thanks.. not sure how that could have happened as that's a standard construct file.

i'lll know what to look for now.

strange DatapawWolf, you use the same source that I do.. to suggest 20-30% usage for IE. I don't know why you feel that 20 of that 28% is be non-game playing.

However each to our own choices. for me it's normal to support and test on all 3 main browsers for a desktop application. (it'd be 4 if I had a mac)

to be honest W3Schools captures a certain segment of the market.. ie the coders.. this does not mean that those dynamics match your players. whatever historical issues IE has had, they are gone. I prefer IE11 because of do-not-track and for privacy reasons. that's my call.

However whatever you believe, I still do not see why if you are going for a desktop application that you should deliberately ignore IE11 users.

now I have seen this before on Scirra.. it's very much a google-orientation here and understandably so, but I'm not sure it's good for all cases.

PS: in my local university all the browsers are modern IE's.. mostly for the business and arts students

R

if the data is being saved as a textfile to the server, why don't you have your clients access the textfiles directly via ajax.. just make sure that they are saved in an accessible location

ggibson1, interesting that it suggests 58%... so let me return the below link which states 25% usage

http://gs.statcounter.com/#desktop-brow ... 307-201407

Don't you love these different analysts. Either way.. I think we're making the same point... being incompatible with IE will reduce your available market on the desktop which can never be a good thing.

I'm working on an MMO-RTS, and I'm just using C2 for the client. server is all custom. C2 is well suited for client development. not sure if this is more than a proof of concept, but it's a learning experience so far.

I know you can use a C2 developed host as a server as described, however I feel that's not scalable. I also strongly feel that it's easier to code complex logic in a proper language than in C2 - at least for me. This is doubly so when you start to use sql databases. Use the right tool for the right job.

I'd also flag that as webrtc is not supported by IE, it would seem to me to be bad practice to dependent on that. I know it's only 25-30% of the desktop market.. but still. if you're doing an android game.. then clearly this is not an issue.

as said, a quick search will find you lots of options

there's webrtc for realtime comms, if you don't want to support Internet explorer

there's scirra's hosted multiplayer plugin and server

there's the ajax plugin which makes RESTful API's easy for turn based or asynchronous games

then there's websocket as well.

so a lot depends on what you want to do )

I caught this error when I ran the following command.

add key "timer" with value max(0,DATA.Get("timer")-1)

The problem arose as I had previously defined the contents of DATA.get("time") as an int, and I had changed it to text. My mistake. Either way, there's a casting issue and this throws up the following error when executed in the browser. So the error could have arisen because of the subtraction, or the MAX function. This might be useful to other people who are trying to decipher the error. If you see an assertion error, check if you are misusing a text variable. It might be worth to Ashley or the developers to put in an error check to help catch the problem or to help a developer diagnose the failure if they've made the same mistake.

Assertion failure: Calling expvalue.set_float without number type

Stack trace:

Error

at assert2 (http://localhost:50000/preview_prelude.js:16:4)

at ExpValue.prototype.set_float (http://localhost:50000/expressions.js:788:3)

at SysExps.prototype.max (http://localhost:50000/system.js:2054:3)

at ExpNode.prototype.eval_system_exp (http://localhost:50000/expressions.js:255:3)

at Parameter.prototype.get_exp (http://localhost:50000/eveng.js:2049:3)

at Action.prototype.run_object (http://localhost:50000/eveng.js:1824:7)

at EventBlock.prototype.run_actions_and_subevents (http://localhost:50000/eveng.js:929:4)

at EventBlock.prototype.run (http://localhost:50000/eveng.js:886:4)

at EventSheet.prototype.run (http://localhost:50000/eveng.js:224:4)

at EventInclude.prototype.run (http://localhost:50000/eveng.js:2318:13)

I love Angryfly's work.. but the links there will not work.. but you can find his stuff at

https://sketchfab.com/angryfly3d

Problem Description

Round2 <img src="{SMILIES_PATH}/icon_e_smile.gif" alt=":)" title="Smile"> System-create does not seem to be working on the exported and hosted version. Taking your advice Ashley, here is a stripped down version that replicates the problem. It is not a "clean build" as I know this configuration normally works... that avoids the bug.. the bug is clearly somewhere else and is likely an edge-case of some sort, or more likely something did not get cleaned up correctly in the tool. I really doubt it's system-create. Please note, I have successfully exported many C2 programmes so if there is something there that is not obvious that may be missing.. I haven't a clue where to look. I thought it was simply export and copy to location?

Attach a Capx

http://149.157.140.178/debug/index.html

http://149.157.140.178/debug/debug.capx

Description of Capx

The capx has 1 event, includes 3 objects. Total size of capx... 7k

Steps to Reproduce Bug

export copy to server click link.

Observed Result

a white screen, one piece of text top-left.

Expected Result

It should change the text on a preloaded text object, create a text object and set it's text, create and resize a sprite. When started, the screen should go black and a text object on the right hand side should appear with text saying "hello text1". there will be a text object top-left that is preloaded and will be set to layout size.

Affected Browsers

• Chrome: (YES)
• FireFox: (??)
• Internet Explorer: (YES)

Operating System and Service Pack

windows 7

Construct 2 Version ID

175

Hi, I got around this my randomly tweaking things, removing things, ya ya ya...

however I hope someone looks at this issue... if you can't trust the local copy to behave the same as the remote copy, then the fundamental concept of Scirra, or any development tool like this, is flawed. I can accept random errors and dependencies causing issues, but the discrepancy between local and hosted versions really has hurt my confidence in the tool.

R

thanks... it was a large project and I just couldn't see it in the big list... thanks!!!

R