mORMot Open Source

Even if I use date only, milliseconds part is fixed in some random number:

  isotime.FromText('1962-11-13');
  dt := isotime.ToDateTime; // fails, because of millisecond part

function TSynSystemTime.ToDateTime: TDateTime;
var time: TDateTime;
begin 
  if TryEncodeDate(Year,Month,Day,result) then
    if TryEncodeTime(Hour,Minute,Second,MilliSecond,time) then  // fails !
      result := result+time else
      result := 0 else
    result := 0;
end;

thank you

I see, very nicely spotted.

thank you

hi,

What about hash performance when it is copied from another list?

listnormal := Collections.NewList<TDateTime>;
listnormal gets populated ...
// want to copy from normal list to fast
listfast := Collections.NewList<TDateTime>([loCreateUniqueIndex]);
listfast.AddFrom(casi); // copy 
   // lots of iterations in some loop
    if listfast.IndexOf(cas)<0 then  // is slower when using AddFromversus just looping through listnormal and .Add
      anotherlist.add(cas);

Do we need to rehash it?

Hi,

Is it possible to convert to loCreateUniqueIndex after mylist := Collections.NewList<TDateTime> creation ?
maybe: mylist.UniqeIndex := True;

thank you

Hi,  (D10.4.2 ent)

Is it possible to use IList<> for records ?
It reports Type is too complex.

thank you,
Vojko

thank you,

is there any other way ?
because we have lots of situations where ridiculously long timeouts stuck the app,
not gui, but termination of threads, restart of threads, ... getting user to see what is happening.

I don't know how others deal with situations when network breaks (which is normal situation), then you retry until it reestablishes the connection.

thank you

The properties are set through SetInt32Option

    // open client connection
    if ConnectTimeout>0 then begin
      SetInt32Option(result,SO_RCVTIMEO,ConnectTimeout);  <== ConnectTimeout = 3000
      SetInt32Option(result,SO_SNDTIMEO,ConnectTimeout);  <== ConnectTimeout = 3000
    end;
    if Connect(result,sin)<>0 then begin    <== waits around 20 - 30 secs
       CloseSocket(result);
       result := -1;
    end;

Great,
Thank you

Thank you,

The exception is generated serverside and trapped on clientside:

server side:

procedure TSomeService.myProcWithException(... ; out Err:TSomeEnum);
begin
  ...
  ...
  if SomeCondition then begin
    Err := MyValueForError;   //==> here I set it serverside, but it's value isn't transmitted to client(it stays default enum) !
    Raise Exception.Create('my error message');  //==>
  end;
end; 

client side:

  try
    ...
    RestSvc.myProcWithException(...., Err);  //== here exception happens, but Err is not valid
    ...
 except on E: Exception do begin
    ReportErrorValue := Err; // here err is not valid !
 end;
 end;

Question is whether out parameter Err should be transitted as set on server side in
case of exception.

Thank you, Vojko

Thank you for replies.

Indeed sicShared only has one instance and does not lock anything.

-- If you have some very long process
It's not so much about long process, here's situation where's a device that must have serialized access (which happen a lot in real life)
and mwthods need to be serialized (wait in que from multiple sources). So we need some kind of control: number of calls in que, timeouts, priorities...

Maybe isn't crazy idea to implement some kind of serialization of per method calls.
It's already impelented with SyncWithMainThread actually (what we use in production succesfully for several years )
that was done gratiously by ab.

Regards, Vojko

I've implemented variant in objects and it work's in production. It can translate even variant arrays
and it saves actual variant type.
Saved string is in hex. It's not JSON friendly but it could be used in some special cases.

we need badly streaming of fullblown variants:
check http://synopse.info/forum/viewtopic.php?pid=7817#p7817

Streaming components into/from JSON? -> Yes
Storing TPersistent properties of a TSQLRecord ORM table? -> Yes

Yes it can be BLOB storage.

thank you, vojko

Hopefully you'll add variant arrays . That would be really nice.

thank you, vojko

It's working now. Thank you.

superbe.

Where can I get these files ? Tried in downloads but there are no changes in there.

maybe I'm missing something., (Delphi2009ent)

We try to use TCrtSocket.
It is said is faster that Indy and blcksock, but the use of latter are running ok in
case of network disruption.

1. Issue: I cannot call close (it's private), if I want to -> I need to recreate object.

2. serious issue     

    FTcp.OpenBind(Host,IntToStr(Port),false);  // this one works ok , it says cannot open socket ....
    // this one is strange
    FTcp.Open(Host,IntToStr(Port)); // after we have access violation ????

How can this be?
We just try to call same FTCP.Open(...) over and over untill it really opens and why AV comes in ?

please note that we used succesfully blcksock in same scenario and in production.

 
3. Could OpenBind have shorter timeout delay (no need to be so long) ?

ok. I undestrand. Thank you.

2. So in current implementation there is no way to have common session, right?

Is there any way I could manage several sessions serverside:
like my own ID for application which has several SetUser('',''), but knows they belong to same group ?