I'm using RAD Studio 10.2.3 Tokyo.
Research
Related to this question, and this issue I run into trouble when debugging my flagship application that extensively utilizes runtime BPLs as a plugin system.
The problem occurs only when compiled and debugged as a Win64 application, and using Delphi's Win64 internal debugger. Delphi's Win32 internal debugger does not have this problem.
When using "load all symbols" in the debugger options, I miss stack and local symbols info for some units but not other units in the same BPL (check screenshots).
However, when I use the option to load debug info for specific BPLs, and only in my runtime BPL all units have local symbols info. Logically, I am missing debug info for all other units.
Of course, I dont like having to specify each symbol table specifically, I would rather load all symbol tables and not have to mess with it.
I have also experimented with using the symbol tables path, but this made no difference in loading the debugging results.
Question
Has anyone else experienced this, and found a better workaround than specifying which symbol tables to load?
Maybe something like an INI file, or Registry setting, or other solution for the Win64 debugger?
Some background info and screenshots below
Debugging view when turning on "Load all symbols"
Debugging view when using specific symbol table for this runtime BPL
Settings for specific symbol table loading
Related
RAD Studio 10.4.2 C++Builder running on Windows 10 Pro 64-bit PC. Target VCL Win64 (ie Clang64 compiler).
I have a large C++ project that is under development using VCL Clang64. I am compiling using static linking of all RTL libraries and all installed component packages.
It was compiling and running OK before. I've added some new stuff, and now it compiles and links without reporting any errors, but when I run it I get an error box and the application doesn't run (doesn't even start up). The error box says.
Error reading OKBitBtn.Kind: Resource BBOK not found
I have used Notepad++ "find in files" to search all directories for a file containing BBOK, but it says "none found".
If I compile with the "Use runtime packages" project option enabled, I get an AV when I try to run the EXE file.
If I compile and link using Clang32 (target = Win32) with static linking, it runs ok - no sign of any errors (but I need a 64-bit application to interface to my PostgreSQL database using FireDAC, so this experiment won't talk to the database, as expected).
I am at a loss as to what can cause this error, and what debugging steps I should take to track it down and solve it.
(thanks to Remy for making edit improvements to my mquestion)
Thanks to help from David Millington and Miguel Moreno (provided by skype conversation I now have the solution:
I have some of my own VCL components that are written in C++ (and work with the classic compiler). I had these de-selected in the "installed packages" list - but they were still listed (with empty check boxes). I know that RAD studio 10.4.2 (and earlier) does not work well with VCL components written in C++ and compiled with clang32 or clang64 compilers.
If I "remove" these components from the list then the problem is solved. At this point I have no idea why this happens.
This is a bit of a pain as it means swapping from projects that use the components with classic compiler (many) to the big project I am spending a lot of time on is not convenient. However it is an acceptable workaround. I am hoping that this issue will be fixed in the coming RAD Studio 11 version.
From time to time, when I load a project into the XE6 IDE, the following error occurs
This error results in the TZConnection component being removed from the Datamodule for some inexplicable reason. Note that the project has been loading without issue for ages and just out of the blue, this occurs.
Has anyone seen this before and know of a way of stopping it from occuring ?
It also occurs with other components, not always the TZConnection component but mostly ZConnection.
Like I said before, this appears randomly. I could close project A, open project B then close it and return to Project A and bang!, the error occurs.
Any clues ? (note that this also used to occur in Delphi 2007)
The Zeos libraries are themselves causing this problem.
To see why, and to fix it, use Delphi to launch a second instance of Delphi, and debug the issue directly.
I wrote a blog post showing exact steps here.
The key is to set the executable that will be run for your zeos package, be sure to build them in debug configuration, and then click the Run button on the Delphi IDE toolbar. A second delphi instance will start. Open the affected form but be sure to be using the IDE instance that is Being debugged as opposed to the one which is currently doing the debugging, when the exception occurs that is causing your component to delete itself, you will be able to step into the package code and see the problem.
