After reading some reports about the inadequacies of the Mono 2.6 garbage collector, I decided to give Mono 2.10 a go. I found that the 2.10 runtime crashes with the following simple F# program:
let rec f x acc =
if x = 0 then acc
else f (x - 1) (acc + 1)
f 10 0
Equivalent looping C# code runs just fine as does an F# hello world program. The F# code also works ok with Mono 2.6 and .Net. Can anyone else reproduce this? Is it a bug or is it just my installation?
Here are the various runtimes I've tried and results I got.
Mono 2.10.2 (compiled from sources on Debian Squeeze)
-- "Stack overflow: IP: 0x4153bb84, fault addr: (nil)"
Mono 2.8 (compiled from sources on Debian Squeeze)
-- "Native stacktrace: ..."
Mono 2.10.2 (Windows binary)
-- "mono.exe has stopped working" dialog.
Mono 2.10.2 (VMware image)
-- Segmentation fault
Mono 2.6.7 (bundled with Debian Squeeze)
-- Works fine
The F# compiler used was from the November 2010 CTP.
I've reported this problem as bug #693905 at https://bugzilla.novell.com/show_bug.cgi?id=693905.
Related
I am new to linux device driver development. I am trying to write a simple hello world module that would display the version of kernel running on the system where I try to insert hello_world module into kernel.
I used LINUX_VERSION_CODE in version.h to get linux version and built the module.
When I try to insert the ko file on a different system other than where it was built, it still shows the version of kernel where it was built
I believe issue lies with using C macro.
Can someone help me how to find linux version of local machine where the ko is to be inserted instead of finding version of kernel where my module gets built
You're right - LINUX_VERSION_CODE is a macro that provides compile-time info about the version of the Linux headers that you're using to compile the module. The macro cannot have any knowledge about the version of the kernel that the module will actually be loaded into.
The utsname() function in <linux/utsname.h> provides a pointer to a new_utsname struct, which has sysname, release and version members that contain what you're looking for.
The information from these members is used in /proc/version, as shown in fs/proc/version.c :
static int version_proc_show(struct seq_file *m, void *v)
{
seq_printf(m, linux_proc_banner,
utsname()->sysname,
utsname()->release,
utsname()->version);
return 0;
}
linux_proc_banner is a string that is currently defined as follows :
const char linux_proc_banner[] =
"%s version %s"
" (" LINUX_COMPILE_BY "#" LINUX_COMPILE_HOST ")"
" (" LINUX_COMPILER ") %s\n";
On my system, reading /proc/version - and thus reading these members - results in obtaining the following string :
Linux version 4.1.6-1-ARCH (builduser#tobias) (gcc version 5.2.0 (GCC) ) #1 SMP PREEMPT Mon Aug 17 08:52:28 CEST 2015
Thus, sysname is Linux, release is 4.1.6-1-ARCH, and version is #1 SMP PREEMPT Mon Aug 17 08:52:28 CEST 2015.
I have installed FsUnit
» nuget install fsunit
Attempting to resolve dependency 'NUnit (≥ 2.6.3)'.
Installing 'NUnit 2.6.3'.
Successfully installed 'NUnit 2.6.3'.
Installing 'FsUnit 1.3.0.1'.
Successfully installed 'FsUnit 1.3.0.1'.
I have created simple unit test:
module Tests
open NUnit.Framework
open FsUnit
[<Test>]
let ``simple test`` () =
1 |> should equal 1
Here is I am launching my test:
» fsharpc -r NUnit.2.6.3/lib/nunit.framework.dll -r FsUnit.1.3.0.1/Lib/Net40/FsUnit.NUnit.dll 01_binomial_tests.fs
F# Compiler for F# 3.1 (Open Source Edition)
Freely distributed under the Apache 2.0 Open Source License
/Users/demas/development/book_exericses/fsharp_deep_dive/01_binomial_tests.fs(7,1): warning FS0988: Main module of program is empty: nothing will happen when it is run
It was compiled fine but I don't know how to launch the tests without VS
Update
I have tried to use NUnit.Runners:
> nuget install NUnit.Runners
> fsharpc -r NUnit.2.6.3/lib/nunit.framework.dll -r FsUnit.1.3.0.1/Lib/Net40/FsUnit.NUnit.dll --target:library 01_binomial_tests.fs
> mono NUnit.Runners.2.6.4/tools/nunit-console.exe 01_binomial_tests.dll
NUnit-Console version 2.6.4.14350
Copyright (C) 2002-2012 Charlie Poole.
