mORMot Open Source

2AB: The problem warleyalex say is really exist. In mORMot roadmap we have item "JSONP support", but actually for now is better to support  CORS (https://developer.mozilla.org/en-US/doc … ntrol_CORS) by define public property AllowedDomains on HTTPServer level and handle OPTIONS request type...

2warleyalex: As solution on your HTTP server where static live (let it be myserver.com) configure redirecting querys from URL myserver.com/root* to mORMot_server_IP:port/root* and do AJAX request to URL's myserver.com/root*

AB, I already have code for direct connection to mORMot services in server side from scripting via HTTP (wrapper around TWinHTTP + simple JS class to implement protocol details) (something like http://nodejs.org/api/http.html ) - I rewrite it in new way and shere when we finish refactoring of main part.

When implementing event is good to use EventEmitter pattern (http://nodejs.org/api/events.html ) - the main different from Delphi OnSomthing is:
1) posibility to add sereval different listiners to one event - very important thing. For example I use it on auth event (fired by auth procedure of my mORMot-based server to fill custom user information. One subsystem do App.on('auth', {obtain additional info from LDAP), other App.on('auth', {obtain additional info from database table) and so on.
2) posibility to create new event types "on the fly" (something like sendMessage in windows with custom code), i.e. Class.fireEvent('myNewEvent', param1, param2) and in other part of code Class.on('myNewEvent', function(){do something})

Anders Hejlsberg... I'll look on TypeScript once again!!

The reason I do not implement debugger yet is new Debuger API I waiting from Mozilla - no reason to implement it in old way. As editor I use (with grate success)  old good SynEdit on delphi side and codemirror on web side (even with code insight)

By the way  Trello is fully JavaScript solution - server is NodeJS + MongoDB. Client - CoffeeScript, Backbone.js, HTML5 History API, Mustache

Great work! Just some tips:
1) To increase mORMot performance it is better to set number of thread = CPUs core count *2 (or even = CPUs core count in your case because you do not do any calculation in server side)
2) As AB sad in prev. post most interesting things started then we use some kind of logic in server side (JSON serialisation, Database access) - in this case I think mORMot is MATCH better compared to others (except node, maybe, or very optimized WCF solution what not use .NET connection pull/serialization)
3) I agree with AB - if using Jmeter you need at last 2 client to utilize all server resources - this is the reason mOTMot, Java, Node and WCF performance is nearly the same.
4) To be honest is good to use nodeJS cluster http://nodejs.org/api/cluster.html with the same thread count as in mORMot
Thanks for your work!

In your example mmo1.Lines.Text is string type. But JSONToObject need RawUTF8 as input parameter. RawUTF8 is UTF-8 encoded string, so direct conversoin Putf8char(mmo1.Lines.Text) is not allowed!
So please, read comments in interface part of mORMot.pas where function JSONToObject defined.
In your case correct code is:
var
  R: RawUTF8;
  isValid: boolean;

R := StringToUTF8(mmo1.Lines.Text); // !! IMPORTANT to put Text into buffer (R) and perform conversion from string to RawUTF8!!!
JSONToObject(d, PUTF8Char(R), isValid);
if isValid then
  mmo1.Lines.Add(d.Name)
else
  mmo1.Lines.Add('invalid JSON string');

2Eric:
1) please, look on this ticket http://synopse.info/fossil/tktview?name=73da2c17b1 - the same in THttpApi2Server.Execute
2) do you have plane to implement authentication?

and question 2AB&Eric - do you have plane to merge THttpApi2Server realization to mORMot repository? (even without auth on HTTPAPI level it give us  ability to solve the problem Chaa temparary solve via RemoteIP in this post http://synopse.info/forum/viewtopic.php?pid=5809#p5809 via THttpApi2Server.connectionID)

Additionally I found good GUI util for manage http.sys - see http://httpsysconfig.codeplex.com/.

I compile mORMot with old sources (1.17) - and the problem is the same. So mORMot is OK, 99% that it because of some Windows updates applied to my computer. On other computer everything work as expected on all mORMot versions.
Also I confirm what using ptIdentifiedInOnFile is stable and very good for understanding - match better then log per thread. Thanks for good solutions!

