chkaufmann

When my application runs in an RDP session I get the following error: exception class : EOSError exception message : System Error. Code: 8. Not enough memory resources are available to process this command. 00469607 Logo12.exe System.SysUtils RaiseLastOSError 004696d7 Logo12.exe System.SysUtils Win32Check 00547faa Logo12.exe System.Classes THandleStream.SetSize 005461af Logo12.exe System.Classes TStream.SetSize64 010931ff Logo12.exe ZipForge 7047 TZFBaseArchiver.ForceUpdate 0109a2ea Logo12.exe ZipForge 9754 TZFBaseArchiver.EndUpdate 0109cd7e Logo12.exe ZipForge 10780 TZFBaseArchiver.DeleteFiles 0125d084 Logo12.exe BSStreams 939 TBSZipStreamWriter.Close The problem is, the error doesn't happen on all clients, it even happens on some clients one day and the other day it's not reproducable. What I know is, it happens when the application creates and writes to a local file during a backup and it happens when the function SetEndOfFile(FHandle) is called. The error only apears when I try to save to a local file (located on any \\tsclient\.... path). When I search the internet I don't find anything helpfull except about changing the settings for the Windows pagefile. Unfortunately the provider of our RDP plattform / servers cannot reproduce the problem. So I'm stuck somehow, I don't even understand if the error message points to the real problem. And I don't understand if this is a problem of resources on the local client or on the server where the RDP session is hosted. Something else I know is, that some customers get the same error message when they launch the File Explorer in an RDP session and try to copy a big file from their local drive to the server. Therefor I think the problem cannot be in my application itself, it has to do something with the OS configuration on the server or on the client. So any hint to isolate and fix the problem is welcome. Thanks. Christian

I changed my code and the error seems to be gone. Thanks a lot. Now my last question is: Are there other potential problems I may have with my application when running in an RDP session? Especially other problems when loading / saving files from local drives? Until now I didn't find a good resource in the internet about where I should be carefull in my application in order to make it run without errors in RDP sessions: Regards Christian

I think I can change the code and create the .ZIP as temporary file and then copy/move this file to the local file on \\tsclient\..... - Normally I use CopyFile() or MoveFile(). What works better with the RDP protocol? - When I look at Windows.pas unit there are many variants of these functions. Should I use a different one? Thanks for your help. Regards Christian

Hi, my provider uses https://rspamd.com/ to check emails before sending out. Now my emails get classified by some modules and I would like to solve this: One is https://rspamd.com/doc/modules/chartable.html Normally I just add text with FMessage.Body.Add() and then I set this: FMessage.ContentType := 'text/plain'; FMessage.CharSet := 'UTF-8'; Is Indy not doing the conversion correctly? The second one is https://www.rspamd.com/doc/modules/mime_types.html this one. The others don't give a big score: R_MIXED_CHARSET 2.89 Mixed characters in a message MISSING_MIME_VERSION 2.00 MIME-Version header is missing in MIME message R_BAD_CTE_7BIT 1.05 Detects bad content-transfer-encoding for text parts 7bit;utf8; Can somebody give me a hint on how to fix this? Christian

I use a cache for sql query results. For this I create a key with THashBobJenkins for the sql string and then I have dictionary with the results with these hashes as key. Unfortunately I run into a problem that two different sql strings give the same hashkey. So I probably have to use the whole sql string as key. Or can somebody provide me a good solution to reduce the sql string to a "smaller" key without loosing uniqueness? Christian

I use the full sql string as key now. Maybe I'm just too old - sometimes still thinking that having several thousand strings as keys may affect performance... Christian

When I use TStringList.SaveToFile() on a hidden text file, I get a create error. Even if I do this as administrator. So the only solution I see is to read the attributes with GetFileAttributes, then save the file and then use SetFileAttributes to hide the file again. Or do I miss something? Christian

When a user starts my application, I would like to know all groups assigned to that user. On the command line I can do this: dsquery user -samid "christian.kaufmann" | dsget user -memberof -expand Now I have a delphi function I could use to run this call and then parse the input. But before I start to implement that I would like to know: - Is a regular user allowed to call this in order to read that information? - Is there another way to get this info from the system? - Or are only administrators allowed to read this information? The background is, I would like to show the user a list of available projects based on the groups he is a member of. Christian

