[PATCH] don't print dmidecode shell error

[Stefan Reinauer]

See patch

[Stefan Reinauer]

* Stefan Reinauer  [110329 21:55]:
> Don't print the error
>
> sh: dmidecode: not found
>
> if dmidecode is not there.
>
> Signed-off-by: Stefan Reinauer
> Index: dmi.c
>
> --- dmi.c	(revision 1282)
> +++ dmi.c	(working copy)
> @@ -75,7 +75,7 @@
>  };
>
> #define DMI_COMMAND_LEN_MAX 260
> -static const char *dmidecode_command = "dmidecode";
> +static const char *dmidecode_command = "dmidecode 2>/dev/null";
>
> static char *dmistrings[ARRAY_SIZE(dmidecode_names)];

Any opinions on a half line patch? :)

[Michael Karcher]

Am Dienstag, den 05.04.2011, 23:48 +0200 schrieb Stefan Reinauer:
> > -static const char *dmidecode_command = "dmidecode";
> > +static const char *dmidecode_command = "dmidecode 2>/dev/null";
>
> Any opinions on a half line patch? :)

Not every OS has a unix command interpreter. This will break on DOS.
Most likely fine if DOS is detected per ifdef.

Regards,
  Michael Karcher

[Stefan Reinauer]

* Michael Karcher  [110406 09:53]:
> Am Dienstag, den 05.04.2011, 23:48 +0200 schrieb Stefan Reinauer:
> > > -static const char *dmidecode_command = "dmidecode";
> > > +static const char *dmidecode_command = "dmidecode 2>/dev/null";
> >
> > Any opinions on a half line patch? :)
>
> Not every OS has a unix command interpreter. This will break on DOS.
> Most likely fine if DOS is detected per ifdef.

Ok, how about this one?

[Michael Karcher]

Am Donnerstag, den 07.04.2011, 18:40 +0200 schrieb Stefan Reinauer:
> Michael Karcher
> > Not every OS has a unix command interpreter. This will break on
> > DOS. Most likely fine if DOS is detected per ifdef.
>
> Ok, how about this one?
> Signed-off-by: Stefan Reinauer

Acked-by: Michael Karcher
> +#if defined (__DJGPP__) || defined (_WIN32)

That might be overly conservative, but I am fine with out. while
command.com does not support 2>something, I am quite sure cmd.exe does.

Regards,
  Michael Karcher

[Owen Mann]

This should work on DOS; the only change is from "/dev/null" to "NUL":
C:\HOTFIXES-ng>command
Microsoft(R) Windows DOS
(C)Copyright Microsoft Corp 1990-2001.

C:\HOTFIX~2>echo foo
foo

C:\HOTFIX~2>echo foo>NUL

C:\HOTFIX~2>dir NUL

Directory of \.

File Not Found

[granted this is on XP; but it should work anywhere]

---- Original message ----
> Date: Thu, 07 Apr 2011 19:16:04 +0200
> From: flashrom-bounces+owen=[e-mail] (on behalf of Michael Karcher )
> Subject: Re: [flashrom] [PATCH] don't print dmidecode shell error  
> To: Stefan Reinauer 
> Cc: [e-mail]
>
> Am Donnerstag, den 07.04.2011, 18:40 +0200 schrieb Stefan Reinauer:
> > Michael Karcher
> > > Not every OS has a unix command interpreter. This will break on
> > > DOS. Most likely fine if DOS is detected per ifdef.
> >
> > Ok, how about this one?
> > Signed-off-by: Stefan Reinauer
>
> Acked-by: Michael Karcher
> > +#if defined (__DJGPP__) || defined (_WIN32)
>
> That might be overly conservative, but I am fine with out. while
> command.com does not support 2>something, I am quite sure cmd.exe does.
>
> Regards,
>  Michael Karcher
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[Owen Mann]

