Retrieving terminal output status?

Hi,

If you are using do shell script, you set a variable to the result. like this:

set myVar to do shell script "ls -al"

Although, on second reading, what you say you are doing may be more convoluted. Could you expand on what the actual process is?

Best wishes

John M

If you're using Terminal.app to run the shell commands (rather than AppleScript's built-in do shell script) then there's no easy way to capture the output since AppleScript has no knowledge of the shell process, or when it completes. The only way to do this with Terminal.app is to periodically poll the process list to see if your process is still running, then grab the output from the respective window.

Calling your shell command via do shell script is much easier.

Regarding your

set exitStatus to do shell script "command; echo $?"

"do shell script" is Apple scripts way of running a unique shell script which has no connection to anything currently running in any window or tab in the terminal app.

This commant is therefore useless.

Sure, but as I stated earlier, that is far from a 'solution'

For one, that returns the entire window contents - potentially a lot of data - and you have to parse through that to find the data you need,
Secondly, How do you know if your process has finished? It might be a task that takes a long time - or, even worse, a variable amount of time - so you have to implement some technique to detect when the process has finished and the 'data you seek' is actually available.

It might be the way to go, but for 99% of tasks do shell script (which doesn't use Terminal.app at all) is a better solution.

do shell script runs a different form of shell, not the same shell as Terminal app.

do shell script runs a unique shell that ends and can never be referenced ever again.

So if your shell commands require any follow up commands, they must be executed in the same line seporated with ; otherwise entering username and passwords for example, at multiple prompts are impossible.

So for 99% of tasks that require reading the output data, "do shell script" is not going to work.

If you need the output, you proobobly intend to follow up with another shell command based on that output.

That cannot be done in a unique applescript shell that ends and can never continue where it left off.

You can check the progress of a running shell with

set commandString to "/bin/ls -AelO" as text
tell application "Terminal"
set windowCount to (count of the windows)
if windowCount is greater than 0 then
repeat with w from 1 to windowCount
tell window 1
set windowID to id
set theTabs to tabs
repeat with thisTab in theTabs
if thisTab is not busy then
do script commandString in thisTab
set selected of thisTab to true
set frontmost to true
return
end if
end repeat
end tell
set frontmost of window id windowID to false
end repeat
end if
tell window 1
do script commandString
set frontmost to true
end tell
end tell

However I prefer the ease of do shell script when you only need one line of shell to run, and do not need any followup commands.

Not only is that THE solution (vs your "far from a solution") it is simple easy, and you mislead people by saying "there's no easy way to capture the output since AppleScript has no knowledge of the shell process, or when it completes."

When I read that I thought you where saying it was impossible, not easy without some output to file that AS must read etc. Then I find it is a one line command to get the entire window. Sure you have to parse the data, but is you are scripting in the shell.. we are above and past calling that difficult.

Regarding your

set exitStatus to do shell script "command; echo $?"

"do shell script" is Apple scripts way of running a unique shell script which has no connection to anything currently running in any window or tab in the terminal app.

This commant is therefore useless.

While the Applescript fragment is not mine (Linc's actually), there is nothing useless about it. It just might not be useful for something you want to do.

Linc's example will tell you if the 'command' you specified completed successfully or not. That can be very useful if you know that the command you are using will return a non-zero status if it did not do what was requested.

Now, reading the other things you have posted in this thread that you are hijacking, it would appear you are not interested in either capturing the output from a command executed from within the Applescript, nor capturing the completion status from a command executed from within the Applescript. It appears you want an Applescript that interacts with a Terminal session.

Given that, you would be better off a) starting a new thread, so you can give points for useful answers (the only payment anyone gets in these forums), and b) clearly stating what problem you are trying to solve.

Perhaps you can give me an example of how "set exitStatus to do shell script "command; echo $?" will ever have any result other than "0"?

Because my understanding is that everytime you call "do shell script" in AS, you get a fresh new & Unique shell to execute the command(s) that have no way of picking up where the last "do shell script" left off. Therefore I must be wrong in what I read because it would mean you could never get any result other than "0" with that command that I reffered to as useless. So please prove me wrong.

Am I hyjacking?

I came to this thread while looking for a method to read the terminal app output from within AS because "do shell script" cannot continue multiple commands from where the last "do shell script" left off. It does however make an easy single line shell that returns a clean result good for a one time only thing. Even for multiple commands on the same line, but not for functions that change depending on the output from the shell.

