Jump to content

A.M. Hoornweg

Members
 View Profile  See their activity

  • Content Count

    446

  • Joined

  • Last visited

  • Days Won

    8

Posts posted by A.M. Hoornweg

  3. Posted

    OK, this is getting even weirder. 

     

    I've placed a tTimer on the form plus an event handler to display the size of the 3 images.  The size of the bitmap is 78x98 pixels in all 3 cases.  

      

    procedure TForm25.check(img: timage; lbl: tlabel; fmt:String);
begin
  lbl.Caption:=format('%s %d x %d',[fmt, img.Width, img.height]);
end;

procedure TForm25.Timer1Timer(Sender: TObject);
begin
 check (image1,label1,'JPG');
 check (image2,label2,'BMP');
 check (image3,label3,'PNG');
end;

     

    Now look what happens if I drag the form to a high-resolution screen (250% zoom). 

     

    The tImage having the JPG did not resize but claims to have done so anyway.

    The tImage with the BMP scaled correctly but still returns the unscaled size.

    And finally, the tImage with the PNG scaled correctly and returns the scaled size.

     

     

    zoom100%.png

    zoom250%.png

  4. Posted · Edited by A.M. Hoornweg

    Hello all,

     

    does anyone else have the issue that a TImage containing a JPG does not re-size when the form is dragged to a screen with a different resolution?

     

    The weird thing: This only happens when the contained picture is a *.JPG.  If the picture is a BMP or PNG then it *does* work OK. 

     

    See attached image: The property "Stretch" is set to True in all 3 cases.

     

    (Edit) compiler version is Delphi 11.1 Alexandria.

     

     

    image_resize.png

  7. Posted

    15 hours ago, Remy Lebeau said:

    Your example is checking for 4 bytes, not 8 bytes.  And also, presumably @TokenStart should be FTokenStart, which is already a pointer. But, in any case, if the strings are greater than SizeOf(UInt64) in length, you are only using those first 8 bytes for the "key", so you might not end up with unique keys for different string values.

    No, "{assuming unicode string}" -> 1 char is 2 bytes so I'm actualy testing for 8.

     

    OP stated  "Usually, the token is part of a larger string being parsed."  So I assumed OP wanted to do a quick lookup in a token dictionary to figure out if there are any entries starting with that substring in the actual dictionary. 

     

     

  8. Posted

    43 minutes ago, MarkShark said:

    Hi all!  I often find myself needing to look up a string token in a dictionary.  Usually, the token is part of a larger string being parsed.  Currently I take a copy of that portion of the string using SetString or Copy and use that new string for the lookup.  The code looks something like this:

      

    
  // Not shown: parse some source string and get a pointer to the first char of the token and the length of said token
  // Now get a copy of that token so we can look it up in the dictionary
  SetString(TempToken, FTokenStart, FTokenLength);
  if not FTokenDictionary.TryGetValue(TempToken, Result) then
    Result := UnknownToken;

    I'm wondering if there's any way to avoid the SetString part of this.  Thanks for any ideas!

     

    You could make the "key" in your tokendictionary a UInt64 instead of a string, but you'd also have to check that your strings have a minimum length of 8 bytes or else you'll get an access violation.
      

    Type pUint64=^Unit64;

{assuming unicode string}
if (fTokenLength >= 4) then  
  if not fTokenDictionary.TryGetValue(pUint64(@Tokenstart)^,result) then ...

     

  9. Posted

    28 minutes ago, Attila Kovacs said:

    Or you can run it as a standalone application as well, sharing the units.

    Yes, but the behavior might not be 100% identical because IIRC in DLL projects some things are not allowed in the initialization and finalization sections of units.  But having said that, I usually test like this as well. A plain vanilla executable is the easiest to debug.

     

  10. Posted

    39 minutes ago, Attila Kovacs said:

    I don't think that attaching 3rd party processes is the official way to debug isapi or other services. Is it?

     

    I believe it is, but there are simpler ways.   IIS is not the only web server that can run ISAPI's. For debugging ISAPI's I personally use Aprelium's Abyss X2 web server. This can run as a normal executable and it comes in an x64 and x86 flavor. In the Delphi IDE I simply specify this program as the host process for my DLL.  

