Hopefully that gets you started. Clarify with your tutor, check your textbook (probably better to try and match their terminology if it differs from mine) and also search for "linked list" and "doubly linked list" on the internet. You should find plenty of resources.