Applescript in Automator behaves differently when executed from inside Automator, or as an Automator application

Hello everyone,
I'm trying to execute an Automator workflow at startup, which I do by saving the workflow as an application, and adding it to the startup list, this works fine.
My workflow works fine when executed from within Automator (ie clicking on the "Execute" button), but does not anymore when the Automator application is executed (ie. simply double clicking on my Automator file)
I have narrowed it down to an Applescript inside my workflow which fails only when the application is executed from the finder (the step works fine from within Automator). The script automatically creates an ad hoc wifi network.
The line which specifically fails is :
set menu_extras to value of attribute "AXDescription" of menu bar items
Anyone has any idea why executing the workflow from within automator, and executing it from the finder yield different results ?
The crazy thing is that all this worked fine up until a few days ago, there might have been an update since (running Mavericks 10.9.2), but nothing game-changing.

You will need to provide more than the one line of AppleScript you quote if you want a definitive answer. (We can't even see which app you are addressing this to.)

Similar Messages

Java application behaving differently when run from netbeans

Hi,
When i run my program from netbeans it receives packets from ip addresses which are not in my subnet
but when i run it from outside netbeans it doesn't receives packets from ip addresses which are not in my subnet.
Can anybody explains why my application is showing different behavior when run from netbeans and when run independently?
Thanks.

Bingo! My script started the rmiregistry and continued on to start the servers. Apparently, it was too quick. I put a 10 second delay after the rmiregistry startup and all is working as it should.
Thanks for your help. You get the Duke dollars if I can figure out how that works.

Executable behaving different when called f/Java

I have an executable that is run from the command line which behaves differently when I run it from my application. The executable takes MP3s files as arguments and merges them together. When I use large files from within my app, it doesn't respond, but using the same files from the command line works fine. I was thinking it might be a memory thing, but the executable should get its own space seperate from the VM. Does anyone have any suggestions? Thanks!

I do not have the source for the executable, so I wouldn't be able to pull lines out of there. It works from my Java app when I pass in smaller files, so I know I have the syntax correct, but with larger files it hangs. When I run the same thing from the command line with the larger files it runs like a champ. I was trying to use my task manager (I'm on Win NT) to look at memory usage since the only variable seems to be file size, but nothing seems to indicate it is a memory issue.

Why does "Link Media" behave differently when used as a keyboard shortcut vs when right clicked???

Not sure whether anyone else is encountering this, but the "Link Media" command behaves differently when right clicked as opposed to when used as a keyboard shortcut.
When you right click to select that option for an offline file, you get only the corresponding disconnected source media. But when you execute "Link Media" via a keyboard shortcut, PP brings up a huge list of all other offline media. Why does it only do that when you use the keyboard shortcut for "Link Media" but not when you right click "Link Media".
I'd like to get the Link Media command to behave via the keyboard shortcut the same way it behaves when right clicked. I don't want to have to deal with relinking a hundred other files, but just the single offline file I'm working on.

Among the alternatives not mentioned... Using a TiVo DVR, rather than the X1; a Roamio Plus or Pro would solve both the concern over the quality of the DVR, as well as providing the MoCA bridge capability the poster so desperately wanted the X1 DVR to provide. (Although the TiVo's support only MoCA 1.1.) Just get a third-party MoCA adapter for the distant location. Why the hang-up on having a device provided by Comcast? This seems especially ironic given the opinions expressed regarding payments over time to Comcast. If a MoCA 2.0 bridge was the requirement, they don't exist outside providers. So couldn't the poster have simply requested a replacement XB3 from the local office and configured it down to only providing MoCA bridging -- and perhaps as a wireless access point? Comcast would bill him the monthly rate for the extra device, but such is the state of MoCA 2.0. Much of the OP sounds like frustration over devices providing capabilities the poster *thinks* they should have.

Event flow in ABAP is different when executed via SE80 vs. via TCODE

Hi,
We have many existing custom ABAP report pgms that have a custom transaction code that executes them. Regardless of how the pgms are invoked, the event sequence in the pgm is INITIALIZATION, AT SELECTION-SCREEN, START-OF-SELECTION. etc.
Now, when I create a new pgm with a transaction code to invoke it, the event sequence when executing the pgm via SE38 is as before (All 3 events above are triggered). However when I invoke the pgm via the custom transaction code, only the event At SELECTION-SCREEN is executed. (Not the INITIALIZATION or the START-OF-SELECTION events).
We have upgraded our SAP GUI from 6.4 to 7.1.   We are upgrading our system from 46.C to ECC6. I get this error in ECC6 and also in the old system 46.C which is causing me to question whether the SAP Gui is involved. We are using SAP GUI 7.1 in both the 46.C and ECC6 systems.
Any help will be appreciated!
Thanks,
Deb

