Lars Fosdal

procedure TForm1.AnsiUpperTest; var s, o:string; begin s:='default'; AnsiUpperTestConvert(s,o); end; Adding a separate output variable helps. Can it be that the const DefaultValue is cleared by the out reference? Alternatives: Use a var param or a function result.

YMMV, by a LOT. We don't use data-aware controls at all, but do all the DB work behind the scenes and pass custom classes to the UIs. These classes have methods to retrieve themselves from the DB, save themselves, as well as various processing methods, and we use helper/proxy-classes to simplify populating the UI as well as taking input from the UI. The source of the data doesn't matter to the UI, and we found it easier to have absolute control over the UI behavior and data updates and creation. Our UIs are much more than CRUD interfaces, as the data types are very complex and interdependent, as well as manipulated from multiple systems. But - in the end - we do a lot of explicit "wiring" in code in the UI proxies. For other solutions, data-aware components can work perfectly, but TBH, I've only worked with one solution that used them, and they had written an entire ORM to handle their data retrieval and input from DB to UI and back,

Without documentation about the parameters of these functions and their types, you will be unlikely to succeed.

Mine too, especially when I've written the semantically correct word, but the compiler barfs because the actual word in the code is misspelled.

The parser only deals with the parsing of a basic syntax. The interpretations of the validity of variations of switches, flags and arguments still has to be built on top of the base parser. I decided not to write those parts, as there are so many (potentially conflicting) ways to write it. It should be easy to add that layer yourself, though.

Double (or multiple) switch chars work. But - as mentioned: the syntax for multiple arguments to an option is (arg1, arg2, ..., arg n) If you need the separator as part of a string, you quote the string. (art1, "this , is in arg2", 'this has a " in it') Parse(' --strings="String A, StringB,StringC, String D and E"') 1 Switch[--] Option[strings] Flag[=] Values["String A, StringB,StringC, String D and E"] AsString('strings'): String A, StringB,StringC, String D and E

What about breaking down the menu into groups? Yes, it is convenient to have everything on top level ... until it isn't. Since I don't use GExperts, my grouping suggestion is only based on menu names, not on actual functionality - so some placements may be totally wrong. File Code Librarian Code Proofreader ... Favorites Open Project Backup ... Class Browser Dependencies Source Export Component Copy Names Find Reference Grid ... Hide / Show Non.-Visual Rename Replace Select Set FocusControl Set Tab Order ... To Code Editor Bookmarks Comment Empty Code Blocks Clipboard History Experts Focus Go to Grep Next Item Previous Item Results Search ... Macro Library Playback Record Start/Stop Tools Ascii Chart IDE Menu Shortcuts... Keyboard Shortcuts Message Dialog PE Information To Do List -- Configuration About Window Add Dock Window Reselect Desktop

For some reason, many of the programmers I know have various degrees of dyslexia - which leads me to mentally auto-correct the misspellings I find, and in many cases, leave them as is. Usually, changing the spelling leads to something unpleasant happening, such as a mismatch between stream content and object property, DB field names, etc.

Early return benefits: The contract clauses of the method are clearly defined at the top of the method code. If you made it past the contract clauses, your code is safe to run You can easily add or remove contract clauses without breaking the code (or having to add yet another layer of indentation) I used to be a nester, but I evolved.

@Mike Torrettinni I updated the gist at https://gist.github.com/LarsFosdal/f8a3e2a7caa5e541e4c04f10dfdfcfff with an extra Analyze function that writes a clearer breakdown of the parsing. and added your example and a modified version of it to the test app. As far as I can tell, it works as expected? Your example: /strings="String A, StringB,StringC, String D and E"' is a single string param. The test output: Parse(' /strings="String A, StringB,StringC, String D and E"') 1 Switch[/] Option[strings] Flag[=] Values["String A, StringB,StringC, String D and E"] AsString('strings'): String A, StringB,StringC, String D and E ClientIdentifier = My modification of your example: /strings=(String A, StringB,StringC, String D and E) is an array of strings. Note the use of parenthesis to indicate multiple arguments. Parse(' /strings=(String A, StringB,StringC, String D and E)') 1 Switch[/] Option[strings] Flag[=] Values[("String A";StringB;StringC;"String D and E")] AsString('strings'): String A ClientIdentifier =

Every request needs evaluation to understand WHY the request is being raised, and if the request actually is what the customer needs - or if he is just attempting to scratch an itch instead of curing the cause of the itch.

