Re: [AMBER-Developers] General test failure question

From: Gustavo Seabra <gustavo.seabra.gmail.com>
Date: Fri, 26 Feb 2010 16:16:10 -0300

On Fri, Feb 26, 2010 at 2:01 PM, Ben Roberts <roberts.qtp.ufl.edu> wrote:
>
> On 26/02/2010, at 11:52 AM, Gustavo Seabra wrote:
>
>> On Fri, Feb 26, 2010 at 1:29 PM, case <case.biomaps.rutgers.edu> wrote:
>>>
>>> On Fri, Feb 26, 2010, Gustavo Seabra wrote:
>>>>
>>>> I suggest a simple "quick-and-dirty" solution: Just put the minuses
>>>> ("-") before each test, so the test doesn't stop at any of those,
>>>
>>> I think(?) that the "-k" flag to make accomplishes the same thing, and
>>> involves less instrusion than adding minus signs everywhere.
>>>
>> Good... I didn't know about this flag. Is there a way to enforce it
>> without the user having to put it himself in the make command?
>>
>> Also, is there a way to force output to a file without the need to
>> redirect in every line?
>>
>> Gustavo.
>
> My solution: Put the make command in a shell script and advise the user to
> run the shell script rather than "make" directly.
>
> Like so:
>
> #!/bin/bash
>
> date_string=`date %Y%m%d%H%M%S`
>
> (make -k -f Makefile_at 2>&1) | tee Makefile_at.serial.${date_string}.log
>
>
>
> That way, there's automatically a log file, timestamped so the user can, if
> he or she so wishes, compare the output of different tests. Also, though I
> haven't implemented this yet, one could use awk (I guess...) to count the
> number of passes/failures/errors at the end of the shell script; provided
> the output strings are consistent enough to be greppable.
>
> Ben
>

That should work as well. The use of that script will also avoid the
need to modify the Makefile. And, it is indeed possible to use awk to
count the failures, as long as we have the reports loosely consistent.

Gustavo.

_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Feb 26 2010 - 11:30:02 PST
Custom Search