Hi All,
Here is my ABAP Code. I have put break points in the code so I can track the sequence.
REPORT ZDEBNEW .
data zdeb(2) type c.
tables kna1.
SELECTION-SCREEN BEGIN OF BLOCK A WITH FRAME TITLE TEXT-002.
SELECT-OPTIONS: I_KTOKD for KNA1-KTOKD.
SELECTION-SCREEN END OF BLOCK A.
INITIALIZATION.
zdeb = '1'.   "break
sy-subrc = sy-subrc.
PAI
AT SELECTION-SCREEN.
zdeb = '2'. "break
sy-subrc = sy-subrc.
START-OF-SELECTION.
zdeb = '3'. "break
sy-subrc = sy-subrc.
END-OF-SELECTION.
zdeb = '4'.   "break
sy-subrc = sy-subrc.
As for the type of transaction I am using...I created it via SE93 and specify SELECTION SCREEN 1000. I classify it as a Professional User transaction. One other thing I have noticed is that when I view the new pgm via SE80 I do not see Screen 1000 in the left navigation panel. If I try to create such a screen in SE80, it says it already exists. I can see Screen 1000 from tran SE51. In one of the existing old pgms, I do see screen 1000 when I view SE80.
Thanks for the help and interest!!!

Code behaves differently when run in debugger

var he:Array = handControl.getHumanInfo(ha, handNumber);
while (he.length != 0)
humanInfo.push(he.shift());
This snippet is part of a large flash project I am working
on. The program works fine by itself, but when I try to step
through it in the debugger, it sometimes crashes, depending on
where I put a breakpoint. If I put the breakpoint on the first line
of code above, I can step into the function getHumanInfo(), trace
it, and step back out, and everything is fine. But if I only put a
breakpoint somewhere inside the getHumanInfo() function, all of a
sudden when I step out of that function, the variable 'he' is
undefined, and so the while loop never ends.
This makes no sense to me. The code is either good or bad, it
shouldn't behave differently based on where I put the breakpoint.
The getHumanInfo() function always creates a new array and
passes it back, so there is no reason 'he' should be undefined.
This has happened many times when I've tried to use the flash
CS3 debugger -- is there a known issue that this sounds like? I'm
so fed up I'm ready to stop using the debugger altogether.

I ran the program without the debugger and put some trace()
statements in, just to make sure. Everything works fine without the
debugger. So it seems to me it has to be a debugger issue. Has
anyone heard of a similar issue?

Old workflow behaving differently in upgrade from 4.6 C to ECC 6.0

We are facing an issue in the workflow . Old and existing workflows before the upgrade have different paths then the new workflow in the ECC6.
Means , the same workflow when started from scratch is working fine , but when the old workflows are copied and run (the way which client is testing or you can say this way they do their work), then the workflow takes a different path. All the forks ,events,if clause are appearing as the last steps in the workflow grapic.
Now my point is, that in upgrade there should not be any major change happened in SAP, seems that we may be missing some step to be performed after upgrade or during upgrade forits standrad as well as custom worflows.

Hello,
I've never done the upgrade myself, but judging from what I've read it's very very important that you try to complete as many workflow instances before the upgrade. That doesn't help you much, but maybe others will take note.
All you can do is search for your particular problem. Are there any error messages in the workflows?
regards
Rick Bakker
Hanabi Technology

Expdp problem when executing from pl/sql transportable tablespaces

Hello
i have a very strange problem for which your assistance is really appreciated. i am applying the transportable tablespaces approach and i have a problem when executing expdp from commandline in comparison to the procedure generated to make the exact same thing.
more precisely when running:
expdp etl/etl dumpfile=TBSP090609_DMP.dmp logfile=TBSP090609_DMP.log directory=EXPDP transport_tablespaces=TBSP090609 transport_full_check=y
the dmp file produced is over 200MB.
when runng the below pl/sql code the volume produced is around 50MB. any idea what i am missing?
job_handle :=
dbms_datapump.OPEN (operation => 'EXPORT',
job_mode => 'TRANSPORTABLE',
job_name => 'export_tablespace',
VERSION => 'COMPATIBLE'
dbms_datapump.add_file (handle => job_handle,
filename => dump_file_name,
DIRECTORY => export_dir_name
dbms_datapump.set_parameter (handle => job_handle,
NAME => 'TTS_FULL_CHECK',
VALUE => 1
dbms_datapump.metadata_filter (handle => job_handle,
NAME => 'TABLESPACE_LIST',
VALUE => '('''
|| tablespace_name
|| ''')'
dbms_datapump.start_job (job_handle);

thanx you for your response. the problem was eventually managed to be resolved as follows:
job_handle :=
dbms_datapump.OPEN (operation => 'EXPORT',
job_mode => 'TRANSPORTABLE',
job_name => 'export_tablespace',
VERSION => 'COMPATIBLE'
dbms_datapump.add_file (handle => job_handle,
filename => dump_file_name,
DIRECTORY => export_dir_name
dbms_datapump.set_parameter (handle => job_handle,
NAME => 'TTS_FULL_CHECK',
VALUE => 1
dbms_datapump.metadata_filter (handle => job_handle,
NAME => 'TABLESPACE_EXPR',
VALUE => 'IN('''
|| tablespace_name
|| ''')');
dbms_datapump.start_job (job_handle);