Try it under Linux - it much faster. In my case I have 2 computer - one with mORMot server + database engine and one with Linux - in case I test mORMot with database access ApacheBench utilize 100% of mORMot server CPU in this configuration.
I run something like this:
ab.exe -c900 -n50000 -p getOneRecFromBigTable.json -T"application/json" "http://mash-w7:888/m3/AS/runList"

I redirect password check to LDAP and not store pwd in my database at all. So if user want to change password he/she change it in LDAP (Windows Domain in my case).

If you  register ( THttpApiServer.AddUrlAuthorize ) your URL with HTTPS=false you must unregister it with HTTPS= FALSE and register again with HTTPS= true

so call
THttpApiServer.AddUrlAuthorize('INFO',aPort, FALSE {UseHTTPS} ,'*', true {OnlyDelete});
and when
THttpApiServer.AddUrlAuthorize('INFO',aPort, TRUE {UseHTTPS} ,'*', FALSE {OnlyDelete});

Hope it's help

Hmmm. Really - TSQLRecord.ID is integer. I work with mORMot in more lower level, and in SynDB level it's int64.
About int overflow - In my project there is one ID counter for all table, ant first 3 digit of it is customer number (I need generate different ID for different customer). So yes - I need bigint.
But if I do not reserve first 3 digit int32 is not enough too - in my country live 42million people, everyone pay taxes every month, so int32 is only for 1 year of production usage of level-country tax system, for example....

About storing int - yes - it much easy - I agree with you

such changes require a huge refactoring of mORMot, I think. Let's wait what AB sad. As for me, the best choice is Int64 as it in mORMot now. But I have remarks for you proposition:
1)  TSQLRecordID=integer; is not very good, because if we possibly move to x64 integer become int64 and recreating primary key of existing DB tables is a problem. So TSQLRecordID=int32 must be used in your case;
2) In my project's I fount what int32 is not enough in real life. Really - I have int32 primaty key overflow in some production DB, so to prevent future problems I recommend use Int64 as primary key. But this is IMHO.
About JSON: JSON doesn't like cardinals because of JavaScript internal realization (JSON is primary using in JavaScript). For example in Mozilla SpiderMonkey internal(I think in V8 to) cardinal stored as double. Using Int64 is also a VERY big problem in JavaScript - it also become double, so only Int64 with 16 digits long is usable.

according to documentation, which you can easy found here: http://synopse.info/fossil/wiki?name=Do … umentation (direct link http://synopse.info/files/pdf/Synopse%2 … 201.17.pdf ) mORMot don't use HTTP auth at all. Please, read documentation or see the comments in source codes (in interface part of SQLite3Commons TSQLRestServer.Create). In sample project you are talking about you can easy turn off auth by change this line of code in unit2.pas
DB := TSQLRestServerDB.Create(Model,ChangeFileExt(paramstr(0),'.db3'), FALSE {aHandleUserAuthentication});

You can address 4Gb. See this topic http://synopse.info/forum/viewtopic.php?id=702
I use this feature in real project.

Thanks for changing. I understand issue not in mormot. So you can not sorry for M$
About making ProviderName property writable: yes, I see the constructor and understand what we need refactoring to make it

Thanks!
This peace of code is funny:

if PWord(P)^=ord('*')+ord('/')shl 8 then begin
   PWord(P)^ := $2020;
   inc(P,2);
   break;
end;

I remember for future use. I see similar when you  compare string for false/true, but forgot

in some case I store my program configuration in JSON format. For better human understanding I add comments in config files, like this

/* 
  This is global configuration file
*/
{
   "serverPort": "888", // HTTP server port
   "serverDomainNames": "+", 
   "handleStatic": true, // handle static files. if true - set staticRoot and staticFolder
   "staticRoot": "m3", // appRoot 
   "staticFolder": /*folder for static */ "X:\\Subsystems\\Components\\Clients\\Web\\"
}

But comment is not part of JSON spec, so I write small function to remove comment from string before pass it to JSON parser.
AB, can you add this function to SynCommons ?