I suspect a DLL-hell path issue. (Multiple copies of Zeos or other core BPL/DLLs in your path.)
Actually, it doesn't sound that inexplicable - it's probably caused by an exception occurring as the DataModule (or some form with db-aware components connected to it via properties) is being loaded into the IDE (see below). Have you tried checking that wherever your ZEOS .BPL files are located is on your system Path? Likewise any .BPLs they depend upon - see the "requires" clause in the .DPK file(s) for Zeos.
This sort of problem arises fairly frequently with flaky DB components, maybe more so than other types of component because db components more frequently involve linkages between datamodules and forms, e.g. when db-aware components on the forms are connected to others on the datamodule.
So, sometimes, whether this sort of problem shows up or not depends on the order in which the IDE will re-open them - try closing the project with only the dm open and then re-opening it. A bit of experimenting with which datamodules and forms are open in the IDE and in which order may help you pin down the problem. If/when you have a reproducible sequence of steps to provoke the problem, report it to the authors.
A fairly reliable way to determine whether the problem is being caused by an exception as a project loads is to run one instance of the IDE inside another. As long as the first ("outer") instance of the IDE has the debugger set to "Break on Language exceptions" it should be able to take you straight to the source of the exception (assuming it occurs, of course) when the project is loaded by the second instance. It may take a few goes to "catch it in the act" of course, but it is hugely satisfying when you manage to. Good luck!
Unlike MartynA I doubt this would be caused by an exception.
I would more likely expect such issues to be caused by windows path environment variable being too long.
Unfortunately still many component vendors and even some programs modify "windows path environment variable` to make their own files accessible by other programs.
And when windows path environment variable fails to provide sufficient information windows will try to find the files in default system directory which is C:\Windows\System32
So I would strongly advise checking the windows path environment variable to check its length.
The easiest way to do this is by simply starting the command prompt and typing in path directive or perhaps path >> D:\path.log to export the path environment variable information into a file for easier reading in case if it is long.
EDIT: BTW I just checked my path environment variable and I see that I will have to do some cleaning because it contains entries for both Delphi XE8 and Delphi XE 10 Seattle file locations even thou I have removed Delphi XE8 from my computer. Not to mention some entries from some programs that I have removed quite some time ago.
I'm using the Synapse library in a Delphi project to do some networking.
When I try and use one of the Types that are defined in the external units, i.e: 'TTCPBlockSocket', it has the red underline and says "Undeclared Identifier 'TTCPBlockSocket'".
The files are all in the correct folders and the 'uses' statement can find them and shows no errors.
The strange thing is that I've had an identical setup in another project and there have been no errors in that project.
As a note: I can't install/edit the software/settings, so I can't add any fixes.
This is a well-known bug in Error Insight. It has existed since Error Insight was first introduced in Delphi 2005 or 2006, and still exists today in Delphi XE5.
It's caused by a different compiler being used for Error Insight, apparently, that doesn't have access to the same symbol set used by the Code Insight compiler (the one that helps you find symbols in the Code Editor) or the command-line compiler (the one used to actually compile your code for an application or package). It therefore only uses files that are actually referenced in the project (.dpr) file to locate symbols, and since the majority of files in the uses clause in your code aren't in the .dpr, it can't find them.
There are two fixes (one that is very easy, and one that works but is a pain in the backside):
(The easy one). Turn off Error Insight totally, in Tools->Options, the Editor Options section, Code Insight; just uncheck the box for Error Insight. I prefer this one because Error Insight doesn't work properly anyway, and it avoids the annoyance of having to use the other option every 10 minutes. This is the first thing I do when I install a new version of the IDE and see the red underlines.
(The pain way). Use Project->Add to project for every unit that contains one of the underlined symbols. This adds a reference to every unit to the project .dpr file, which causes it to increase in size drastically. It's a pain because you typically have to do that for every single unit (in my experience, including those that are part of the standard VCL/RTL) that hasn't already been added, and it very quickly becomes irritating. Error Insight doesn't tell you anything that a quick Ctrl+F9 won't anyway, IMO.