  12. Posted

    Duh.   It looks like we got bitten by visual form inheritance.   The tDatetimepicker is on a base tForm that is inherited from. This base form is perfectly OK but I see in an inherited DFM that "tDateTimePicker.Kind" is reverted to dtkDate.  I can't explain how this happened. 

     

     

  15. Posted

    OK, this took me some time to figure out, but it is actually possible:

     

    - Put a tDateTimePicker on your form

    - Set property "Kind" to "dtkDateTime"

    - Set property "Format" to "dd/MM/yyyy HH:mm:ss"       (the capitalization of MM and HH is important!)

     

    -Setting the tDateTime is easy, just use property datetime.

    -Use the following routine to GET the tDatetime.

      

    Function ExtractDateTime(Picker:tDateTimePicker) :tDatetime;
var buf:pWideChar; LastBlank,i,len:integer;
    Text,TimePart:string;
begin
     len:=picker.GetTextLen+1;
     buf:=Stralloc(len);
     picker.GetTextBuf(buf,len);
     Text:=buf;
     strdispose(buf);

     LastBlank:=length(Text);
     for i:=length(Text) downto 1 do
     begin
       if Text[i]=' ' then
       begin
         LastBlank:=i;
         break;
       end;
     end;

     TimePart:=Copy(Text, LastBlank+1, length(Text));
     result:=picker.Date+strtotime(TimePart);
end;

     

     

     

     

     

  17. Posted

    15 hours ago, jeroenp said:

    This is true: some of the time it is up.

     

    Over the last 24 hours downtime is "more" than 55% (not the kind of SLA I would be satisfied with, but hey: I'm not Embarcadero IT), see the saved status at https://archive.ph/2HXRI:

    From the history on that page (or browsing the live status at https://stats.uptimerobot.com/3yP3quwNW/780058619), you can see that both uptime and downtime periods vary widely between roughly 5 and 90 minutes.

    To me that sounds like intermittently failing or underdimensioned hardware.

     

     

     

    No it is not.

     

    Guys, just take a look at the PHP exception trace.  More specifically, the line that reads  "LoadBalancer->getConnection(0,Array,false)", that's where the shit hits the fan.

     

    That line triggers a "connection refused" exception. One of the servers which the load balancer can choose from in this array is inaccessible. 

    Not all of them are inaccessible because when you press F5 often enough in your browser the load balancer will sooner or later hit one that works. 

     

    So this is a simple web site management issue.  The person in charge should simply remove broken db servers from the list which the load balancer uses to choose from.

     

     

    So... Why the heck does Idera not manage this properly?  

     

     

     

     

     

     

     

     

     

    loadbalancer.png

    • Like 2
    • Thanks 1

  18. Posted

    11 minutes ago, Lajos Juhász said:

    We that already using Delphi should know to work without a proper documentation.

    I was hoping to save some time by finding out if an existing class is suitable for my needs. More specifically, I wanted to know if tStringbuilder has functionality to use it as a fifo buffer.  

  19. Posted · Edited by A.M. Hoornweg

    11 minutes ago, Lajos Juhász said:

    There is no need to panic. It's out for about 2 two weeks. There was a recent webinar from Ranorex that modern IT companies now have a monthly schedule for release. So before we all panic we have just to wait another 2 weeks and see whenever that time frame is used in Idera or not.

    Breaking a production system in the process of releasing something is simply not done. One first gets the new system ready, only then is the old one taken offline. A two weeks outage is dramatic with respect to the perceived reliability of a company.

     

    (edit - typo)

  20. Posted

    How long has it been now that the docwiki is dysfunctional, one week? Two? 

     

    The site will only work once in five attempts or so, all the other attempts show an error log saying the database backend can't be reached.  There's some kind of load balancer in front that doesn't know which of the servers behind it is actually working and which is not....   The end result is like Russian Roulette ! 

     

    @Embarcadero: Guys, this is how not to do load balancing. At least make sure your load balancer knows which servers actually work!

     

     

  21. Posted

    26 minutes ago, Uwe Raabe said:

    That is quite some valuable information. 👍

    Thanks.  I have stumbled upon a ton of bugs in dpi-awareness so far but I haven't had the time yet to report them to QC because that usually requires test cases.   

