mikerabat

I'm not sure about exactly this library but some predecessors indeed were able to sort pointer list whereas the "key" is a float/int 32/64 . The key can be specified as an "offset" to the starting address -> this allows to sort pointers to records/classes.... at least that is the way they do it in "Fast quicksort Implementation using AVX Instructions" The definitions in the library for AVX512 are actually using arrays of templates so I guess that means that you can sort records right not only Double/float... ? This problem is solved by an algorithm that allows to extract the kth largest element (not only the median) -> basically it does the partitioning of quicksort but does not sort hte "unnecessary" half. That in fact allows to pick the median in O(n) (roughly) time.... And if you want to go fancy you can do a rolling median over an array of doubles. The algorithm can be found here: https://github.com/mikerabat/mrmath/blob/master/RollingMedMean.pas

I just released an assembler optimzed (AVX2 set) Base64 encoding/decoding unit. check it out under https://github.com/mikerabat/fastbase64/ My reference implementation achieves up to 10 times speedup against the reference Indy implementation. Let me know what you think or if you encounter some errors...

@David Heffernan That would be against the direction I develop here - first I need the working ASM 😉 then the tool can use the disassembly to create the db instructions....

Thanks for the input - I will beef up the tool to get at least the else part aligned on the right side 😉 that really looks nicer.

Jus FYI: That is intentional - the left side shall show the original assembler call the right side the DB translation -> my tool, that derrives the DB statements from the assembler code does a line by line comparison.

I'm puzzled... Since Delphi 11 the tab key does not work as it has been the case in all the previous versions... I tried to setup the editor options the same way as my Delphi 2010 (or 10.4) installation but the tab key behaves differntly.... What I actually want is that when I press tab the cursor shall be placed after the last non whitespace from the previous line e.g. if you have var i : integer; begin |<- cursor here now when pressing tab the cursor shall jump the character after the n Also in Delphi 2010 it was standard if: if abc then begin while i < 10 do begin |<- e.g.: cursor here end; now the first tab jumps just right below the b of begin, a second tab to the end of begin That was standard for us for up until Delphi 10 ... but now in Delphi 11 I cannot get this working. it either works or it jumps the number of characters that has been defined in the TabStop field. anyone has a clue what there coudl be the problem? A also have DDevextensions and gexperts installed.... kind regards Mike

That is actually interesting... If I use the same settings I got a different result (aka the annoying one). (sorry I cannot create such a neat little video 🙂 ) I'll check if GExperts or DDevextension interfere here...

Sorry... the tab key is broken or at least behaves completly different than in Delphi 2010 (or afaik Delphi 10.4) which is ... anoying. e.g.: Consider the following code Result := Format('%s;%s;%s', [ DateToStr( fDay, Fmt), TimeToStr( fFrom, Fmt ), TimeToStr( fTo, Fmt ) ]); when you place the cursor just behind the '[' bracket and then press enter the cursor is placed just below the R of Result... so far so good. But... In previous Delphi versions - when tab is invoked - the cursor jumps to the ":" then to "F" of format. Now the tab just moves the cursor by 2 (acording to the setting) which is actually anyoing. Another case: Consider the following code (we like to have the then else block that way in case the statement is only one line aka without begin end): if fLunchLen > 0.01 then Result := Result + Format(';%s;%.1fh', [TimeToStr(fLunch, Fmt), fLunchLen], Fmt) else Result := Result + ';;'; // no free lunch... it bloody friday! When you place the cursor behind then, press enter -> press tab in previous delphi versions the cursor was placed just right below the then... now it only jumps by 2. any comments on how to get that old behaviour back is highly appreciated... kind regards Mike

sorry that doesn't help ... I have to hit tab quite often such that it jumps to the desired place. Anyone else that has problems with that? Or ... does anyone excatly know what changed in this version?

You could do something like CBOR encoding does. Basically there they encode values based on the NEEDED space. If integer then encode it in an int8 - to int16. You can also check if the encoding would fit into a 16bit float (half a single) In addition it seems that a differential approach for each block could be feasable ... does it change much?? And of course at the end just compress the stream 😉

First: how often does the license change? Never had any case like that... Even windows update did not change it... Actually the articel that is referenced in the code has a few hints for such unique ID's.