I have found the solution was to use terminal, and the problem was "do shell script" does not execute the same bash as terminal, nor can it continue where it left off.

My "hyjacking" was to correct the incorrect statements made by others so that other people who are trying to solve something and find this thread find the answer as well as know some of the information posted on this thread was dead wrong.

I am not looking for points, payment, kudos or political he-said-she-said.

Just correcting people that post pointless code such as the one you quoted me commenting on as being useless because it will always return "0" no matter what. Right?

I stand corrected, you are getting a "1"!

Ok

So how is echo returning any output that you aren't already getting returned on that same line?

It seems (almost) redundant but it is much cleaner then text with an error though. And I still stand corrected since you are getting a "1"

So how is echo returning any output that you aren't already getting returned on that same line?

The difference is, that the error is caught, but the error status is saved in the variable "errorstatus". Your script can continue.

Without the "echo" the script will be terminated with an error message.

set exitStatus to do shell script "cat /Users/nonexisting"

return "Done"

terminates with an error, the "return" will not be executed.

set exitStatus to do shell script "cat /Users/nonexisting; echo $?"

return "Done"

terminates without an error, after executing the following commands, but you save the error status in "exitstatus"

I just go with try, on error, end try

I found this thread while looking for output collection, not error trapping.

Echo works in one line rather than multiple try , end try lines being added I guess.

Interesting way to do that.

Perhaps you can give me an example of how "set exitStatus to do shell script "command; echo $?" will ever have any result other than "0"?

Because my understanding is that everytime you call "do shell script" in AS, you get a fresh new & Unique shell to execute the command(s) that have no way of picking up where the last "do shell script" left off. Therefore I must be wrong in what I read because it would mean you could never get any result other than "0" with that command that I reffered to as useless. So please prove me wrong.

LeonieDF gave you an example, which you acknowledged.

The semicolon between the command the the "echo $?" allows more than one command to be executed in the same shell session. The shell saves the completion status of every command in the shell variable $?, the echo command it just sending the contents of $? to standard out (stdout), and the Applescript is putting the output from stdout into the exitStatus variable.

You could have a long serious commands separated by semicolons, vertical bar | pipe symbols, and they would all execute as part of the same shell session. You could even invoke a shell script that would be 1000's of lines long and it would be invoked in that shell session (along with subshells invoked by the initial shell session).

You are connect that each and every "do shell script" starts a new shell session. But you are wrong in that multiple commands issued in the same shell session start unrelated shells.

Am I hyjacking?

Yes in that you are taking control of this thread, but are unable to give any credit to anyone for information you find valuable. Then again, maybe you have not found anything valuable from any of the contributes answering your questions.

I came to this thread while looking for a method to read the terminal app output from within AS because "do shell script" cannot continue multiple commands from where the last "do shell script" left off. It does however make an easy single line shell that returns a clean result good for a one time only thing. Even for multiple commands on the same line, but not for functions that change depending on the output from the shell.

I have found the solution was to use terminal, and the problem was "do shell script" does not execute the same bash as terminal, nor can it continue where it left off.

But that was not necessarily what the original author, Hi], was asking for. The Hi] seemed to be satisfied with the answer given by John Maisey. Hi] gave John 10 points for his answer.

The last paragraph can be parsed 2 ways. a) the version of bash execute by the Terminal and by "do Shell Script" are different, and b) the Terminal is running a separate session from the session "Do Shell Script" invokes. a) would be wrong, in that it is the same bash program run by both, and b) would be correct in that the are running unrelated shell sessions.

You appear to want the ability to have co-processes where the process running the Applescript can interact with the process running the shell script. While this is possible, co-processes are more difficult to setup and control.

You approach was to have a Terminal session running one shell process and Applescript asking the Terminal session to read its display buffer.

My "hyjacking" was to correct the incorrect statements made by others so that other people who are trying to solve something and find this thread find the answer as well as know some of the information posted on this thread was dead wrong.

But you assume Hi] did not like his answer, and there is no evidence this is the case. You are trying to impose you view of this thread on Hi] and the users that posted answers that Hi] seemed to like.

If you have started your own thread, you could have defined the discussion by stating your needs, and not trying to alter Hi]'s request.