My application runs on a server as remote desktop application. Now I want a function to add files from a local the drive. The RDP file is configured to map local drives and I can see/select these in the file open dialog. Now I would like to limit this dialog to the clients local drives only. - Is there a global option for that? - Is it enough to set a OnFolderChanging event and check the path to be a local drive? - Are there other things todo? Since the file open dialog is almost a complete Explorer today, I would like to limit functionality to open/load a file. Regards Christian

My problem is, that user managment is done by an external company with quite expensive people. Not my choice, and I just try to make things user friendly with the part I do. One final question: The shared drives as UNC path look like \\tsclient\c\..... Is this "\\tsclient" a default by Microsoft in a terminal session or can this prefix be customized by the administrator? Christian

Yes I know. In the end it's more to help the user do use only his local drives to save/load files. Christian

My application has to connect to another application. It waits for messages in XML format, delimited with SOH and EOT. I wrote a listener thread. So far it works, but I wonder, if this is the way to do it. Here is my code: - Is it correct to call ReadLn and then the thread is blocked until any data is sent by the server? - I cannot test the error handling with SendData because I don't succeed to deactivate the server in the moment when the data is sent. Will it work that way? Or should I check another property of my TIdTcpClient to see if the server was deactivated? - Do I need more cleanup if I destroy my TTsClient object? Christian type TTsClientThread = class(TThread) strict private FClient : TIdTCPClientCustom; FMessages : IBSQueue<IBSXmlElement>; FPartMsg : String; strict protected procedure Execute; override; public constructor Create(AClient: TIdTCPClientCustom; const AMessages: IBSQueue<IBSXmlElement>); end; TTsClient = class(TIdTCPClientCustom) strict private FListener : TTsClientThread; FMessages : IBSQueue<IBSXmlElement>; public constructor Create(const AIpAddress: String; APort: Integer); destructor Destroy; override; function IsAlive: Boolean; function Pop(var AMessage: IBSXmlElement): Boolean; procedure SendData(const AMessage: IBSXmlElement); end; { TTsClientThread } constructor TTsClientThread.Create(AClient: TIdTCPClientCustom; const AMessages: IBSQueue<IBSXmlElement>); begin FClient := AClient; FMessages := AMessages; inherited Create; end; procedure TTsClientThread.Execute; var s : String; begin while not Terminated do begin try s := FClient.Socket.ReadLn(Char.EOT, IndyTextEncoding_UTF8).Substring(1); if not s.IsEmpty then FMessages.Append(NewBSXmlFromString(s)); except on E: EIdSocketError do begin Terminate; Exit; end; end; end; end; { TTsClient } constructor TTsClient.Create(const AIpAddress: String; APort: Integer); begin inherited Create(nil); FMessages := TBSGenerics.GenQueue<IBSXmlElement>; Host := AIpAddress; Port := APort; Connect; FListener := TTsClientThread.Create(Self, FMessages); end; destructor TTsClient.Destroy; begin FListener.Free; inherited; end; function TTsClient.IsAlive: Boolean; begin Result := not FListener.Terminated; end; function TTsClient.Pop(var AMessage: IBSXmlElement): Boolean; begin Result := FMessages.Pop(AMessage); end; procedure TTsClient.SendData(const AMessage: IBSXmlElement); begin try Socket.Write(IndyTextEncoding_UTF8.GetBytes(Char.SOH + AMessage.AsXmlDataString([]) + Char.EOT)); except on E: EIdSocketError do FListener.Terminate; end; end;

Thanks for the reply. I have one final question. First I tried with this call: var tmp : TIdBytes; FClient.IOHandler.ReadBytes(tmp, -1, False); FPartMsg := FPartMsg + IndyTextEncoding_UTF8.GetString(tmp); .... Here I noticed, that if ReadBytes had nothing to read, tmp was not reset to length zero, but once some bytes could be read, tmp was set to the number of bytes. What is the reason for that? Christian

For many enumeration type I have I define the same set of methods. Here is an example: TBSSwGender = (sgeNone, sgeMen, sgeWomen, sgeMixed); TBSSwGenderHelper = record helper for TBSSwGender public function AsInteger: Integer; function Code: String; function Name: String; end; function TBSSwGenderHelper.AsInteger: Integer; begin Result := Ord(Self); end; function TBSSwGenderHelper.Code: String; begin Result := '#Enumeration.BSSwGenderCode%d'.Fmt([AsInteger]).Translate; end; function TBSSwGenderHelper.Name: String; begin Result := '#Enumeration.BSSwGenderName%d'.Fmt([AsInteger]).Translate; end; Now my question is, instead of writing the same code for all types. So something like this: TBSEnumHelper<T> = record helper for T function AsInteger: Integer; ... end; TBSEnumHelper<TBSSwGender> = record helper for TBSSwGender Christian

