AlexBelo
Members-
Content Count
59 -
Joined
-
Last visited
-
Days Won
1
Everything posted by AlexBelo
-
Hi all. Please forgive my ignorance, but I can't figure out how to use the new bug tracker (https://embt.atlassian.net/), for example, get a list of all open bugs. Maybe there is a tutorial somewhere?.. TIA.
-
Thanks, it works ... now. I did the same thing before but didn't get any results. This is weird ... In general, the home page is a perfect example of very poor usability.
-
Hi all. I recompile my old 32-bit codebase to use it in Win64 projects. At moment I need to recompile a component with ide designer part (it lives in separate unit). Question: how to configue my package to avoid compilation of this unit?
-
Hmm... But how I use such a component for more then 15 years?.. OK, most likely because I always compile monolithic executables. 🙂
-
But AFAIUI it was not obligatory. Now it's "must" for more-than-Win32 components.
-
Thank you. Googling "separate packages for runtime and design" returns old but still relevant detailed article: Creating Packages https://delphi.fandom.com/wiki/Creating_Packages
-
Hi all. Win64 platform, small console test project. After some days with 11.1.5 I permanenetly get the following message: Debugger fatal error: debug kernel not responding. Cleaning of project, rebooting IDE and Windows do not help. Any advice? --- Upd: it looks like something is wrong in project itself. I can run and debug another test project. Anyway I still need advice what to search/fix in project ...
-
__fastcall for overlapped virtual destructor of Delphi base class in CLANG compiler
AlexBelo posted a topic in General Help
HI all. RAD 11.1.5 Consider the following code: class dExc : public TObject { public: static int Counter; dExc() { Counter++; printf(" dExc\n"); } /* __fastcall */ ~dExc() { Counter--; printf("~dExc\n"); } }; int dExc::Counter=0; If I compile this in "classic" mode without __fastcall I get: [bcc32 Error] Main.cpp(19): E2113 Virtual function '~dExc::dExc()' conflicts with base class 'TObject' CLANG-32 compiler shows "Success", but program crashes in destructor of ~dExc. Adding __fastcall produces workable exe in both cases. (CLANG-64 version works with and without __fastcall because 64-bit uses its own call convention AFAIK) Is this a bug in CLANG-32 compiler? -
__fastcall for overlapped virtual destructor of Delphi base class in CLANG compiler
AlexBelo replied to AlexBelo's topic in General Help
Thank you for additional information. __fastcall for overlapped virtual destructor of Delphi-based class in CLANG32 compiler https://quality.embarcadero.com/browse/RSP-38783 -
Hi all. I'm c-builder programmer and I need some Delphi help. In old pas file which I'm trying to compile I see (simplified): function SetItemCell(ACol,ARow: longint; List:TList; value: Pointer): pointer; var t: pointer; sublist:TList; begin t:=NIL; ... t:=sublist.items[ARow]; sublist.Items[ARow]:=value; ... SetItemCell:=t; // give back the pointer to the previously stored element to let the caller dispose it end; OK, this code puts new pointer into list and returns old contents. Now usage with pointer to string: procedure TStringAlignGrid.SetHintCell(ACol,ARow: longint; const Value: string); var v: PString; begin v:=NewStr(Value); // <--- new AnsiString instance on heap v:=SetItemCell(ACol,ARow, FHintCell, v); if v<>NIL then DisposeStr(v); // <--- delete old string end; No problem in "old Ansi" Delphi versions but now compiler scolds NewStr and DisposeStr as incompatible with Unicode. So a question is: how to create string on heap?
-
Sources are very old and not mine. All compiles fine in 2007 (last ansi version) but now I'm preparing a codebase for unicode and 64 bits. Thanks all for help. BTW I see nothing too strange in code: function SetItemCell operates with pointers on different types (some structures and strings as special case). Nothing nonlegitimate or obscure.
-
Is this correct New(v); v^:=value; Dispose(v);
-
Umm... procedure TStringAlignGrid.SetHintCell(ACol,ARow: longint; const Value: string); var v: PString; begin // v:=NewStr(Value); // <--- new AnsiString instance on heap v:=New(Value); [DCC Error] E2008 Incompatible types
-
Hi all. I'm new in 64-bit RAD development. At moment I'm trying to find out how to search memory leaks. According to docs FastMM's ReportMemoryLeaksOnShutdown does not work in C++ projects at all. madExcept does not work with 64-bit projects ... Do you know any tool for 64 bits?
-
Thank you for tip. According docs it supports all types of applications (including 64 bit BCB). But at moment I have problems: 64 bit app does not start at all, 32 bit app does not show any results ... I'll try asking EL support.
-
Hi all. I'm investigating memory leak in new builder "quality" release using madExcept's resource leak feature. Everything goes fine in win32 compilation but in win64 I see: [ilink64 Error] Fatal: Unable to open file 'MADEXCEPT.O' No such file in madExcept installation. All paths in IDE to madExcept binaries looks valid. Any advice? Also what can I use to see memory leaks in win64 version?
-
Yes, it was missed libs (*.a). But unfortunately even successfully linked madExcept does not work in my test 64-bit CB exe (nor "mad" exception handling, nor leaks detection) ...
-
Here is info from Mathias (aka madshi) So CB64 is simply out of official support.
-
Hi all. I'm on CB2007 for many years. After last special BCB "quality release" I've eventually decided to try a modern version. void __fastcall TfrmMain::btnHelloClick(TObject *Sender) { ShowMessage(L"Hello, World!"); for(int i=0; i<10000000; ++i, ++t) { try { try { throw Exception(L"Message Message Message Message Message Message Message Message Message Message Message "); } catch(const Exception& e) { throw; } } catch(const Exception& e) { } } ShowMessage(L"Good bye, World!"); } CLANG compiler, static linkage. Memory Initial Final (data from Task Manager) Win64 17680K 18328K Some leak?.. Win32 5080K 1734332K !!!!!!!!!!!!!!!!!! madExcept reports memory leak on every rethrowing in catch (tested without loop, of course) in Win32. Unfortunately I can't test it in details in Win64 version (because it looks like madExcept still doesn't support bcb64) but subtle hint on memory leak is also visible in test results. Before posting QP report I would like to see your comments/results.
-
https://quality.embarcadero.com/browse/RSP-38717?jql=
-
Testing your cases in last version, Win32 build: ACCESS VIOLATION DURING STACK UNWIND No AV but a lot of memory leaks. Not Firing Destructors When It Should Destructors are on place but memory leaks. TRY/CATCH LEAK This very basic case is really fixed. Fire Destructors Twice No double destruction but memory leaks. AFAIUI all memory leaks are unproperly handled exception objects.
-
Hi all. I'm trying "modern" IDE (I still use 2007 version) in "C-Builder mode" (if this matters). I see vertical lines at the left side which marks blocks of code (from '{' to '}' ). Is this possible to hide them?.. (Can't find anything in Tools->Options).
-
> all the options I've done it before posting here - except definitely useless in this context "Colors" clause. 😄
-
> creator normally answers OK, I'll try there.
-
Yes, thank you. (Why they placed it in "Colors" section? I see no reason for this...) > 3rd party tools No, fresh installation without any additions.