Hey! We actually use the following code to get the system serial - maybe that can help: function GetSystemSerial : string; var FSWbemLocator : OLEVariant; FWMIService : OLEVariant; FWbemObjectSet: OLEVariant; FWbemObject : OLEVariant; oEnum : IEnumvariant; iValue : LongWord; begin CoInitialize(nil); try result:=''; FSWbemLocator := CreateOleObject('WbemScripting.SWbemLocator'); // ATT: we had a case, where on a Win2016 server, the system service failed to // execute this command. Changing the service to a local user solved the problem. // There has been 2 cases where the WMI service was corrupted or was not running. // -> to check: enter WmiMgmt.msc in the console and right click on the local wmi - control -> Properties // if an error occurs the repository seems to be corrupt // one time it helped to reset and rebuild the repository: // https://techcommunity.microsoft.com/t5/ask-the-performance-team/wmi-rebuilding-the-wmi-repository/ba-p/373846 try FWMIService := FSWbemLocator.ConnectServer('localhost', 'root\CIMV2', '', ''); except on e : Exception do begin OutputDebugString(PChar('WMI service not properly running: '+e.Message+#13#10+ 'https://techcommunity.microsoft.com/t5/ask-the-performance-team/wmi-rebuilding-the-wmi-repository/ba-p/373846')); raise; end; end; // todo: if we still support winxp this wmi class is not supported?!? FWbemObjectSet:= FWMIService.ExecQuery('SELECT UUID FROM Win32_ComputerSystemProduct','WQL',$00000020 {wbemFlagForwardOnly}); oEnum := IUnknown(FWbemObjectSet._NewEnum) as IEnumVariant; if oEnum.Next(1, FWbemObject, iValue) = 0 then result:=String(FWbemObject.UUID); finally CoUninitialize; end; end; There is also another place where you could get some unique id: // from http://www.nextofwindows.com/the-best-way-to-uniquely-identify-a-windows-machine with OpenHKLMReadOnly('SOFTWARE\Microsoft\Cryptography', True) do begin AktMachineGuid:=ReadString('', 'MachineGuid', ''); Log('Machine guid: ' + AktMachineGuid); Free; end; hope that helps! kind regards Mike

Hi all! In our system we load a bunch of files on startup in parallel so for each of these files we craete a thread that does that in the background. basically this is what how it is done: TEVentChannel is the data object that holds the data -> on the data object creation a thread (TEvtLoadThread = class(TThread)) is created which does the loading -> of course it is ensured that data access and loading is mutal exclusive. The snippset of code that is involved is: constructor TEventChannel.TEvtLoadThread.Create(parent: TEventChannel); begin fParent := parent; inherited Create(False); end; procedure TEventChannel.TEvtLoadThread.Execute; begin NameThreadForDebugging('Load '+AnsiString(fParent.ChanName)); // ########################################### // #### Just start loading the data try fParent.Reload; except on E : EChannelException do begin fParent.fLoadFailed := False; fParent.fDataLoaded := lsBadRecId; end; else // damaged file?? fParent.fDataLoaded := lsAll; fParent.Clear; fParent.fLoadFailed := True; end; end; My problem here is that on a system on which that code ran for over 2 years without problems now happens to fail with the error "Cannot call Start on a running or suspended thread" This error is issued in the TThread.Afterconstruction call and it seems that our system fails to create that thread... Looking at the TThread code it is most likely that the Thread could not be created so what on earth could be could interfere here? We tried to disable FSecure (or at least they claim no security system is monitoring our program) and Windows Defender without any change. Is there anything known out there that could limit an external program in such way? kind regards Mike

Thanks for all the input. Finally it turned out that a white listing software from https://seculution.de/ was the problem at all. They thought we did some bad stuff and their heuristics failed miserably. Please note that they also tried to just disable that software to rule them out but disabling was not enough ☹️

Thanks Remy!! Right... the call to BeginThread actually suceeds - the callstack clearly indicates that it's the ResumeThread call that fails. So... the only thing I found online was that there is a right " THREAD_SUSPEND_RESUME " which may have some influence here... But I have absolutely no idea what could change that. Note that the code actually works on around 8k of installations but there are only a few (2 for now) that seem to make trouble in that regard. The thing is that if there are way fewer files to be loaded (e.g. 6-8) it works most of the time (aka not always)... Also the system worked for 2 years now without any problems... so I'm quite stunned that now something pops up. So... Is there any 3rd party or any configurable restrictions on Windows that disallow something like releasing many threads at once? kind regards Mike

Short answer ... yes - if you got simultaneously more connections from clients than you have requests handles in Delphi then an exception will be raised... Actually I find that property very stupid at all... e.g. apache creates on startup those 50 threads beforehand waiting for incoming connections. On each request a WebModule object will either be taken from the cache or created and put into the cache so this number should be either growing to the demand or be set to the number of the maximum active connections the server is allowed to serve.

