[TxMt] Re: Capturing error messages

[Martin Kühl]

On Sat, Jul 2, 2011 at 23:14, Greg <web at web.knobby.ws> wrote:
>
> On Jun 30, 2011, at 10:42 PM, Martin Kühl wrote:
>
>> On Thu, Jun 30, 2011 at 21:58, Greg <web at web.knobby.ws> wrote:
>>> I'm running a Ruby script from TextMate and within that script call a
>>> Perl script (gpsPhoto.pl) and would like to capture the error messages
>>> and act on them. The messages show up in the TM Running window.
>>
>> There's nothing TextMate-specific about this, you can capture those
>> messages any way you would otherwise in Ruby. Which way that is depends
>> on how you call the other command, e.g. with backticks you would append
>> `2>&1` to the command line.
>>
>> If you're unsure, [1] has a survey of ways to run subprocesses in Ruby.
>>
>> HTH,
>> Martin
>>
>> [1] http://tech.natemurray.com/2007/03/ruby-shell-commands.html
>>
> Neophyte that I am it took a little bit of thinking. Mainly was to just set the call to Perl like so:
>
>      perlOutput = `perl \"#{gpsPhotoLoc}\"   --image \"#{imageFile}\" --gpsfile  \"#{gpxFile}\"   --timeoffset #{geoOffset} --maxtimediff #{maxTimeDiff} --maxdistance #{maxDistance}  --geoinfo wikipedia --city auto --state guess --country guess`
>
> I didn't have the perlOutput = before. $STDERR has info when the Perl
> script has an error and perlOutput gives info on what happens when the
> script runs. So found what I was looking for.
>
> One question: would the `2>&1` comment still apply? It didn't make any
> difference when I added the 2>&1 to the end of backtick quoted line.

Short answer: No, probably not.

Longer answer: Scripts have two "channels" for messages, one for output
and one for error messages. After your command, `perlOutput` contains
messages from the former. If you append `2>&1` to the command (inside
the backticks) it will contain messages from either. If a script doesn't
use the error channel it won't make any difference.

> Learning slowly.

Isn't that the only way? :-)

Martin