I'm writing a dynamically linked library that includes both C and Haskell code. I'm using the GHC. It's working fine except that when there is a stack overflow in Haskell, the program crashes. It looks like I should be able to catch these errors by adding a hook when I set up the RTS. (https://downloads.haskell.org/ghc/8.10.5/docs/html/users_guide/runtime_control.html#hooks-to-change-rts-behaviour)
However, the hook isn't run. The documentation says that it won't work if the library is dynamically linked in Windows, but I'm running in Linux. Am I missing something or is the documentation incomplete?
The hook runs in a stand-alone executable.
A little more detail: I'm writing a plug-in for PostgreSQL to allow a user to run haskell code. I'm using Hint as the interpreter. Since the user can input possibly bad code, a stack overflow is an issue that must be handled gracefully.
Related
Closed. This question needs debugging details. It is not currently accepting answers.
Edit the question to include desired behavior, a specific problem or error, and the shortest code necessary to reproduce the problem. This will help others answer the question.
Closed yesterday.
Improve this question
I'm trying to add CUDA functionality to an existing code. The desired result is that if the user has cuda runtime installed on their machine, the code will use their cuda runtime (using dlopen) to check if a CUDA enabled GPU is available and then run the CUDA code on it if that's true. Otherwise, run the original non-GPU accelerated code. However, there are some gaps in my understanding of libraries and CUDA that make this tricky for me.
The code compiles just fine if I specify the location of the required CUDA libraries (cudart and cublas) and dynamically link them. However, I tried not linking these libraries and instead wrapping 'everything' I need using dlopen and dlsym to get handles to the functions I need. However, compilation fails when it gets to actual device code (definitions for angle bracket code) because it's looking for things like __cudaRegisterFunction during compile time. I've replaced the angle bracket calls with a wrapped version of cudaLaunchKernel but still get this issue, possibly because the definitions of the machine code themselves require some special calls.
Some fundamental things I'm unsure about are when the symbols in a shared lib have to be resolved. For example, let's say the user does not have cudart.so, is it possible for me to just not run any cudart/cuda code and avoid any runtime issues involving finding references to functions contained in this library? Or do all cudart.so functions need to be found in the .so file regardless of whether or not they're used? If the answer to this question is that only functions that are used need to be resolved, would this not obviate the need for wrapping functions via dlopen/dlsym? Another question somewhat related to this is: can you compile cuda code without linking to cudart? I may be confusing two separate issues in that it might be necessary to link to cudart.so when compiling CUDA code but that does not mean you are actually using cudart.so during runtime.
It's entirely possible I'm going about this the entirely wrong way so hopefully the general statement of what I'm trying to do can get me to some working answer.
I took the habit of managing my post-install setup things in the Setup.hs. The advantages are clear: it's coded in haskell, and it has access to the app data directory, where the binary will get installed.
I'm using it now for instance, to copy the QML data files and icons that a haskell QML application needs to the installation folder.
This all worked well until recently, but now stack started optimizing the build a lot, and if it detects that no haskell files changed, it won't re-trigger the Setup.hs. That obviously doesn't work for me, because I'd like my QML files and icons to get copied every time I run stack install during development.
I can't find a flag for stack that would trigger just the Setup.hs and nothing else (or little else -- certainly clean and --reconfigure force rebuilding too many things). How are people doing that? Should I write a shell script? But it'll have trouble finding the target directory for the installation, plus it's less portable that the haskell in my Setup.hs?
I feel stack are probably right in their reasoning and if my current way was really correct they would have enabled this scenario by now. So either there's a stack flag that I don't see, either I'm doing it wrong I think.
NOTE: stack is so smart that not even touch Setup.hs or of another .hs file helps to re-trigger its execution. I must physically change the contents of one file.
NOTE2: Ah, I re-discovered stack path to find out target path information. I could use that from a shell script to install the files. Still feels like a shame to write that in shell when I have haskell...
NOTE3: posted the question on the stack mailing list now...
NOTE4: actually opened a stack bug now
This question might be too detailed for this forum, but I could not find a mailing list for duktape. Maybe this question will be useful for others trying to get duktape running on more obscure hardware.
I am trying to get duktape to work on an old ColdFire CPU, using an OLD gcc compiler (2.95.3). The board has limited resources (flash/RAM) but I seem to have enough of both. I must live with the old compiler.
I believe the duk_config.h is calculating the right options regarding endianness, etc. I am using a number of the duktape options to reduce code and data size. I have successfully used the same configuration on 64 and 32 bit Ubuntu and it works fine.
