mORMot Open Source

In fact, there is minor issue/bug on older SMS versions 1.2 | 2.0 | 2.0.1 | 2.2 and 2.2.2 - the TObject::Destroy method would just delete the property named "prop" rather than the actually all the properties of the object itself. For instance, obj.Free is doing nothing because the method Destroy was buggy.

var TObject={
    (...)
    Destroy: function (s) { for (var prop in s) if (s.hasOwnProperty(prop)) delete s.prop },
    Destroy$: function(s) { return s.ClassType.Destroy(s) },
    (...)
}

Look it should be: delete s[prop] --> It will remove all properties from JavaScript object itself. The delete operator is more about freeing memory than deleting an element. This issue was fixed on the DWScript trunk. On newer SMS version 3.x, they fixed the TObject::Destroy method.

Some changes are needed on the SynCrossPlatform units, Eric Grange has made some changes on the DWscript, nowadays, in my tests, it compiles and works in both SMS 2.2.2.04694 (lastest WinXP compatible) and SMS 3.1.0.80 (development version)!
I forgot to say, the examples 27 and 29 (specific SMS) should be updated on the current mORMot branch. There is an issue at the example 29. When you open the project, there's an ugly message: "Error while loading project file with the message: WebForm.pas doesn't exist". So, you can not open the project 29, because there's an bug in the SMS 2.x (Make Project/Item external) this issue, you can miss units on the project.

The Pas2JS with Delphi/TMS it's a bit pricey!

1. TMS WEB Core:= 295 EUR introductory price
2. TMS WEB Studio := 595 EUR introductory price (TMS WEB Core + TMS FNC controls + TMS XData)
3. It can be installed in Delphi XE7 to Delphi Tokyo 10.2.x

We proposed to Pas2JS team another "open source" alternative using with the Lazarus IDE as base library.

...just analyzing the TMS Web Core emitted javascript output, and and quite a bit of effort, doing some reverve engineering I believe I've created a clone of the following components:

TLabel, TEdit, TButton, TComboBox, TMemo, TListBox, TSpinEdit, TCheckBox, TPaintBox, TImageControl, TPanel, TSplitter, TGridPanel, TSpeedButton, TToolBar, TDBEdit, TDBLabel, TDBSpinEdit, TDBMemo, TDBNavigator, TMessageDlg, TClientDataSource, TClientDataSet.

     ,  ,  , , , 
    <(__)> | | | 
    | \/ | \_|_/ 
     \^  ^/   | 
     /\--/\  /| 
    /  \/  \/ | 

- there are enough similarities that it's clear that they have the same source. I didn't illegaly copied the source code anyway, since TMS didn't release the product yet; some reverse reverse engineered the exact Pascal code from the TMS JS code, pas2js compiler will create code that looks exactly like the TMS JS code. The Pas2JS team think it could be treated as disassembling instead of reverser engineering. "It's far too near a copyright violation, AKA it was merely disassembled, using this as a base for a library would definitely
give bad publicity", said Pas2JS team.

Any idea about this - the code copyright "javascript piracy"?

I was probably drunk last night, Idera didn't buy FMSoft! (that's bad news for UNIGUI).

about Idera absorbing sencha company.
I remember since the Art Landro has joined the Sencha's company as its new CEO. The worst mistake was letting Sequoia near it. From that point it was all lies and sticky-tape. First decision: immediately, the super architect Aconran and others core employees were laid off and comes with the decision of the diabolic the requirement to purchase "5-dev license for only 5K", and wow, they simply deleted all the comments as "no longer relevant" about ExtJS. Sencha was treating their community like sh*t.
The latest version available to us has known bugs that are fixed already. I remember I fixed some bugs myself and submit countless fixes back to Sencha but don’t get to track their integration via a pull request or use the framework versions that incorporate them in my open source projects. These CEO left a large hole in the boat.

Let’s not pretend this acquisition is good news for anyone using Sencha products. It's indirect confirmed, IDERA has already dismissed all developers and will outsource the development process to places like India, and they have already told the Sencha services guys to look for new jobs. Presumably IDERA got Sencha for a fire sale price and now intends to milk the licensing revenues while they last without investing much. The acquisition comes along with the announcement on Single Developer Licences for Ext (US$960), this is indeed annual subscription license that includes 1 year of development and distribution rights. For perpetual license (minimum 5 licenses for $4,340).
--------
Look, this BIG predator, just care about how they can earn more and eliminate their competitors. I think they will propose a single developer perpetual license linked with Delphi to use ExtJS. That more expensive Delphi licenses await us (bad news). If so, only UniGUI will be the extra cost.