This is what I did. The builder class itself has no RTTI. But I created a generic record. IBSLookupElements is a simple collection of (Key, Caption, ParentKey) elements. Fmt() and Translate() are string helper methods for Format and language specific translations. Then TBSCatalogBuilderSimple is a general subclass of my TBSCatalogBuilder class. Here is my record definition: type TBSCatalogBuilder<T> = record strict private FCatalogId : Integer; FDefault : Integer; FName : String; FTextIdent : String; FTypeData : PTypeData; public constructor Create(ACatalogId: Integer; const AName: String); function Gen: TBSCatalogBuilder; function SetDefault(const AValue: T): TBSCatalogBuilder<T>; function SetTextIdent(const AValue: String): TBSCatalogBuilder<T>; end; { TBSCatalogBuilder<T> } constructor TBSCatalogBuilder<T>.Create(ACatalogId: Integer; const AName: String); var i : PTypeInfo; begin i := TypeInfo(T); Assert(i.Kind = tkEnumeration, 'Passed type is not an enumeration.'); FTypeData := GetTypeData(i); FCatalogId := ACatalogId; FDefault := 0; FName := AName; end; function TBSCatalogBuilder<T>.Gen: TBSCatalogBuilder; var tmp : IBSLookupElements; begin tmp := NewBSLookupElements; for var ix := (FTypeData.MinValue + 1) to FTypeData.MaxValue do tmp.AddOrSet(FTextIdent.Fmt([ix]).Translate, ix); Result := TBSCatalogBuilderSimple.Create(FCatalogId, FName, tmp, FDefault); end; function TBSCatalogBuilder<T>.SetDefault(const AValue: T): TBSCatalogBuilder<T>; begin FDefault := TValue.From<T>(AValue).AsOrdinal; Result := Self; end; function TBSCatalogBuilder<T>.SetTextIdent(const AValue: String): TBSCatalogBuilder<T>; begin FTextIdent := AValue + '%d'; Result := Self; end; Then the usage of the helper looks like this: type TElmeSchlafmodus = (esmNone, esmKeinSchlafmodus, esmTeilzeit, esmVollzeit); ADef.RegisterCatalog(TBSCatalogBuilder<TElmeSchlafmodus>.Create(CAT_ELME_BETRIEB_SCHLAFMODUS, 'Schlafmodus') .SetTextIdent('#ElmeBetrieb.Schlafmodus') .SetDefault(esmVollzeit).Gen); Thanks for all help. Christian

Ok, here is the long story. In my system I define catalogs that can be used as lookup for values. There are complex catalogs with dynamic content that can be edited by the users. And there are simple catalogs that are hard coded in my application. For these hard coded catalogs I have builder class: TBSCatalogBuilder = class abstract strict protected function AddItem(AItemId: Integer; const AName: String; AIsDefault: Boolean = False; const AParentItem: IBSCatalogItem = nil): IBSCatalogItem; procedure DoBuildItems; virtual; abstract; public constructor Create(ACatalogId: Integer; const AName: String; AOptions: TBSCatalogOptions = []); end; So to define a catalog I create a subclass of TBSCatalogBuilder, overwrite the DoBuildItems method and call AddItem several times to define the catalog. Now some hard coded catalogs just represent the items of an enumaration type and instead of creating a subclass I would like one class with a constructor that takes the type of the enumeration and a second parameter with the default value. Right now the definition looks like this: TBSCatalogBuilderSetType = class sealed(TBSCatalogBuilder) strict protected procedure DoBuildItems; override; final; public constructor Create(ACatalogId: Integer; const AName: String; ASetType: PTypeInfo; ADefault: Integer = 0); end; and then I can use it like this: type TBSColumnType = (ctUnknown, ctBoolean, ctBytes, ctDateTime, ctFloat, ctInteger, ctMemo, ctString, ctGuid); begin myColumnTypeCatalog = TBSCatalogBuilderSetType.Create(1, 'Column Types', TypeInfo(TBSColumnType), Ord(ctString)); end; This works and it does what I want, but the TypeInfo() and the Ord() just don't look nice and I was wondering if there is not a way to avoid these. And I didn't want to define a generic class because like this the code is duplicated for every catalog. But maybe this is the only way to do it: TBSCatalogBuilderSet<T> = class(TBSCatalogBuilder) public constructor Create(ACatalogId: Integer; const AName: String; ADefault: T); end; Christian