Unix script not working when executed from SAP PI communication channel

Hi Experts,
I am currently using the Run OS command after message processing functionality of the file adapter.
However, the unix script doesn't seem to work when run via communication channel.
The script runs properly when executed manually.
The communication channel logs don't show any error as well.
Do you have any ideas as to where the error might be.
Thanks,
Mike

Hi All,
Thank you for your replies. I appreciate this.
@Anand, here is the script:
#!/usr/bin/sh
#version 1
hostname=$(uname -a | cut -f 2 -d " ")
MAILTO="[email protected] "
homedir=$PWD
echo $homedir | mailx -m -s "OB10 $hostname" $MAILTO
/opt/java6/bin/jar -xvf OB10.zip
ll | mailx -m -s "OB10 $hostname" $MAILTO
What's puzzling me is that I am able to receive the email alerts.
This means that the user has sufficient authorization to execute the script.
But the unzip command is not working.
I have seen the unzip work when run manually.
Thanks,
Mike

Jar works different when executed with webstart

Hello, I have a problem when executing an app with webstart. When I execute the jar directly on windows everything works fine, but when I execute the webstart link it executes but not correctly. The application has some timers that don't work. The rest of the application seems to be ok. Anyone has a clue about this problem? Thanks

Hello, I have a problem when executing an app ...What application? What does it do? Do you control
the source code?
..with webstart. What is the URL the the JNLP? Or failing that,
what is the JNLP content?
Is there any stacktrace in the Java console?
Do you have the console pop-up for web start apps.?
...When I execute the jar directly on windows
everything works fine, but when I execute the
webstart link it executes but not correctly. The
application has some timers that don't work. Timers don't work? Maybe they are just lazy!
Or to put that another way, what is supposed
to happen at the end of the time that does not
happen in the web start app.? Are exception
or error stacktraces printed to the console?
...The rest
of the application seems to be ok. Anyone has a clue
about this problem? Without answers to the questions above,
we are really just guessing.
Message was edited by:
AndrewThompson64

SSO problem when redirecting from a JSP page to an external application

Hi,
I try to make a redirect from a JSP page (that is under a SSO protected application on iAS) to another page from another application, on an external iAS server, also protected by (a different) SSO. After the redirection is done, the login window appears, I enter the login name and the password and after that I obtain the followin error:
"Oracle SSO Failure - Unable to process request
Either the requested URL was not specified in terms of a fully-qualified host name or OHS single sign-on is incorrectly configured.
Please notify your administrator."
In the logs og the server I found the following:
[OSSO] W05: Requested URL is not specified in terms of fully-qualified host name or invalid SSO partner configuration. Host from request
mycompany.com:7777, registered host 144.147.147.200:7778.
(the ip address being the address of the mycompany.com host).
Any clue about this? Thanks a lot in advance!
Regards,
Marinel

Hi Carlo,
Thanks for your answer.
The JSP original page is not added as a partner application to the second SSO server.
The idea was that the user should insert first the login name/passwd for the first server, after being logged in, then redirected to the second application (on a different server), insert the login name/password for the second application and then load the 2nd application page. It seems that is not working after inserting the password for the 2nd application.
Coming to a more general question that could help me to avoid this complicated approach: is it possible to have two different applications deployed on two different iAS servers and the two applications to use the same SSO (let's say the one from the first iAS server)? I have to mention that the process scenario is the following: the user load a page from the first application (protected by SSO), then, after successfull login and some processing in the first app, he will be automatically redirected by the first app to the second application, on the second server. I want to have also the second application, on the 2nd server, protected by SSO (ideally would be the same SSO as the first one!). Ideally the scenario would be: if it is redirected from the first app and the user is already authenticated, the automatic redirection should be done transparently for the user (without enetring the password again). If the user goes directly from the browser to a page of the second app, the SSO login window should be displayed and the user should provide his password.
Is such a scenario possible on two apps deployed on two different servers?!
Thanks a lot again!
Regards,
Marinel