Copyright (C) 2002-2004 James W. Newkirk, Michael C. Two, Alexei A. Vorontsov.
Copyright (C) 2000-2002 Philip Craig.
All Rights Reserved.
Runtime Environment -
OS Version: Unix 14.4.0.0
CLR Version: 2.0.50727.1433 ( Mono 3.5 ( 3.10.0 ((detached/92c4884 Thu Nov 13 23:27:38 EST 2014) ) )
ProcessModel: Default DomainUsage: Single
Execution Runtime: mono-3.5
Could not load file or assembly '01_binomial_tests, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null' or one of its dependencies. The system cannot find the file specified.
You have a couple of options:
Add dependency on NUnit.Runners NuGet package, which comes with a stand-alone runner
Use Xamarin Studio, which seems to have Unit Tests panel for running tests
Use FAKE which is a nice F# build tool that can run unit tests
I personally use FAKE because it nicely automates everything. To do that, you need to put something like this in your FAKE build.fsx script:
Target "Test" (fun _ ->
!! "/tests*.dll |> NUnit (fun p ->
{p with
DisableShadowCopy = true;
OutputFile = testDir + "TestResults.xml" })
)
You'll need dependencies on FAKE and NUnit.Runners and then FAKE should also find the runner automatically (so you do not have to set it explicitly in your build script).
I am working on compiling legacy Fortran 77 code on a Mac OS X 10.8.4 using gfortran 4.6.2. The gfortran was installed from http://gcc.gnu.org/wiki/GFortranBinaries#MacOS site.
According to the gcc.gnu.org wiki for GFortran, Release 4.2 includes this:
"The DEC extensions %VAL, %LOC and %REF are now supported."
I am getting the error below when I compile:
IPL(I) = %LOC(R8D(INITX(I)+1))
1
Error: Invalid character in name at (1)
Is there a workaround?
It seems that the new intrinsic LOC can be used instead:
The LOC() intrinsic works the same way as the %LOC() construct
http://gcc.gnu.org/onlinedocs/gfortran/LOC.html
Here is a bit of code that won't win awards for complexity:
[<EntryPoint>]
let main argv =
let i = 1I
printfn "One is %A\n" i
0 // return an integer exit code
It's compiled as follows: "c:/Program Files (x86)/Microsoft SDKs/F#/3.0/Framework/v4.0/Fsc.exe" --out:numericstest.exe --debug:full --target:exe --standalone Program.fs
Under Windows it produces the expected result. However under Mono 3.0.7 compiled under Ubuntu it instead says:
mono numericstest.exe
Unhandled Exception: System.InvalidProgramException: Invalid IL code in System.Numerics.BigInteger:get_One (): method body is empty.
at Program.main (System.String[] argv) [0x00000] in <filename unknown>:0[ERROR] FATAL UNHANDLED EXCEPTION: System.InvalidProgramException: Invalid IL code in System.Numerics.BigInteger:get_One (): method body is empty.
at Program.main (System.String[] argv) [0x00000] in <filename unknown>:0
What am I doing wrong please? Many thanks.
There's nothing wrong with your code -- that exception is due to your code anyway. It looks like there's something wrong with the System.Numerics.dll assembly on your machine; either it's not installed correctly, it's getting compiled incorrectly (e.g., by the Mono C# compiler), or it's doing some kind of type forwarding which is not working like it should, etc.