/// remove comments from string before pass it to JSON parser
// may be used for prepare configuration files for loading
// for example we store server configuration in file Config.json and want to put some comments in this file
// then code for loading is:
// var cfg: RawUTF8;
//  cfg := StringFromFile(ExtractFilePath(paramstr(0)) + 'Config.json');
//  removeCommentsFromJSON(PUTF8Char(cfg));
//  pLastChar := SQLite3Commons.JSONToObject(sc, PUTF8Char(cfg), configValid);
// handle 2 type of comments:
// starting from // and till end of line and /* .....  */
procedure removeCommentsFromJSON(P: PUTF8Char);
var
  nP: PUTF8Char;
begin
  if P=nil then exit;
  while P^<>#0 do begin
    case P^ of
      '"': begin // skip string
        inc(P);
        while not (P^ in [#0, '"']) do begin
          if P^ = '\' then begin  // escaped "
            inc(P); if (P^ in ['"', '\']) then inc(P);
          end else
            inc(P);
        end;
      end;
      '/': begin
         inc(P);
         case P^ of
           '/': begin // this is // comment - replace by ' '
             dec(P);
             repeat
               P^ := ' '; inc(P)
             until P^ in [#0, #10, #13];
           end;
           '*': begin // this is /* comment - replace by ' ' but keep CRLF
             dec(P); P^ := ' '; inc(P);
             repeat
               if not (P^ in [#10, #13]) then P^ := ' '; // skeep CRLF for correct line numbering (for error for example)
               inc(P);
               if (P^='*') then begin
                 nP := P + 1;
                 if (nP^<>#0) and (nP^='/') then begin
                   P^ := ' '; nP^ := ' ';
                   inc(P); inc(P);
                   break;
                 end;
               end;
             until P^ = #0;
           end;
         end;
      end;
    end;
    inc(P);
  end;
end;

I try to work with  class and property attributes in XE via  new RTTI (TRttiContext e.t.c.) classes and got some problems:
1) An error EInsufficientRtti  occurs - not for all cases new RTTI is present
2) VERY slow performance of new RTTI. Very slow.
3) It not work in FPC

So in my opinion using new RTTI is early now.

I think it`s bad idea. For example, if I install any database server or linux distributive  there is no guest user in it, is it?

You are right about nginx proxy. I use nginx as proxy and load balance with great success in real project. In my current project I use nginx for static files and proxy for mORMot server. If THttpServer become compatible with FPC I will try to use it with nginx proxy - this better as for me than single thread FastCGI solution.
But I don't understand how it possible (use THttpServer  under linux) - TSynThreadPool use IOCP as i see (CreateIoCompletionPort  TSynThreadPool .Initialize). Some wrapper exist in FPC for this call? Or under Linux you made other realization?

FPC support is a greate news! But one question - what about THttpServerGeneric descendants under Linux? I mean THttpApiServer (need http.sys) or THttpServer (need IOCP). If we do THttpServerGeneric without asynchronous IO we need to change architecture from fixed number of thread to unlimited thread pool (it's bad  - every thread in current realization store database connection. creating many DB connection is very resource cost)? Or you have other idea?
For example my lovely HTTP server nginx use this models: http://nginx.org/en/docs/events.html

Also if we migrate to Linux we can do PostgreSQL direct support - under Linux it's a best production DB for now. Almost (but even better) as Oracle in performance on big databases (1Tb and up) but free and OpenSource.

About inserting - I agree with you - this is not good in server side.
Always insert rowCount=000000 (6 digits) is the way. I code this feature, make performance tests and post results here. After we get test result we can decide add this feature or not.
About remove rowCount for small requests IMHO 15 additional bytes it doesn't matter. But if we don't remove rowCount we will get identical behavior in all cases - it is more important.
For example in JS client part I simple write

 if (result.rowCount>0) {...} 

If rowCount is optional I must write something like this

 if (result.rowCount?(result.rowCount>0):(result.data.length>0)) {...} 

If we write "rowCount": xx in TSQLTable.GetJSONValues and in TSQLTableJSON.ParseAndConvert analyze it exists and if exist - use rowCount value instead of calculating row count this be enough for me.