Flex app behaves differently when Adobe automation files loaded

We have an application built with Flex 3.1. I am attempting to create QTP tests for this application. The instrumentation files are loaded at run time, not compiled into the application.
I am using QTP 9.5 with the Adobe Flex Add-in 3.0. I am able to see the controls an run tests to some extent.
The main issue I am trying to resolve is that there is a difference in behavior of the application when it is started with the instrumentation files. The differences occur both inside and outside of the QTP environment.
The observed differences are:
The initial page is a login screen. After entering the id and password, I can click ‘Enter’ or click the “Log in” command button. With the instrumentation files loaded, the ‘Enter’ key no longer works.
When I log out of the application, a progress screen displays while the environment clean-up occurs. When the progress bar hits 100%, the original login screen is displayed. With instrumentation files loaded, activity stops when the progress bar hits 100% and the original login screen is never displayed.
After logging in, a window with perspective tabs displays. Each perspective tab page contains a window containing a list of resource associated with the perspective. When instrumentation files are loaded, the tab pages exist, but the perspective windows are not created.
I realize I am dealing with application-specific behavior here and I am pursuing this with our developers. What I am looking for is any information related to differences in application behavior when instrumentation files are loaded.

I can't seem to reproduce what you are seeing with this sample application:
<s:Application xmlns:fx="http://ns.adobe.com/mxml/2009"
               xmlns:s="library://ns.adobe.com/flex/spark">
    <s:states>
        <s:State name="state1" />
        <s:State name="state2" />
    </s:states>
    <s:controlBarContent>
        <s:Button label="change state" click="currentState=currentState=='state1'?'state2':'state1'" />
    </s:controlBarContent>
    <s:Button x="50" y="50" label.state1="state1" label.state2="state2" skinClass="CustomButtonSkin" />
</s:Application>
Can you provide a similar sample application that demonstrates the issue?

Views always appear as different when engineering from Logical to Relational

I created some views in my logical model. After I engineer them to relational model, they appear in relational with an exclamation mark near the upper left corner. If I try to generate ddl for them, I get an error (-- ERROR: Invalid View VwTest). I realized that I have to "validate selected views", in order for the eclamation mark to disappear, and then I am able to generate ddl for them. After that, when I engineer from logical to relational again, it always shows me that my views are different. The difference is that in logical, my entities are shown between double quotes ("), while in relational, after the validation, the tables are not between double quotes. Is there a way to overcome this?
Thanks

Hi,
Thanks for reporting this problem. It appears that this is happening when the Entity name contains lower-case letters.
David

Error in prepare cursor stmt, when executed from a cgi called from web apln

We are facing with the following error message:
Error Prepare cursor SELECT message FROM OPS$SVEERAMA.ERRORCODE ORDER BY message; error code: 549755810774, offset: 180388626432
549755810774 is the sqlca.sqlcode when printed as long and is -3114 (negative) when printed as int.
We are trying to execute a cgi from the web application, the cgi in turn tries to connect to the oracle database. The cgi script, tries to establish an oracle connection, and tries to execute a prepare cursor statement.
We are using oracle 10g and proc - Pro*C/C++: Release 10.1.0.5.0, on linux x86_64 machine. We have two sqlca.h header files, one which has the struct sqlca's variables defined as long and the other as int. we are using the int one.
Any help is much appreciated. Also if you can let us know what are essential for establishing a db connectivity from a cgi script.(any env variables, etc)

All oracle error messages are negative number.
/home/oracle>oerr ora 3114
03114, 00000, "not connected to ORACLE"
// *Cause:
// *Action:I suspect that you have not successfully established a connection to the database. Unfortunately, I don't actually know what you need for CGI. I would check for other errors in the portion of the code where you establish the connection.
John

Classpath issues when executing from a jar

I have a jar that I built for an RMI server that has a need to access another jar. (log4j). I run the jar from the command line using:
java -jar -Djava.rmi.server.codebase=my.jar my.jar
and the code runs fine but when I try to add in the log4j.jar I can't find it. I have placed it in my.jar in a lib directory within my.jar, I have placed it in the current directory, I have placed it in other directories I know are on the classpath, I have done all kinds of variations with the -classpath switch on the command line and nothing works.
How do I package, configure and/or call such that the Logger class in log4j.jar can be found by my RMI server when I execute my.jar?
Thanks

Running from the jar-file ignores the CLASSPATH. See Bug #4459663
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4459663
You have a few options:
First, you can unjar all of the jars you depend on and rejar 'em up with your jar. This is the simplest solution.
Second, you can skip running "from the jar" and just put that jar in your CLASSPATH.
Third, you can put jars you depend on in the $JAVA_HOME/jre/lib/ext directory.
Fourth, you can modify the manifest to specify a classpath, and then put the libraries you depend on at the places specified.
And, if you think There Ought To Be A Better Way, vote for the RFE!

Applescript in Automator behaves differently when executed from inside Automator, or as an Automator application

Similar Messages

Maybe you are looking for