Hey Jo! Actually there is already a 64bit compiler so.... I don't know what you are waiting for 😉 anyway.... the probelm is not the compiler but rather the api used aka WinRT and the fixes are put in there in Delhi11 . So either update to this version or check the fixes. I can remember that I also put them on stackoverflow: https://stackoverflow.com/questions/61890445/delphi-bluetooth-le-heart-rate-profile-access

Actually the fix I posted to Embarcadero was exactly that and a few things more. Basically they used an outdated API and needed pairing. The code there should fix that. Let me know if you can find it - if not let me know. Maybe I can send you the fixed Windows BlueTooth LE stuff directly (I guess I may not post stuff here 😉 ) kind regards MIke

Which Delphi version are you using? I once fixed a few things in Bluetooth LE here https://quality.embarcadero.com/browse/RSP-20296 - let me know if that might help you. kind regards Mike

Dear all! Our Software supports new uploading large files (100MB + ) to a server. To get a nice feedback we wanted to have some kind of progress telling the user how far we are already with the upload. For the upload we use the TidHTTP component + an openssl io handler. To get the progress working we used the OnWork event that is published by the component. While we were using the component for testing in our internal network (which is quite fast) the progress seemed ok but when we tested that on a slow network (around 2MBit) we saw that the first 99% were very fast (20MB per second +) and then the last bytes "hang" until the full file was transmitted. Basically the progress bar showed a progress of 99% for around 4minutes until the file was finally uploaded. What am I missing here? How can I determine how much data is transmitted to the server? The code involved is: function CreateLibertyHTTP( method : TIdSSLVersion = sslvTLSv1_2; proxySet : ISuperObject = nil ) : TIdHTTP; var sslIOHandler : TIdSSLIOHandlerSocketOpenSSL; begin Result := TIdHTTP.Create(nil); sslIOHandler := TIdSSLIOHandlerSocketOpenSSL.Create(Result); sslIOHandler.SSLOptions.Method := method; Result.IOHandler := sslIOHandler; if (proxySet <> nil) and (proxySet.S['Server'] <> '') then begin Result.ProxyParams.ProxyServer := proxySet.S['Server']; if proxySet.I['Port'] <> 0 then Result.ProxyParams.ProxyPort := proxySet.I['Port']; if proxySet.s['Username'] <> '' then begin Result.ProxyParams.ProxyUsername := proxySet.s['Username']; Result.ProxyParams.ProxyPassword := proxySet.s['Pwd']; end; Result.ProxyParams.BasicAuthentication := proxySet.B['BasicAuthentication']; end; end; httpReq := CreateLibertyHTTPReq( some params here ); DL.LLog(dlInfo, 'Liberty upload'); libertyParams.Clear; libertyParams.Add('SID=' + SessionId); libertyParams.Add('blockearly=1'); libertyParams.Add('csrfToken=' + csrfToken); // eg 'http://192.168.0.193:81/tom24x/uploadFile?SID=45784095&blockearly=1' s := httpReq.Post(LibServer+'tom24x/uploadFile', libertyParams, IndyTextEncoding_UTF8); if Pos('error', LowerCase(s)) > 0 then raise Exception.Create('Error initializing upload'); uplFileResp := SO(s); fUploadFileSize:=FileSizeByName(DawFileName); //s:='?SID='+SessionId+'&blockearly=1&overwrite=1'; // overwrite recording in case it's already uploaded Stream := TStringStream.Create(''); Params := TIdMultipartFormDataStream.Create; httpReq.OnWork:=LibertyUploadProgress; try Params.AddFile('pdf', DawFileName, 'application/octet-stream'); // New in 2.9.3: The Uploadfile issues a token which can be used as session id (one time thing) // so the param does no harm at all. s := LibServer+'rec/UploadRecording.exe?SID=' + uplFileResp.S['session'] + '&blockearly=1&overwrite=1'; // e.g. 'http://192.168.0.193:81/rec/UploadRecording.exe?SID=45784095&blockearly=1&overwrite=1' try httpReq.Post(s, Params, Stream); except on E: Exception do raise Exception.Create('Sending DAW file failed: ' + E.Message); end; DL.LLog(dlInfo, 'Liberty says: '+Stream.DataString); procedure TfrmDBBrowser.LibertyUploadProgress(ASender: TObject; AWorkMode: TWorkMode; AWorkCount: Int64); function BytesToStr( numBytes : Int64 ) : string; begin if numBytes > 1000000 then begin Result := Format('%.2f MB', [numBytes/(1000*1000)]); end else if numBytes > 10000 then begin Result := Format('%d kB', [numBytes div 1000]); end else Result := intToStr( numBytes ) + ' B'; end; begin if fSxProgress.Cancelled then Abort else begin // note: the upload file size may be a bit higher than the actual file size due to header overhead fSxProgress.lblTopLabel.Caption := fLibertyUplLabel + ' ' + BytesToStr( Min(fUploadFileSize, AWorkCount ) ) + '/' + BytesToStr( fUploadFileSize ); fSxProgress.lblTopLabel.Update; fSxProgress.SetProgress(-1, MulDiv(AWorkCount, 100, fUploadFileSize)); end; end; kind regards Mike