It's just a question of readability. My version works fine, but now I would like to avoid these TypeInfo() and Ord() when I call Foo(). And if I can validate the second parameter to be a valid member of the first (tkEnumeration) Parameter, this would be a nice additional check. Christian

Yes "generic" is maybe the wrong word because I don't mean a generic method. Maybe I should have said untyped or typeless. I would just like to call it like this: Foo(TSet1, a13); Foo(TSet2, a21); Christian

I have two applications that communicate in a LAN using UDP. Both applications have TIdUDPClient and a TIdUDPServer. Until now the user must configure the port and the IP address. Now I would like a solution where client computers "find each other" automatically. What is the best way to do that? Regards Christian

In my Windows applications I use VirtualTreeview for grids with inplace editors. Sometimes I load up to several 10'000 rows (nodes). For this I use rather light node classes. When I look at TTreeviewItem in FMX, then each node is a control and I don't think, this will work the same way. So are there alternatives. I need - tree check functionality (simple, tristate) - multiple columns - inplace editors - lazy load for child nodes Christian

Hi, I have an application that still uses an MS Access as database. With the library I have I can run it on other databases as well since all requests are pure SQL (no TTable or things like that). Now I'm looking for a replacement and I wonder, if there is a solution that could be used for cloud and offline. The databases are not that big (about 25 tables, maximum 25'000 records each). But each customer may have up to 50-100 databases. Ideal would be a solution where the database could be in the cloud, but there is an offline copy in case the internet connection is broken, so I can continue to work. When I'm back online, all changes would be synchronized to the server. Now I know there is probably no easy solutions, but I'm open for ideas on how to find a good solution. Christian

Hi, I just fixed an error in the following function. I had to add the last line where I assign tmp to Result. Now I wonder, why I didn't get a compiler warning like "return value of function might be undefined.". I get such errors in the same project with other methods, so the options are set correctly. Christian class function TLnxResult.GenForEvent(AData: TLnxObject; AEventId, AAgegroupId: Integer): IBSEnumerable<IBSSwLenexResult>; var data : TLnxAbstractObject; clubs : TLnxCollection; tmp : IBSList<IBSSwLenexResult>; begin tmp := TBSGenerics.GenList<IBSSwLenexResult>; data := AData; while Assigned(data) and (data.Tag <> lxCOMPETITION) do data := data.Parent; clubs := TLnxCollection(TLnxObject(data).ChildObject[lxCLUBS]); if Assigned(clubs) then for var ixC := 0 to clubs.ChildCount - 1 do begin for var e in TLnxEntryResult.GenForClub(lxRESULTS, TLnxCollection(clubs.Objects[ixC].ChildObject[lxPERSONS]), AEventId, AAgegroupId) do tmp.Add(e as IBSSwLenexResult); for var e in TLnxEntryResult.GenForClub(lxRESULTS, TLnxCollection(clubs.Objects[ixC].ChildObject[lxRELAYS]), AEventId, AAgegroupId) do tmp.Add(e as IBSSwLenexResult); end; tmp.Sort(function(const AResult1, AResult2: IBSSwLenexResult): Integer begin Result := AResult1.Place(nil) - AResult2.Place(nil); if Result.IsZero and (AResult1.Place(nil) = BSSW_MAX_PLACE) then Result := AResult1.Status.AsInteger - AResult2.Status.AsInteger; end); Result := tmp; // <<<< this line was missing. end;

So just that I get it right: Because my return value is an interface type, the compiler adds something like this as first statement of the method: Result := nil; Chrisitan

In 10.3 (Rio), class constructors are implemented the same way as initialization sections of units. This means, I have no control about the order. So if I create an object of my class in an initialization section, I cannot be sure, that the class constructor for this class was already called. Just wondering: is this problem fixed in 10.4? Or is it still done the same way? Christian

The problem is to create a reproducable case. The order of initialization sections and class constructor is at least partial random. So things work fine but if you add a unit or change one of the uses in a project, the order is different. So I don't use class constructor anymore.