I don't know if I'm misunderstanding the use of shell() or if this is a bug. Trying to run some programs with shell error with a could not close file. Sadly it I can come to any reasonable conclusion as to why.

Example output from running writeln(shell("gcc"))
gcc: no input files
std.contracts.ErrnoException: std/stdio.d(336): Could not close file
`gcc' (Success)

Note that the program was executed correctly printing the error message, and the appended (Success) massage...? Also shell("gcc -v") works as with proper output and no error.

Similar errors trying to run shell("dmd") unsure if this is related to the programs not being in /bin.

Jesse Phillips wrote:
> I don't know if I'm misunderstanding the use of shell() or if this is a bug. Trying to run some programs with shell error with a could not close file. Sadly it I can come to any reasonable conclusion as to why.
> 
> Example output from running writeln(shell("gcc"))
> gcc: no input files
> std.contracts.ErrnoException: std/stdio.d(336): Could not close file `gcc' (Success)
> 
> Note that the program was executed correctly printing the error message, and the appended (Success) massage...? Also shell("gcc -v") works as with proper output and no error.
> 
> Similar errors trying to run shell("dmd") unsure if this is related to the programs not being in /bin.


I think it's a bug in Phobos. Note that you get the exact same result even if you remove the enclosing writeln(), which means that the program's standard output stream hasn't been intercepted. Although I'm unsure about the details, it seems that the std.stdio.File object that should encapsulate the program's output stream isn't opened correctly, and thus an exception is thrown when an attempt is made to close it.

Probably you don't get an error for "gcc -v" because the output of this command is printed to standard error instead of standard output.

-Lars

On Tue, 07 Jul 2009 12:04:43 +0200, Lars T. Kyllingstad wrote:

> Jesse Phillips wrote:
>> I don't know if I'm misunderstanding the use of shell() or if this is a bug. Trying to run some programs with shell error with a could not close file. Sadly it I can come to any reasonable conclusion as to why.
>> 
>> Example output from running writeln(shell("gcc")) gcc: no input files
>> std.contracts.ErrnoException: std/stdio.d(336): Could not close file
>> `gcc' (Success)
>> 
>> Note that the program was executed correctly printing the error
>> message, and the appended (Success) massage...? Also shell("gcc -v")
>> works as with proper output and no error.
>> 
>> Similar errors trying to run shell("dmd") unsure if this is related to the programs not being in /bin.
> 
> 
> I think it's a bug in Phobos. Note that you get the exact same result even if you remove the enclosing writeln(), which means that the program's standard output stream hasn't been intercepted. Although I'm unsure about the details, it seems that the std.stdio.File object that should encapsulate the program's output stream isn't opened correctly, and thus an exception is thrown when an attempt is made to close it.
> 
> Probably you don't get an error for "gcc -v" because the output of this command is printed to standard error instead of standard output.
> 
> -Lars

Thanks, I try and narrow down when this happens. I first assumed it was because running GCC without anything outputs to stderr and not stdout, but that wouldn't explain the DMD issue. ("ggc -v" goes to stdout)

Jesse Phillips wrote:
> On Tue, 07 Jul 2009 12:04:43 +0200, Lars T. Kyllingstad wrote:
> 
>> Jesse Phillips wrote:
>>> I don't know if I'm misunderstanding the use of shell() or if this is a
>>> bug. Trying to run some programs with shell error with a could not
>>> close file. Sadly it I can come to any reasonable conclusion as to why.
>>>
>>> Example output from running writeln(shell("gcc")) gcc: no input files
>>> std.contracts.ErrnoException: std/stdio.d(336): Could not close file
>>> `gcc' (Success)
>>>
>>> Note that the program was executed correctly printing the error
>>> message, and the appended (Success) massage...? Also shell("gcc -v")
>>> works as with proper output and no error.
>>>
>>> Similar errors trying to run shell("dmd") unsure if this is related to
>>> the programs not being in /bin.
>>
>> I think it's a bug in Phobos. Note that you get the exact same result
>> even if you remove the enclosing writeln(), which means that the
>> program's standard output stream hasn't been intercepted. Although I'm
>> unsure about the details, it seems that the std.stdio.File object that
>> should encapsulate the program's output stream isn't opened correctly,
>> and thus an exception is thrown when an attempt is made to close it.
>>
>> Probably you don't get an error for "gcc -v" because the output of this
>> command is printed to standard error instead of standard output.
>>
>> -Lars
> 
> Thanks, I try and narrow down when this happens. I first assumed it was because running GCC without anything outputs to stderr and not stdout, but that wouldn't explain the DMD issue. ("ggc -v" goes to stdout)


