[petsc-dev] [petsc-users] make test failed

[Satish Balay]

On Tue, 19 Jun 2012, Barry Smith wrote:

> 
> On Jun 19, 2012, at 1:47 PM, Satish Balay wrote:
> 
> > On Tue, 19 Jun 2012, Barry Smith wrote:
> > 
> >> 
> >> On Jun 19, 2012, at 1:27 PM, Satish Balay wrote:
> >> 
> >>> Ok - I think the sources were extracted with a windows tool - and not
> >>> cygwin 'tar' - so the scripts [which got extracted in dos format] are
> >>> misbehaving.
> >> 
> >>   Satish,
> >> 
> >>  Ok, how do we prevent this from ever happening again? 
> >> 
> >>   When ./configure is run it should first check if the scripts are corrupted from a Windows tool and stop with a nice error message telling the user to use cygwin tar to process the tar file.
> >> 
> >>    Please figure out how to make this a patch for 3.3
> > 
> > Perhaps there is only one script [mpiexec.uni] thats relavent on windows.
> 
> petscmpiexec? petscrun?

Perhaps gradually migrate? They are not in the critical to 'configre
&& make && make test' path. [as mpiexec.uni is]

> all the bash code inside the makefiles?

Make is handling procesing those strings and passing to bash - so
perhaps its doing the correct thing? [as the build was successful]

> What exactly is the windows tool changing in the sh file? Maybe just
>  reformatting the file will solve the problem?

Its the 'dos' mode vs 'unix' mode thingy where you have '\n\r' vs
'\n'.

It appears that the extra '\r' char is causing bash to process strings
incorrectly.

Presumably 'winzip' or similar tool created all text files
[sources,makefiles,scripts] etc in the dos mode.

Satish


> 
>   Barry
> 
> 
> 
> > 
> > If this is the case - perhaps converting it it from bash to python is
> > the way to go? [as python is unaffected by dos format - configure ran
> > fine]
> > 
> > Satish
> 
>