---- Original message ----
> ---- Original message ----
> > Date: Thu, 07 Apr 2011 19:16:04 +0200
> > From: flashrom-bounces+owen=[e-mail] (on behalf of Michael Karcher )
> > Subject: Re: [flashrom] [PATCH] don't print dmidecode shell error  
> > To: Stefan Reinauer 
> > Cc: [e-mail]
> >
> > Am Donnerstag, den 07.04.2011, 18:40 +0200 schrieb Stefan Reinauer:
> > > Michael Karcher
> > > > Not every OS has a unix command interpreter. This will break on
> > > > DOS. Most likely fine if DOS is detected per ifdef.
> > >
> > > Ok, how about this one?
> > > Signed-off-by: Stefan Reinauer
> >
> > Acked-by: Michael Karcher
> > > +#if defined (__DJGPP__) || defined (_WIN32)
> >
> > That might be overly conservative, but I am fine with out. while
> > command.com does not support 2>something, I am quite sure cmd.exe does.
> >
> > Regards,
> >  Michael Karcher
> >
> > flashrom mailing list
> > [e-mail]
> > http://www.flashrom.org/mailman/listinfo/flashrom
>
> Date: Thu,  7 Apr 2011 13:28:16 -0400 (EDT)
> From: [e-mail] (on behalf of Owen Mann )
> Subject: Re: [flashrom] [PATCH] don't print dmidecode shell error  
> To: [e-mail]
>
> This should work on DOS; the only change is from "/dev/null" to "NUL":
> C:\HOTFIXES-ng>command
> Microsoft(R) Windows DOS
> (C)Copyright Microsoft Corp 1990-2001.
>
> C:\HOTFIX~2>echo foo
> foo
>
> C:\HOTFIX~2>echo foo>NUL
>
> C:\HOTFIX~2>dir NUL
>
> Directory of \.
>
> File Not Found
>
> [granted this is on XP; but it should work anywhere]
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

Sorry, I missed the "2" (and sorry for the top-post).
This works in command.com (in XP) too:
C:>cat foo
'CAT' is not recognized as an internal or external command,
operable program or batch file.
C:>cat foo 2>NUL

[Stefan Tauner]

On Thu, 07 Apr 2011 19:16:04 +0200
Michael Karcher  wrote:
> Am Donnerstag, den 07.04.2011, 18:40 +0200 schrieb Stefan Reinauer:
> > Michael Karcher
> > > Not every OS has a unix command interpreter. This will break on
> > > DOS. Most likely fine if DOS is detected per ifdef.
> >
> > Ok, how about this one?
> > Signed-off-by: Stefan Reinauer
>
> Acked-by: Michael Karcher
> > +#if defined (__DJGPP__) || defined (_WIN32)
>
> That might be overly conservative, but I am fine with out. while
> command.com does not support 2>something, I am quite sure cmd.exe does.

stepan told me to commit this, but i hesitate. i want to integrate
dos/win support first.

i would drop the __DJGPP__ flag there because /dev/null is supported by
it: http://www.delorie.com/djgpp/doc/eli-m17n99.html#Unix%20devices
additionally i would use 2>NUL for win32. for xp there is official
documentation. a friend of mine also tried it on win2k so i would say
it is fairly safe to use it?

[Bernd Blaauw]

Op 27-6-2011 14:44, Stefan Tauner schreef:
> stepan told me to commit this, but i hesitate. i want to integrate
> dos/win support first.
>
> i would drop the __DJGPP__ flag there because /dev/null is supported by
> it: http://www.delorie.com/djgpp/doc/eli-m17n99.html#Unix%20devices
> additionally i would use 2>NUL for win32. for xp there is official
> documentation. a friend of mine also tried it on win2k so i would say
> it is fairly safe to use it?

It would be kind to be able to detect a missing DMIDECOD.EXE (DOS) or 
DMIDECODE.EXE (LFN platforms, likely thus only Windows/ReactOS) only 
once, instead of ending up with
6 times the line "bad command or filename" (sent to stdout only, unable 
to redirect to textfile). Searching current directory, flashrom's 
directory as well as entire path, should do, I guess.

[Stefan Tauner]

On Mon, 27 Jun 2011 14:54:49 +0200
Bernd Blaauw  wrote:
> Op 27-6-2011 14:44, Stefan Tauner schreef:
> > stepan told me to commit this, but i hesitate. i want to integrate
> > dos/win support first.
> >
> > i would drop the __DJGPP__ flag there because /dev/null is supported by
> > it: http://www.delorie.com/djgpp/doc/eli-m17n99.html#Unix%20devices
> > additionally i would use 2>NUL for win32. for xp there is official
> > documentation. a friend of mine also tried it on win2k so i would say
> > it is fairly safe to use it?
>
> It would be kind to be able to detect a missing DMIDECOD.EXE (DOS) or 
> DMIDECODE.EXE (LFN platforms, likely thus only Windows/ReactOS) only 
> once, instead of ending up with
> 6 times the line "bad command or filename" (sent to stdout only, unable 
> to redirect to textfile).

does it with the previously attached patch still?
> Searching current directory, flashrom's 
> directory as well as entire path, should do, I guess.

patches are welcome ;)

[Bernd Blaauw]

Op 27-6-2011 15:29, Stefan Tauner schreef:
> does it with the previously attached patch still?

Can't say, I'm a simple Windows/Dos user, thus leeching on Idwer's DOS 
compilation.
Someday I'll start with Linux and compiling, likely the day I purchase 
that Asrock E350M board for Coreboot attempting to create a 'full' 
opensource machine.
> > Searching current directory, flashrom's
> > directory as well as entire path, should do, I guess.
>
> patches are welcome ;)

Thanks for the offer :)