Strange enough on a second installation (aka our test site) the OnWork property works as expected... uploading around 1MByte per second and not waiting at all at the end. It looks like there is some kind of proxy in between that first receives the full message - relays that to the actual server - and waits until something comes back... I have though never heard of something like this and they couldn't clarify that either... Is there actually some kind of hardware around that could reproduce such a behaviour?

That is only the one half of the story. This is the base - in the next abstraction layer Emba decided to go to the different platforms - and for windows and Bluetooth LE the last file that actually implements stuff is System.Win.BluetoothWinRT.pas That is also not entierly true. The imlementation from Embarcadero lacks this ability yes - I managed though to extend the WinRT implementation such that this is possible too. See the attached file on: https://quality.embarcadero.com/browse/RSP-21711 hope that helps 😉

And how do you do that in win32? I wasn't able to find any BT LE classes or functions besides the WinRT stuff that is covered in Delphi. Please note that my code here patches the actual TBluetoothLE component that ships with Delphi 10.4.2! And yes I use this code in conjunction with a standard VCL app 😉

I ran into the same problem and managed to fix that with an external timer that calls CancelDiscovery. It works but leaves a massive memory/thread leak since every time a new discovery is started the timer variable gets overwritten and not freed leaving the pending thread and it's resources.... The problem actually is: the problem is within the Delphi implementation of the Thread that triggers the CancelDiscovery function. In Windows there is a Thread (TWinRTBluetoothLEAdapter.TDiscoverThreadTimer in System.Win.BluetoothWinRT.pas) that actually just waits for a specific amount of time and calls the given timer procedure. The problem here is that the Timer procedure cleans up the thread (calls .Free) which waits for the end of the Thread -> deadlock. To circumvent that problem I needed to adjust System.Win.BluetoothWinRT do the following: Augment the class TWinRTBluetoothLEAdapter by private const WM_BTLE_CANCELDISCOVERY = $400 + 666; // WM_USER ... private fHDL : HWND; procedure BTLETimerHandler( var msg : TMessage ); protected procedure ThrDoCancelDiscovery; .. implementation procedure TWinRTBluetoothLEAdapter.BTLETimerHandler(var msg: TMessage); begin if msg.Msg = WM_BTLE_CANCELDISCOVERY then DoCancelDiscovery; end; procedure TWinRTBluetoothLEAdapter.ThrDoCancelDiscovery; begin // just post otherwise the thread hangs! PostMessage( fHDL, WM_BTLE_CANCELDISCOVERY, 0, 0); end; constructor TWinRTBluetoothLEAdapter.Create(const AManager: TBluetoothLEManager; const ARadioInfo: TBluetoothRadioInfo); begin inherited Create(AManager); fHDL := AllocateHWnd(BTLETimerHandler); ... end; destructor TWinRTBluetoothLEAdapter.Destroy; begin DeallocateHWnd(fHDL); ... end; function TWinRTBluetoothLEAdapter.DoStartDiscovery(Timeout: Cardinal; const AFilterUUIDList: TBluetoothUUIDsList; const ABluetoothLEScanFilterList: TBluetoothLEScanFilterList): Boolean; begin ... // new code if Assigned(fTimerThread) then begin fTimerThread.Cancel; fTimerThread.Free; end; // changed from DoCancelDiscovery to ThrDoCancelDiscovery FTimerThread := TDiscoverThreadTimer.Create(Self, ThrDoCancelDiscovery, Timeout); ... end; I guess one could also use TThread.Queue instead of the "complicated" window allocation in the timer and that should do the trick too!

Hi! I'm currently migrating our projects from Delphi2010 to Delphi 10.4. One of my biggest changes is actually using StrLCopy and other Ansistring functions that used to be in SysUtils. They are now marked as deprecated and moved to System.AnsiStrings so ... I followed the compiler hint. Now the base problem is that the compiler is confused which function to use so I always prefix the call with System.AnsiStrings.StrLCopy ... which fixed that. But that looks actually very "ugly" and I have the feeling that this is not the way to go idera ment to go. Is there a "best" or anticipated way to handle this? kind regards Mike