(mORMot with Ext) and (UniGUI with Ext) is like rocket and turtle if we spoke about speed.

Let's suppose you have an iPod, it's so beautiful and easy to play with it, but it is a tight coupling player, once the battery dies you might as well buy a new iPod because the battery is soldered fixed and won't come loose, thus making replacing very expensive. A loosely coupled player would allow effortlessly changing the battery.

Would you rather spend less money just for a battery and have a working iPod again or you would prefer to spend a lot of money for a whole iPod?

It's a trade-off when you try to pick between an tight coupled framework and a loosely coupled player set of tools that work together. The latter gives you more freedom but requires you to make more architectural decisions, it also means the learning curve is high an flexibility suffers. This loosely coupled framework was designed to be highly
optimized to speed. It uses a lot of established conventions/patterns and once you are familiar enough with them, it can make you very productive.

I've got the following response: http://pastebin.com/raw/wpzPZdfx   (see the page is renderized)

The response using https://www.hurl.it/ is:
See the response picture See this image before page render

...I had to just unpack, another clean mORMot and patched the units and... tandãã... it worked as expected!

/\(^_^)/\

Mr. AB could take a look at this code.

I think there an issue on the 2nd reconnection.
http://i.makeagif.com/media/3-09-2017/T2uxmP.gif

I can confirm the issue.
If you enter the correct credentials at 1st try; work as expected.
If you enter the wrong credentials at 1st try; the subsequent tries... it keeps failing - you are not re-authenticated;

AFAIK, currently it is not possible to communicate in raw binary mode (e.g. send raw video streaming), a new kind of record would be necessary to send binary data from the server to the client.
All communication is is done through base64-encoding (using interface based services), this will adds 30% to 50% overhead.

No. This pattern (use interface parameters to interface-based service methods to send a callback to the server) can also be via client TSQLRestClient (which may be over HTTP).

{ login }
GET /root/TimeStamp HTTP/1.1
GET /root/Auth?UserName=Supervisor HTTP/1.1
GET /root/Auth?UserName=Supervisor&Password=e20f974a72b2e457d4d1b070ee7a8130e4e34d4fa2bd50422c684f3281917a84&ClientNonce=7218250DB57A2F308A8E4274E0859CB23F8572D76A6244C1B0FA8B28C451F421 HTTP/1.1
GET /root/TimeStamp?session_signature=012B5CCBF1452A52B3B4A862 HTTP/1.1

{ ServiceDefine on the client side }
POST /root/MyOtherService._contract_?session_signature=012B5CCBF1452A53A3E3DCBD HTTP/1.1
[]

POST /root/MyOtherCallback._contract_?session_signature=012B5CCBF1452A53674BAAE9 HTTP/1.1
[]

{ call the Withdrawal client method which will call the server callback MethodWithCallback(amount,self) }
POST /root/MyOtherCallback.Withdrawal?session_signature=012B5CCBF1452A53332FF6F1 HTTP/1.1
[2000]

{ client release }
GET /root/Auth?UserName=Supervisor&Session=19619019&session_signature=012B5CCBF1452A55EEFF2E27 HTTP/1.1

I can't find a way to invoke the Withdrawal method on the ajax client; for initialization of the registered interface, for instance, i believe the resolve method is required on the client side to work.

Using the latest newPascal, not exacty flawless but I've got good results in a very old computer with WinXP

Total assertions failed for all test suits:  7 / 24,316,034

See: http://pastebin.com/raw/SZB4vPED

Unfortunately I didn't read the book yet. The book itself seems to be of good quality. One thing I really didn't like was the front
cover art of the book. I would expect a demoniac cool marmot illustration on the cover of a mORMot book.

Do the source code files for programs shown in the book are available to download?