I have upgraded a project from RAD Studio 2010 to RAD Studio XE4. The project is mostly C++ using the C++ Builder half of RAD Studio, with smatterings of Delphi.
Under 2010, the Release build was 22MB. Under XE4, the same Release build is 55MB.
This is a problem because:
55MB is remarkably large for an EXE
It takes noticeably longer to start the program (not much, but some.)
Many of our customers download new versions from remote areas or on board ship while at sea. Size matters.
What could cause this, and how do I fix it?
Notes
Oddly enough, the debug build is still only 23MB. Project settings seem very similar. I diffed an export of the option sets and saw only <UsePackages>True</UsePackages> as something present in Release not Debug; removing it made no difference. Beyond that there's code optimisation turned on and lack of a _DEBUG define.
Debug info is generated for the Release build, but goes into an external .tds file. (This will be used for EurekaLog in future, and currently allows us to debug release builds of the program.) I wondered if the linker was linking in debug info, but as far as I know the C++ linker always places the debug info in an external file, which exists, plus EurekaLog is not yet enabled in the build.
The project files (.cbproj) were created in XE4, not upgraded from old 2010 ones. I made new projects and added the old .cpp and .pas files. This was to avoid issues caused by upgrading - on the Embarcadero forums it's often recommended to create projects anew when upgrading IDE versions. Until this occurred, I had indeed encountered fewer issues than in 2010.
There are many mentions of Delphi XE2+ EXEs being noticeably larger than those produced by old versions of Delphi. (1, 2, 3, 4, 5, 6.) While the C++ linker is different to that used by Delphi, it is possible the causes are similar.
The cause there appears to mostly be RTTI used by the Delphi RTL code, which in a Delphi project can be removed by specifying {$WEAKLINKRTTI ON}. The XE4 documentation does not mention an equivalent C++ linker pragma. There is #pragma explicit_rtti (the C++ analog of {$RTTI}) and __declspec(delphirtti) (the C++ analog of {$M}/{$TYPEINFO}).
The project is linked using runtime packages and the dynamic RTL. It is a 32-bit VCL forms application. I have XE4 Update 1 installed on Windows 7.
Edit: David Heffernan asks in a comment below about the .map file sizes of each project.
The release .map size is 17MB, and .tds size is 80MB.
Debug has now stopped linking, with the linker reporting it is out of memory (an old, old bug that should have been fixed years ago.) There are no .map or .tds files generated and so I can't give a comparison size accurately. From memory the .tds was about 100MB, and unfortunately I don't recall the .map size. If I get the project linking I will update the question.
Edit again: I found the answer (sortof) - turning 'Expand inline functions' OFF reduces the EXE size to the expected value. This is puzzling itself though - see my answer below. Note that the debug build already had this turned off and still won't link, giving an out of memory error.
I haven't marked my answer as correct yet, in case someone can provide some insight into exactly what is going on, which would be a much better answer than just 'turn this option off'.
Solved (sortof.)
The Release build had "Expand inline functions" turned on. Turning this off reduced the EXE size from 55MB to 17MB.
This is an astounding difference and I don't know why the difference is so big. I find it unlikely we have 38MB worth of inlined functions, even counting templates and headers, let alone ones using the inline keyword. If there is a way to examine the linker steps, output, or obj files to see what is being expanded, please comment - I'd find this really valuable.
We are trying to split up our monolithic EXE into a combination of an EXE and several packages. So far, we have one package that we're trying to use, and when running the EXE Codeguard shows the following error on startup:
CG Error
Two different CRTLDLLs are loaded. CG might report false errors
(C:\Windows\system32\CC32100MT.DLL)
(D:\Projects\Foo\Bar.bpl)
OK
I read this as two different runtime libraries being loaded - one, the correct one (CC32100MT.dll), one incorrect, which is the package we're trying to use.