Then you have a different version of GCC than I. From the man page of GCC 4.3.3:

  -v  Print (on standard error output) the commands executed to run the
      stages of compilation.  Also print the version number of the
      compiler driver program and of the preprocessor and the compiler
      proper.

-Lars

Jesse Phillips wrote:
> On Tue, 07 Jul 2009 12:04:43 +0200, Lars T. Kyllingstad wrote:
> 
>> Jesse Phillips wrote:
>>> I don't know if I'm misunderstanding the use of shell() or if this is a
>>> bug. Trying to run some programs with shell error with a could not
>>> close file. Sadly it I can come to any reasonable conclusion as to why.
>>>
>>> Example output from running writeln(shell("gcc")) gcc: no input files
>>> std.contracts.ErrnoException: std/stdio.d(336): Could not close file
>>> `gcc' (Success)
>>>
>>> Note that the program was executed correctly printing the error
>>> message, and the appended (Success) massage...? Also shell("gcc -v")
>>> works as with proper output and no error.
>>>
>>> Similar errors trying to run shell("dmd") unsure if this is related to
>>> the programs not being in /bin.
>>
>> I think it's a bug in Phobos. Note that you get the exact same result
>> even if you remove the enclosing writeln(), which means that the
>> program's standard output stream hasn't been intercepted. Although I'm
>> unsure about the details, it seems that the std.stdio.File object that
>> should encapsulate the program's output stream isn't opened correctly,
>> and thus an exception is thrown when an attempt is made to close it.
>>
>> Probably you don't get an error for "gcc -v" because the output of this
>> command is printed to standard error instead of standard output.
>>
>> -Lars
> 
> Thanks, I try and narrow down when this happens. I first assumed it was because running GCC without anything outputs to stderr and not stdout, but that wouldn't explain the DMD issue. ("ggc -v" goes to stdout)


It happens whenever the program you try to run has a nonzero return value. I'll file a bug report, if you haven't done it already.

-Lars

On Thu, 09 Jul 2009 11:20:54 +0200, Lars T. Kyllingstad wrote:

> Jesse Phillips wrote:
>> On Tue, 07 Jul 2009 12:04:43 +0200, Lars T. Kyllingstad wrote:
>> 
>>> Jesse Phillips wrote:
>>>> I don't know if I'm misunderstanding the use of shell() or if this is a bug. Trying to run some programs with shell error with a could not close file. Sadly it I can come to any reasonable conclusion as to why.
>>>>
>>>> Example output from running writeln(shell("gcc")) gcc: no input files
>>>> std.contracts.ErrnoException: std/stdio.d(336): Could not close file
>>>> `gcc' (Success)
>>>>
>>>> Note that the program was executed correctly printing the error
>>>> message, and the appended (Success) massage...? Also shell("gcc -v")
>>>> works as with proper output and no error.
>>>>
>>>> Similar errors trying to run shell("dmd") unsure if this is related to the programs not being in /bin.
>>>
>>> I think it's a bug in Phobos. Note that you get the exact same result even if you remove the enclosing writeln(), which means that the program's standard output stream hasn't been intercepted. Although I'm unsure about the details, it seems that the std.stdio.File object that should encapsulate the program's output stream isn't opened correctly, and thus an exception is thrown when an attempt is made to close it.
>>>
>>> Probably you don't get an error for "gcc -v" because the output of this command is printed to standard error instead of standard output.
>>>
>>> -Lars
>> 
>> Thanks, I try and narrow down when this happens. I first assumed it was because running GCC without anything outputs to stderr and not stdout, but that wouldn't explain the DMD issue. ("ggc -v" goes to stdout)
> 
> 
> It happens whenever the program you try to run has a nonzero return value. I'll file a bug report, if you haven't done it already.
> 
> -Lars

Thanks.