The TSQLRest class has a 
   property Services: TServiceContainer read fServices;
   which you can access to the interface-based services list, so can get a list such as:

[{
"contract" : "MyService",
"implementation" : "shared",
"methods" : [{
		"method" : "Add",
		"arguments" : [{
				"argument" : "Self",
				"direction" : "in",
				"type" : "self"
			}, {
				"argument" : "n1",
				"direction" : "in",
				"type" : "integer"
			}, {
				"argument" : "n2",
				"direction" : "in",
				"type" : "integer"
			}, {
				"argument" : "Result",
				"direction" : "out",
				"type" : "integer"
			}
		]
	}, {
		"method" : "GetSampleRecord",
		"arguments" : [{
				"argument" : "Self",
				"direction" : "in",
				"type" : "self"
			}, {
				"argument" : "id",
				"direction" : "in",
				"type" : "int64"
			}, {
				"argument" : "SampleRecord",
				"direction" : "out",
				"type" : "TSQLSampleRecord"
			}, {
				"argument" : "Result",
				"direction" : "out",
				"type" : "boolean"
			}
		]
	}
]
}]

The project is here.

I just had a closer look at this post

It seems there is some interesting concepts about sync that I think could be applied to.

#001. Each client (the mobile app) must decide when to sync with the server, e.g. when a user clicks "DO_SYNC Refresh". The app will sync by calling a method DO_SYNC(). This method will call syncFromServer() first and then syncToServer().  The app will always get new data FROM the server first and then send new data TO the server.

#002. The syncFROMServer() method on the client, will ask the server for all data changes since servercounter_lastsync. The server will answer this request by fetching all objects with object.counter_lastupdate > servercounter_lastsync and sending them to the client. The server will also send its current counter to the client.
The client will store this new value in servercounter_lastsync (because this is the server counter for which the
client has received all data changes).

Example Sync objects of type "booking" FROM server TO client (DOWNLOAD):
GET http://127.0.0.1/root/booking/list/12345

12345 = client.SERVERCOUNTER_LASTSYNC (counter value on server for start of sync)

JSON returned FROM server TO client:

{
	objects : [{
			guid : 65E5B3EB-72B3-4768-B86D-B0D07DF81C3E,
			pk : 2014-05-10,
			name : apples,
			value : 9,
			timestamp_created : 2014-03-20 13:15:00,
			timestamp_lastupdate : 2014-03-20 13:15:00
		}, {
			guid : 77310491-2B68-4EF2-AD8A-5DA75D0BB168,
			pk : 2014-0-12,
			name : oranges,
			value : 3,
			timestamp_created : 2014-03-20 12:18:08,
			timestamp_lastupdate : 2014-03-20 15:17:23
		}, {
			guid : 93F58F2C-8837-4CE3-A3DB-CA0AD01632B7,
			pk : 2014-07-18,
			name : cherries,
			value : 6,
			timestamp_created : 2014-03-20 13:16:32,
			timestamp_lastupdate : 2014-03-20 15:17:58
		}
	],
	result : {
		statuscode : OK,
		SERVERCOUNTER : 453
	}
}

#003. The syncTOServer() on the client, will send all its data changes TO the server since COUNTER_LASTSYNC.
The client will fetch its objects with object.counter_lastupdate > COUNTER_LASTSYNC and send them to the server.
When the server confirms have received the data (result = ok), the client will update its COUNTER_LASTSYNC with
the current value of its counter.

Example Sync objects of type "booking" FROM client TO server (UPLOAD):

POST http://127.0.0.1/root/booking/update

JSON in POST FROM client TO server: 

{
	objects : [{
			guid : 65E5B3EB-72B3-4768-B86D-B0D07DF81C3E,
			pk : 2014-05-11,
			name : apples,
			value : 9,
			timestamp_created : 2014-03-20 13:15:00,
			timestamp_lastupdate : 2014-03-20 13:15:00
		}, {
			guid : 77310491-2B68-4EF2-AD8A-5DA75D0BB168,
			pk : 2014-06-12,
			name : oranges,
			value : 7,
			timestamp_created : 2014-03-20 12:18:08,
			timestamp_lastupdate : 2014-03-20 15:23:25
		}, {
			guid : 93F58F2C-8837-4CE3-A3DB-CA0AD01632B7,
			pk : 2014-07-18,
			name : cherries,
			value : 9,
			timestamp_created : 2014-03-20 13:16:32,
			timestamp_lastupdate : 2014-03-20 15:19:43
		}
	]
}

JSON returned FROM server TO client : 
{
	result : {
		statuscode : OK,
		SERVERCOUNTER : 456
	}
}

The idea of EACH DEVICE (client) will keep track of its sync status using these 2 parameters:
• COUNTER_LASTSYNC: value of the local counter of the client on the last sync TO the server (upload)
• SERVERCOUNTER_LASTSYNC: value of the counter on the server, on the last sync FROM the server (download)

I wish I could implement the Sync using this approach.