     

    Another subtle one: If you use tScrollbox, you need to set tScrollbox.Vertscrollbar.Position to zero before a DPI change, otherwise some components on the tScrollbox may become unreachable after the scaling (the ones near the top). 

  22. Posted

    I'll see if I can concoct an example next week. It's hard to extract from my current program. I can tell you that it definitely happens in some forms of mine that have some panels with constraints and other panels with align=alClient, it causes these forms to "explode" when they are dragged between monitors. 

     

    By trial and error I could stop it from happening by protecting WMDPICHANGED against recursions. 

      

    procedure tMyForm.WMDPIChanged(var Msg: TWMDpi);
begin
  inc(fChangingDPI);
  try
    if fChangingDPI = 1
    then
      inherited    
    else
      Msg.Result := 0;
  finally
    dec(fChangingDPI);
  end;  
end;

     

  23. Posted

    One observation: 

     

    in unit Vcl.Forms there's an optimization in tCustomForm.WMDPICHanged which ensures that a form is only re-scaled when Message.YDPI is unequal to fCurrentPPI.

     

    Assuming that fCurrentPPI may not be perfectly reliable, I have removed that optimization just for testing.  And bingo, my form now appears to re-scale properly every time.

      

        //unit vcl.forms.pas method tCustomForm.WMDPICHANGED
    
    if (* (Message.YDpi <> fCurrentppi) *) and Scaled then
    begin
      DoBeforeMonitorDpiChanged(fCurrentppi, Message.YDpi);
      OldPPI := fCurrentppi;
      ScaleForPPIRect(Message.YDpi, Message.ScaledRect);
      FCurrentPPI := Message.YDpi;
      fPixelsPerInch := FCurrentPPI;
      DoAfterMonitorDpiChanged(OldPPI, FCurrentPPI);
    end;

     

     

     

    Also I've found out that constraints may be problematic in Delphi.

     

    Normally, when a form is dragged to a new monitor having a different scaling factor and its center reaches the new monitor, Windows sends a wm_dpichanged message to the window's wndproc to inform it of the new resolution and also gives it a pointer to a rectangle containing a proposed new position and size. The form must then redraw itself using the proposed position and size to ensure that  the scaled window is definitely on that new monitor.

     

    However, constraints in VCL components can enforce size limits to components and to the form itself.  So it may very well be that after the positioning and resizing, the center of the form is suddenly back on the previous monitor ...

     

     

    • Like 1

  24. Posted

    Hello World,

     

    I'm trying to debug some dpi-awareness problems in an application of mine (using Delphi 11 on a Windows 10 X64 VM).  I have a 3-screen setup with different scaling factors for testing (from left to right: 300%, 100% and 175%). I test dpi awareness by dragging my main form between these screens. 

     

    Most of the time the re-scaling kinda works (albeit slow), but sometimes things get wonky and I wanted to know why. This main form of mine is a rather complex one and it needs some time to redraw when the screen DPI changes. I'm trying to reduce that amount of time but in this case the VCL components themselves are kinda sluggish.

     

    For diagnostic purposes I have overridden the method tForm.DoAfterMonitorDpiChanged() to output some debugging information and the outcome puzzles me.

      

    procedure tMyForm.DoAfterMonitorDpiChanged(OldDPI, NewDPI: Integer);
begin
  inherited;
  OutputDebugString(Pchar(Format('Currentppi / Pixelsperinch  %d / %d',[currentppi, pixelsperinch])));
end;

    Most of the time the values of CurrentPPI and PixelsPerInch are identical and reflect the value of the screen where the form was dragged to, but occasionally they go out of sync. 

     

    When that happens, one of the two contains the DPI of the new screen where I moved the form, whilst the other one reflects the dpi of the screen where the form used to be. This happens especially if the dragging was "fast".

    When I look into the source code of unit vcl.forms, I see that the developer usually sets both properties to the same value. So I assume this behaviour to be a bug. Also, I fail to understand why tForm needs both these properties if they are intended to always be the same value. Unfortunately I can't consult the on-line help because https://docwiki.embarcadero.com has been down since yesterday.  Can anyone shed some light to this, and maybe try to reproduce the behavior?

     

     

     

     

    dpi out of sync.png

×