TList<T>.Pack( function (const L, R: T): Boolean begin Result := (L.Condition); end);

Did the refactor solve your "always same I" problem?

Nope. F5 or Ctrl-R and the visual error remains, and I still can't edit the post again.

Is it about the pitfalls of variable captures? The simplest solution to that is to simply divide and conquer. Also - the queuing code you wrote, will only do five scrapings since you never descrement nCount when a task completes, nor do you retry the loop until there are no lines not containing 'done'. procedure TForm1.QueueScraping(I, Q: Integer); Async( procedure begin Memo1.lines.add('I=' + I.ToString() + ' Q=' + Q.ToString()); GetWebContent(I); end ).Await( procedure begin Memo1.lines.add('Done I=' + I.ToString() + ' Q=' + Q.ToString()); end); end; procedure TForm1.BitBtn1Click(Sender: TObject); var I, nCount : Integer; begin I := 0; nCount := 0; for I := 1 to StringGrid.RowCount-1 do begin if StringGrid.Cells[0, I] <> 'done' then begin inc(nCount); if nCount > 5 then Break; QueueScraping(I, nCount) end; end; end;

Did you call BluetoothLE1.DiscoverServices before calling GetService? http://docwiki.embarcadero.com/Libraries/Rio/en/System.Bluetooth.TBluetoothLEDevice.GetService http://docwiki.embarcadero.com/Libraries/Rio/en/System.Bluetooth.TBluetoothLEDevice.DiscoverServices

Pessimist = Experienced optimist 😛

@rvk - My question was due to the comment in the docwiki link that explicitly mentions Firebird and Interbase. Likewise, in http://docwiki.embarcadero.com/Libraries/Rio/en/FireDAC.Comp.Client.TFDCustomConnection.Transaction - it says "Note: At the moment, this behavior applies only for InterBase and Firebird connections."

Note to @Daniel - See the above comment. The quote gets wonky after an edit, and once it is wonky, it can't be edited again.

Is it correct that the custom transactions only apply to Interbase and Firebird? I.e. for f.x. SQL Server, transaction handling would have to be in the query itself?

I prefer individually named methods instead of overloads. Optional parameters can work in some cases, but when things get really complex with a lot of variations, I use variations on the theme that @Der schöne Günther introduced. A record, an object, or an anonymous method.

Are you trying to prevent running the same app twice? https://lonewolfonline.net/prevent-multiple-instances-delphi-application-running/ https://stackoverflow.com/questions/8688078/preventing-multiple-instances-but-also-handle-the-command-line-parameters As mentioned above, FindWindow need window class names or window titles. https://docs.microsoft.com/en-us/windows/win32/api/winuser/nf-winuser-findwindowa Unless you want to enumerate processes to find the main window(s), calling taskkill could be an alternative. https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/taskkill

We are trying to make a habit of assigning events / config in code, and not rely on dfm / unit event handler references. Frames are indeed a great way to plugin/unplug code. Although - I have been told that Frames are not quite as desirable in FMX? Can anyone offer their experience with FMX and dynamic creation of frames?

We partly use the same model as Rollo62 does. Functionality that is shared across apps, is implemented as frames. The amount of work required to create a frame from a regular form, depends on the complexity - but quite often it is more or less just a cut and paste.

It could be BLE version / chipset dependent. That out-of-range issue is a killer, really. Not being able to do a proper recovery there is pretty darn hopeless.