On Thu, 2008-05-15 at 12:52 +0530, Subrata Modak wrote:
> On Wed, 2008-05-14 at 17:25 -0700, Garrett Cooper wrote:
> > On Wed, May 14, 2008 at 5:50 AM, Subrata Modak
> > <[EMAIL PROTECTED]> wrote:
> >         On Wed, 2008-05-14 at 17:57 +0530, Rishikesh K. Rajak wrote:
> >         > On Wed, 2008-05-14 at 17:31 +0530, Subrata Modak wrote:
> >         > > Hi,
> >         > >
> >         > > I was thinking why not enable LTP to generate the OUTPUT
> >         and the LOG
> >         > > files as default. Let me know the pros and cons of this. I
> >         have created
> >         > > the attached patch for implementing the same:
> >         >
> >         > HI Subrata,
> >         >
> >         > Thanks for patch. But, just i was thinking it would be a
> >         better idea
> >         > to leave to the user, where and in which format he wants to
> >         create
> >         > result file. Just an idea.
> >         
> >         
> >         Hi Rishi,
> >         
> >         Thanks for your comment. This patch (and the consequent
> >         enhancement)
> >         does not take away user freedom to generate his/her own away
> >         of creating
> >         output and log files. He/she can still use the options -l and
> >         -o to
> >         generate the same. What this patch provides is to create a
> >         output and
> >         log file only when the user has not mentioned them on ltp run.
> >         
> >         The idea of creating default output/log is from a user only,
> >         who felt
> >         that it is LTP´s drawback of not generating so.
> >         
> >         Regards--
> >         Subrata
> >         
> >         
> >         >
> >         > - Rishi
> >         > >
> >         > > Signed-off-by: Subrata Modak <[EMAIL PROTECTED]>
> >         > >
> >         > > Regards--
> >         > > Subrata
> > 
> > 
> > Pro: You have your output all the time, so no more going "oops! I
> > forgot to specify that -o / -l option!"
> > Con: As we all know, many embedded systems have a limited amount of
> > space, be it physical or virtual, s.t. more logs will clog up a
> > running embedded system faster. This is especially true because tar
> > uncompressed is used instead of with some form of compression (bzip2,
> > cpio, gzip, etc).
> > 
> 
> True. So, i think, we can skip generating the output file (under
> ltp/output by default), which actually rakes up huge space. Let the user
> specify this.
> 
> But we can go ahead in generating a log file (under ltp/results) by
> default, as, log file contains just a line summary of PASS/FAIL for each
> test case, and occupies very less space. This really will come handy in
> situation where the user had really oopss.......s..ed, discovering at
> the midst of the run, as he has forgot to specifiy the options at
> runtime. A handy default logfile will really simplify the job. What do
> you say ? Patch attached.

On further offline discussion we found that the necessity for generation
of a logfile may be mandatory as it´s size is not a constraint for most
of the systems. I have applied this patch and hence future LTP will
automatically generate a log file as default.

Having said that, i would like to see patches which disables this
(default logfile generation) in architectures where even logfile size is
a constraint to save, given the nature of embedded hardware they have.

Regards--
Subrata

> 
> Regards--
> Subrata
> 
> > Why can't the user specify that they want logs instead of forcing the
> > rest of the users to conform to this behavior? Then again
> > using /dev/null or /dev/std{err|out} would work too I suppose...
> > 
> > -Garrett
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft 
> Defy all challenges. Microsoft(R) Visual Studio 2008. 
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________ Ltp-list mailing list 
> [email protected] 
> https://lists.sourceforge.net/lists/listinfo/ltp-list


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to