I must've drunk two liters of caçacha. I've been posting a lot of non-sense posts.

Apologies for my ignorance, this is simply a javascript callback . Thanks!

I've been thrown head first into a task which I'm feeling pretty disorientated about, trying to piece it all together...

and finally I got successfuly decrypt an encrypted AES-256 string from CryptoJS using SynCrypto

I'm using the tool mORMot/JS/AES to generate the key and IV, I just have to copy 'n paste the key/iv into SynCrypto and you can decrypt the cipherText.

Unfortunately, I could not decrypt the cipherText from SynCrypto

var
  key : TSHA256Digest;
  aes : TAESCBC;
  s    : RawByteString;
begin
  SynCommons.HexToBin(PAnsiChar(
    '33743b03c28fc783b01119d8b8c6b2564108318d465a2fb4ff319010c4aa6493'),@key,sizeof(key));

  aes := TAESCBC.Create(key, 256);
  try
    s := StringToUTF8(edtCrypted.Text);  ----> cipherText from CryptoJS

    edtDecrypted.Text := UTF8ToString(
      aes.DecryptPKCS7(Base64ToBin(s), true)
    );
  finally
    aes.Free;
  end;
end;

I suspect the iv, key and salt should match with the CryptoJS, SynCrypto generates very different output to similar config.

CryptoJS generates similar cipherText
U2FsdGVkX191ODhWOVL5TAGl0gvhK/VqGyyV90UimNg=   // test
U2FsdGVkX1+WwR/zthxXTCId8IE/CiSPmWjhXL/9YGU=    // test
U2FsdGVkX1/upKAyGuoFwDFdz5qrO6izE6TnYnunbYo=    // test
U2FsdGVkX19bNLYeAjWVjLxy0B8D89yW8gzSNEt95dU=   // test

whereas SynCrypto generates different cipherText, such as:

Xb6j9dHzEBqQk49Fnly3jvzRp3yLmeS3G5km5nJi2JM= 
xxVMU/6ytwYKZjuPVeEZP2EB6cEk403OK+WMiQAtvPA=
kpz1uR8oQTJzRIGgln+aplh+jEaJsiBKF/qaLgiQaKw=
AwJNKNx8OH37e7LQhTE2QM3WnpVMY8umnNhNwaLS4r0=
81MSSvWFN91HUeXdoVx8R2Dd0dZgOqFJomRwRIiH8+I=

Hum, dynamic linking. I just applied this trick, and sprint faster than Usain Bolt, mormot or squirrel... to see the tests on the very old computer with Ubuntu 14.04.1 LTS and almost everything is green. This is awesome!

FPC3.1.1+ CrossCompile+i386-linux32 

Nowadays, it's very difficult to find Kylix on the internet, but I it's especially powerful with mORMot.
Most developers would not notice or be concerned, after 16 years, the little mORMot still like this compiler. You can create small executables, the resulting executable just seemed to perform so much better if you compare with FPC.

This is from the NewPascal (a month ago) pack with Lazarus. It look likes there's a new pre-release which I didn't test;

In order to use cross compile from Win32 to Linux I had to unpack the CrossLinuxi386.rar archive in c:\NewPascal

then in project Options
(x) Build Modes (default)
-->Compiler Options
--> Paths --> Libraries (-Fi): C:\NewPascal\cross\lib\i386-linux\; C:\NewPascal\cross\bin\i386-linux\
--> config and Target --> Target OS: Linux
                                    Target CPU: i386
                                    Target Processor: default ;
--> Compilation and linking: ( 1 ) quick, debugger friendly -O1

How do you perform cross compile from win32 to linux succeed using dynamic linking?