The "properties string" that is formed and set in duk_hthread_create_builtin_objects() is:
"bb u pnRHSBOL p2 a8 generic linux gcc" which seems correct (not sure of the effect of the "generic" tag for architecture).
I am getting a failure when calling duk_create_heap(). I have isolated the problem to a what I believe is a JS compile error related to duk_initjs. If I undef DUK_USE_BUILTIN_INITJS, initialization works. The error is a syntax error (not sure where yet). By running "strings" on my executable, I can see that the javascript program source string is there. As a side issue, when this error occurs, the longjmp doesn't work (setjmp never called?) so my fatal handler gets called, but I don't care about for now.
I thought it might be my small C stack (as it appears the js compiler uses recursion) but making the stack much larger didn't help.
I am starting to dig into the JS compiler, but this must be an issue with the architecture or my environment. Any suggestions appreciated!
EDIT: I just now noticed a post of a similar issue, and there was a request to repeat with "-DDUK_OPT_DEBUG -DDUK_OPT_DPRINT -DDUK_OPT_ASSERTIONS -DDUK_OPT_SELF_TESTS" I will try to use these options (if possible, I am very close to a relocation limit on my executable).
There was a bug in 1.4.0 release (https://github.com/svaarala/duktape/pull/550) which caused duk_config.h to incorrectly end up with an unpacked value representation even when the architecture supported packed representation. This might be an issue in your case - try adding and explicit -DDUK_OPT_PACKED_TVAL (which forces Duktape to use packed representation) to see if it helps.
I'm working on a fairly simple text-editor for Haskell, and I'd like to be able to highlight static errors in code when the user hits "check."
Is there a way to use the GHC-API to do a "dry-run" of compiling a haskell file without actually compiling it? I'd like to be able to take a string and do all the checks of normal compilation, but without the output. The GHC-API would be ideal because then I wouldn't have to parse command-line output from GHC to highlight errors and such.
In addition, is it possible to do this check on a string, instead of on a file? (If not, I can just write it to a temp file, which isn't terribly efficient, but would work).
If this is possible, could you provide or point me to an example how how to do this?
This question ask the same thing, but it is from three years ago, at which time the answer was "GHC-API is new and there isn't good documentation yet." So my hope is that the status has changed.
EDIT: the "dry-run" restriction is because I'm doing this in a web-based setting where compilation happens server side, so I'd like to avoid unnecessary disk reads/write every time the user hits "check". The executable would just get thrown away anyways, until they had a version ready to run.
Just to move this to an answer, this already exists as ghc-mod, here's the homepage. This already has frontends for Emacs, Sublime, and Vim so if you need examples of how to use it, there are plenty. In essence ghc-mod is just what you want, a wrapper around the GHC API designed for editors.
I have a stack which was originally built in Hypercard then migrated to Metacard. Obviously, it has expanded greatly over that time. Some core features broke when I tried to migrate to Runrev which is why I've waited till now to finally do that. I'm keeping it as a stack rather than an exe so I can save changes to it. I've built a standalone player to launch it and that is working. I've included the revmessagebox.rev stack in the Standalone Stack settings. This does add it but, incorrectly. I can put messages to it from my stack but, it won't run commands and it's missing all it's icons. I'm also included the revimagelibrary.rev and revtools.rev stacks in the hopes of fixing this but, no dice. I was also hoping that including revimagelibrary.rev would get my old Metacard icons to display but, no dice. I appreciate any help I can get on this.
Rich
I don't think you can. The message box is part of the IDE and requires the development environment to run. When you build a standalone your scripts etc are compiled and an interpreter for commands is no longer present.
To replicate it in a standalone you could use a simple window with a field to accept text and would require you passing the text entered to a "Do" command. The other functions present with the message box (accessed via the icons you mentioned) are also development tools and don't make much sense in a standalone.
The message box is not only integrated into the IDE, the engine also has hooks that directly support it. I'm not sure those hooks are included with the engine that is built into a standalone, so even if you adapt the existing message box for your standalone it still may not work correctly.
The solution, as others have said, is to build your own stack that functions as a pseudo-message box. It is easy to display messages in your own stack, and pretty easy to execute simple commands using the "do" command. It is somewhat more difficult to execute complex or multi-line commands. But I agree with Dunbarx that I'd assess the need for such a thing if you are planning this standalone for distribution. It's a non-standard interface element.
What James said. But note that though the msg box is indeed integral to the IDE, it is still just a stack, and that stack can be replicated to whatever extent you need.
That said, the msg box is usually used as a development tool, to test short scripts (usually one-liners), to get or set property values quickly, as a simple calculator, that sort of stuff. If you need that sort of functionality, you should probably integrate it more comprehensively into the structure of your project.
Craig Newman