Continuing to run the program shows odd errors, especially casting between classes or passing a pointer to a class as a parameter in a method that crosses the EXE/DLL boundary. Codeguard itself doesn't show any other errors at all though. Edit: This is now resolved, and wasn't related. The program appears to run correctly, but the warning Codeguard shows is still worrying.
How do we solve this?
Some more details
We've looked at as many things as we (the developer working on this and I) can collectively think of:
Each project is built using runtime packages. The EXE host lists Bar in its package list.
Each project is set to compile with dynamic RTL. However, changing this does not solve the problem.
The package is linked to the EXE via its BPI file, but linking via a LIB makes no difference either.
The EXE and BPL are compiled with the same project settings, where the same options exist for both types of project. We think, anyway :)
There is only one copy of the BPL and BPI on the system: it's definitely linking to the right one.
Examining the EXE and BPL with Depends and TDump show they are both using C:\Windows\system32\CC32100MT.DLL. They should both be using the one RTL.
Creating a new project (a plain VCL forms application) and linking to the BPL (via its BPI) works fine. Something in the process of adding all the files and LIBs that make our EXE contain the code it needs to changes this, but we haven't been able to figure out what.
The LIBs all either correspond to DLLs we use (flat C interface, usually look as though they were built with MSVC) or are simple projects with lots of related files, compiled to a lib for the purpose of linking into the EXE - these correspond roughly to the areas of the program we want to split to BPLs, by the way. There don't seem to be project options for the LIB projects that would affect RTL linking, unless we've missed them.
I have exhaustively hunted through Depends and looked at all RTL and CC32*.dll files the EXE and every single DLL references. All are identical: rtl140.bpl and CC32100MT.DLL. Fully qualified paths show they are the same files, too. Everything should be using the one same run-time library.
Edit: The final EXE is complex, built with several libs, several DLLs, etc. All these, when built with C++Builder, are built with the current version. Is it possible there's something in one of these DLLs or LIBs that could cause a problem? I don't know enough about how the RTL is linked in to be sure about where to look... my (naive?) assumption is that the linker would normally link in one set of RTL functions, but that of course doesn't seem to be happening... and I don't know how things change when using packages. Is it possible this error has always existed and Codeguard has not flagged it before, because we haven't used something dynamic like a package?
Perhaps another question is, Why would a package have its own RTL anyway, or what would make it count as 'a RTL DLL' to Codeguard?
We're stumped. Absolutely stumped. We've had other problems using BPLs (they seem to be surprisingly tricky things, especially using C++) but have managed to solve them all. This one we've had no luck at all and we'd really appreciate any insights :)
We're using C++Builder 2010 (as part of RAD Studio actually, but with little Delphi code apart from components.)
Edit: Started a bounty. I'd really like to solve this!
Edit 2: Thanks to David Dean for his help (marked as answered below.) Via email, he pointed out this issue was reproduced in a simple test case by someone else, and is logged in Embarcadero QC as report 86335. Currently there is no fix, but the warning does not appear to indicate a genuine problem (ie, it's probably a spurious error, and while it's a pity to have to click past the dialog when you run, hopefully there's nothing in the error to worry about.)
Since one of these is coming from a .bpl, did you try turning off "Build with runtime packages" in the project options?
We had a similar problem. We tracked it down to a (non VCL) .cbproj that was created without the "Multithreaded" option.
As far as I can tell, the only time you get chance to set this option is when you create a new .cbproj, it cannot be changed afterwards using the GUI. We ended up "hacking" the .cbproj to include the following:
<Multithreaded>true</Multithreaded>
To determine which dll is causing the issue, it should be the last dll loaded in the output window just before you see the CG message.
Did you check if you use _TCHAR as char. We had some similar problems with RAD Studio and we found a workaround using _TCHAR as char. As soon as one DLL or BPL Project is compiled with wchar_t, this code guard error appears.
We also figured out, that EXE projects can be compiled with TCHAR = wchar_t without any problem (the main function will be WIDE).
The settings does not affect the GUI being able to handle UNICODE.
A customer logged a similar case in our public bug tracking system and the bug has been identified and fixed in the latest release.