What happens if you run the code without using BigInteger (via the I suffix)?
I tried your code in my Ubuntu (12.04, 32-bit) VM running under VirtualBox. The code compiled and ran as expected. Here's the output if you want:
Compile/Run
jack#jack-linux:~/Desktop$ fsharpc --out:JoeHuha.exe --debug:full --target:exe --standalone JoeHuha.fs
F# Compiler for F# 3.0 (Open Source Edition)
Freely distributed under the Apache 2.0 Open Source License
jack#jack-linux:~/Desktop$ mono JoeHuha.exe
One is 1
Mono version info
jack#jack-linux:~/Desktop$ mono -V
Mono JIT compiler version 3.0.5 (master/1643364 Fri Feb 22 19:31:07 EST 2013)
Copyright (C) 2002-2012 Novell, Inc, Xamarin Inc and Contributors. www.mono-project.com
TLS: __thread
SIGSEGV: altstack
Notifications: epoll
Architecture: x86
Disabled: none
Misc: softdebug
LLVM: supported, not enabled.
GC: Included Boehm (with typed GC and Parallel Mark)
So it turns out lua for windows install has some earlier 2.1 version of luasql on it, and I need to be able to use luasql 2.2 (particularly mysql). I've spent all day trying to install this thing with luarocks but keep hitting a wall. This is the 3rd wall I've hit but the first one that I just have no clue where to even begin.
I run: luarocks install luasql-mysql MYSQL_DIR="E:/Programs/MySQL/MySQL Server 5.5"
It goes through and outputs (removing a lot of the output where it's just repeating the same thing but different file):
Extracting luasql-2.2.0\src\jdbc
Extracting luasql-2.2.0\src\jdbc\Makefile
... (lots of these here)
Extracting luasql-2.2.0\vc6\sqlite.def
Everything is Ok
Folders: 17
Files: 84
Size: 358091
Compressed: 440320
cl /MD /O2 -c -Fosrc/luasql.obj -IF:/Code/Lua/5.1/include src/luasql.c -IE:/Prog
rams/MySQL/MySQL Server 5.5/include
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 16.00.30319.01 for 80x86
Copyright (C) Microsoft Corporation. All rights reserved.
luasql.c
cl /MD /O2 -c -Fosrc/ls_mysql.obj -IF:/Code/Lua/5.1/include src/ls_mysql.c -IE:/
Programs/MySQL/MySQL Server 5.5/include
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 16.00.30319.01 for 80x86
Copyright (C) Microsoft Corporation. All rights reserved.
ls_mysql.c
e:\programs\mysql\mysql server 5.5\include\mysql_com.h(291) : error C2061: synta
x error : identifier 'SOCKET'
e:\programs\mysql\mysql server 5.5\include\mysql_com.h(337) : error C2059: synta
x error : '}'
... (lots of these here, same errors just different lines)
E:/Programs/MySQL/MySQL Server 5.5/include\mysql.h(374) : error C2143: syntax er
ror : missing '{' before '*'
E:/Programs/MySQL/MySQL Server 5.5/include\mysql.h(374) : fatal error C1003: err
or count exceeds 100; stopping compilation
Error: Build error: Failed compiling object src/ls_mysql.obj
Any idea where to start? I've had to a) install standalone mysql (was previously using one that came with WAMP since I was already using that before), and b) install visual studio 2010 (c# and c++ versions so far), and c) use visual studio command prompt to run the luarocks stuff, if I use regular command problem it breaks way before this.
I'm hoping this is an easy issue to fix for someone familiar with compiling C which is what this seems to be doing.
Add the following to ls_mysql.c after line 17:
#include <windows.h>
So that windows.h is included after winsock.h and before mysql.h.
References:
http://forums.mysql.com/read.php?45,37472,37472
C++ Redefinition Header Files (winsock2.h)