I am not looking for points, payment, kudos or political he-said-she-said.

Who said anything about you getting points. Then again, if Hi] wants to give you points, he is welcome, as he started the thread, and he is the one that gives out the points for this thread.

Any questions you asked (like how is it possible to get a non-zero status returned), you cannot give LeonieDF any credit for providing a solution. Only Hi] can do that. If you have started your own thread, you could have described your needs, and any useful information you received, you had the option to give that people giving those replies points (the only payment they get for donating their time and knowledge to this forum).

That is hijacking a thread.

Just correcting people that post pointless code such as the one you quoted me commenting on as being useless because it will always return "0" no matter what. Right?

Unfortunately you were not correcting the post.

BobHarris,

At this point you are nothing more than a hostile troll attempting to defend yourself because you are wrong.

Why are you pointing out that I acknowledged something I agreed is possible unless to build a long hostile debate to try to repair your sad ego?

You are connect that each and every "do shell script" starts a new shell session. But you are wrong in that multiple commands issued in the same shell session start unrelated shells.

I am correct? fine, I don't care, I can admit when I am wrong, how about you? Or are all your posts an effort to defend your ego?

I NEVER said that "multiple commands issued in the same shell session start unrelated shells".

Where did you even get that fabricated lie? That is a retorical question because I already know I never said that.

What I did say was that each seporate "do shell script" starts a seporate shell session unrelated to any other session before it. If you use multiple commands on the same line seporated with ; then of course they are in the same shell session. In fact you are mimicing what I already said, because I already said that while explaining the limitations of "do shell script"

Look I am not going to read or respond to the rest of your comment because you are a troll.

I am not "hyjacking".

I am not "taking control". If I took control of anything, I would have deleted and banned you for trolling, sir.

My giving a comment does not give me "power", but apparently you think it does which is why it is so important to you.

I am just correcting the facts with what does, and what does NOT work.

I could pick appart your responce and find more lies in it, but the conversation has degrated into you bickering and nobody cares and nobody will read this non technical babble or yours. The thread is dead now. You killed it. Nobody wants to read you fussing about how right you are and wrong I am. Get over it.

Not only is that THE solution... it is simple easy, and you mislead people by saying "there's no easy way to capture the output

Well, not only does your code not come close to the OPs original question (it doesn't - try it some time), but it's also a FAR cry from 'simple'.

Sure, some people might you consider your (flawed) 25-line script to be simple, but the nature of the question from the original post clearly implies the OP is a novice (no offense, hi]).

And if you think your (flawed) 25-line script is 'simple' compared to a one-liner:

set myresult to do shell script '/blah/blah -blah"

then you and I have different perceptions of the word 'simple'.

Camelot,

My code is off topic and not even close?

Interesting because the OP did ask "terminal output status and store it into a variable once a command has been executed?" And I am the ONLY person to provide the method to get "Terminal output... into a variable"

Terminal is an application in the Utilities folder.

Shell is not something that "implies the original poster is a novice"

You are attacking the wrong reply of mine. It is not 25 lines, it is 1 line.

I just found that you can use "set shellOut to contents of window 1" inside your tell application "Terminal" script to get the data you seek!

aka

set shellOut to contents of window 1

Sure you have to put that inside

Tell application "Terminal"

do script "command;command;" in window 1

set shellOut to contents of window 1

end tell

That is one line not 25 and it is in Terminal, not AppleScript faux Shell.

Last I checked he asked for Terminal output, not AS shell output.

Next time you want to attack someone, get your information right.

You are just saying that because you are dead wrong.

As I said before you mislead people when you write

Camelot wrote:

then there's no easy way to capture the output since AppleScript has no knowledge of the shell process, or when it completes

No easy way? Try one line. set shellOut to contents of window 1

No knowledge of when it completes?

if thisTab is not busy then

Applescript has a wonderfull feature called the Dictionairy. Read it sometime and you will see it does know when the shell process completes.

Stop trying to pretend you arnt DEAD WRONG.

I found this because I wanted to google the same thing as the OP and you told me it was impossible.

Your wrong.

So I provided the answer and corrected you for posting lies.

Now you throw a hissy fit and say I am off topic because I told you AS knows when the shell is completed, as well as how to get the Terminal output in a single line.

That is the OPs question, and I provided it for everyone else that finds this via google.