jarto

Thank you. I can reproduce the problem. I'll have a look.

Yes, the mustache template was the same. Could you also send a compiled version of the server to jt at smartmobilestudio.com?

Have you tried rebooting?

This is weird. The paths are correct in the dpr and I can connect to the server with the mORMotClient.pas that I can download from http://localhost:8888/root/wrapper If I try the mORMotClient.pas that you attached, the login does not go through. Comparing the two files, it seems that my version just does not require any authentication. I'll attach my version of mORMotClient.pas, which works through connect but does not compute. This is good to know. mORMotClient.pas

@IgorSavkicThank you for the great instructions. I was able to download, compile and run Project14ServerHttpWrapper, but ran into problems after that. GetClient in the downloaded version of mORMotClient.pas differs from the original one that is included in mORMot: //OLD one: procedure GetClient(const aServerAddress, aUserName,aPassword: string; onSuccess, onError: TSQLRestEvent; aServerPort: integer=SERVER_PORT); //NEW one: procedure GetClient(const aServerAddress: string; onSuccess, onError: TSQLRestEvent; aServerPort: integer=SERVER_PORT; aServerRoot: string=SERVER_ROOT; aHttps: boolean=false); The new one has no UserName and Password. When I remove them, I am able to compile and connect to the service. However, I run into issues with this: XMLHttpRequest with the synchronous flag set to true is deprecated due to its impact on user-perceived site performance. It's probably because TServiceClientAbstract.Create does a synchronous call in CallRemoteServiceInternal and CallGetResult returns an unidentified result. fClient.CallRemoteServiceInternal(Call,self,SERVICE_PSEUDO_METHOD[imContract],'[]'); result := CallGetResult(Call,dummyID); And if I ran the compiled app, I am able to reproduce the issue where the internal browser does not load the project. I'll dive into that now, but I'd really appreciate help with this synchronous-issue. @warleyalex

Thank you. Gonna have a look at this.

Eric has not reacted yet, so I'll include my own fix in the next update.

@Daniel Eiszele Thank you, I'll add it. There is also TVariant.ForEachProperty, which can be used for traversing all the properties. It's handy as you get both the name and the value in the same callback. However, it had a bug in it and always returned the value as undefined. TVariant.ForEachProperty(MyVariant, function (Name: string; Data: variant): TEnumState begin WriteLn('Name of the property is: '+Name); WriteLn(Data); result:=esContinue; end); I'm going to include the new TVariant.Values and the bug fix to TVariant.ForEachProperty in the next update.

Do you guys have any advice on how I could help? I myself don't have any mORMot servers that I could use to test against. However, I'd love to get familiar with this and eliminate the problem with the SMS Browser.

The reason this bug happens is, commit 5e92a541002cd781ad335f0e7a4cd64615ca0f0b in DWS: Here the code checks if SetRandSeed is used. If not, it uses Math.random instead of alea and removes Random from DWS's dependencies, which results in alea.js not being linked in. The problem is, that Eric should also have checked if Randomize is used, before removing the dependency. So I basically know how to fix this myself, but prefer to get an official fix from Eric. Currently you can trick DWS to work properly by adding a call to SetRandSeed before calling Randomize.

It's an interesting idea, that we could you a static html page as a start point. I think the challenges would be to identify the controls from the "alien" html and tie them to functionality. With simple buttons and edits it's easy, but more advanced, dynamic components are harder (like UL/OL-lists to ListBox)

I'll add it.

A new update is now available in the alpha-channel. It contains fixes to bugs reported by @JimKueneman IDE: Fix bugs in handling Project Search Path Bug fix to Search&Replace

Yep, this is a bug in DWS. I've reported it to Eric.

Thanks. I'll have a look. Seems like the alea-function is not linked in during compile.