I need to detect if a batch file has been executed from the local shell or by double-clicking within P4V. Is there a way I could check if P4V is open?
At first I thought to do it in this way,
but as always, TIMTOWTDI. Any thoughts?
Short answer: no, not really.
Long answer:
You can try to fake it, as you suggest, by checking for a running p4v.exe process. This will only tell you if P4V is open, though. If you had P4V open in another window, and you ran your batch file from a shell (e.g. double clicking in Explorer, or running it from a cmd prompt), you'll get a false positive.
Another approach you might take would be determining whether the script is within a Perforce client workspace. That could potentially be done by having the script run a "p4 where" on its own path to see if it's within the workspace -- but it depends on the script's executing environment having correct connection settings, which depends on the client machine being configured "nicely" (e.g. with P4CONFIG files and/or "p4 set", which are accessible to all Perforce client apps -- if you use P4V exclusively its connection settings won't be readily accessible to your script).
Related
I have some trouble keeping alive a background process when launched by TFS.
Usually I use a batch that launch a java server (new window), as long as I keep this window open it works properly.
C:\Users\TFSService\mbs-iot-sdk\osgi\bin\vms\jdk\server.bat
In order to make my process automatic, I include this in TFS. In the step I call a batch that contains the following:
cd C:\Users\TFSService\mbs-iot-sdk\osgi\bin\vms\jdk // necessary to find the batch
start C:\Users\TFSService\mbs-iot-sdk\osgi\bin\vms\jdk\server.bat
In my task manager, I can see in background tasks that java is launched (no new window is opened), exactly as it behaves when launching directly the batch. But after a few seconds, when TFS switches to the next step, it stops.
Then the next step carries on but fails as it requires the server to be launched.
Is there a particular way of doing it in TFS ?
thank you
Alexandre
It's suggest to launch the .bat file from a relative path not directly use cd to hard code the path.
Also recommend you to use Run Batch File task not Run Command Line task to launch the .bat file.
According to your description, seems you are using a run command line task in your build pipeline. Then run the command under the working directory c:\Build_work\5\s, the command cd to C:\Users\TFSService\mbs-iot-sdk\osgi\bin\vms\jdk\ on the build agent, find the server.bat, run the server.bat.
First check if the .bat file is located at the path you are specifying on the build agent. Not sure if the bat file have to run under C:\Users\TFSService\mbs-iot-sdk\osgi\bin\vms\jdk\, guess you are also hard code the path in your server.bat file. Suggest you change all the path to relative path, you could use some built-in variable in TFS.
As for your workaround in comment, seems you want to chain builds in TFS. The official docs literally say "not yet" and have a uservoice in planed. However you could use some workaround, such as create or use other's customize extension (use rest api) to call another build. Detail ways please refer huserben's answer in this question: How to chain builds in TFS 2015?
Note sure you have to go deep into this area for your original issue. Just add some related info in case you are interested or need.
Well,
Just in case someone else goes through the same kind of issue, I found a workaround:
I wish to mix different command line steps, some of them launching Python scripts:
I have one step for launching the server that is required for my testing tool, one step for my testing tool and one Python step for differential testing
I realized that I could embed everything in a Python script.
It can handle server launching process in a separate window (with subprocess), launch my Python part and launch another process for my validation tool.
I have to test the whole chain but, at least, I solved my problem of launching a background process and detach it from TFS
I have Teamcity currently setup to run a batch file, this batch file executes a fortify scan. It seems the environment variable 'PATH' had loaded correctly on one attempt and started to perform the scan. On the next build attempt the batch file couldn't locate one of the executable(sourceanalyzer.exe). When looking into the user defined parameters it seems different environment variables were loaded than the previous build attempt. It looks like the system environment variables load on the successful attempt and then a users environment variable loaded on the failed attempt. Is there a way to permanently set the environment variable PATH in the agent to load only the system environment variables?
UPDATE:
I have tried several things, to include passing in a Fortify environment variable, which does allow TeamCity to start running the scan. However, it looks like I hit another snag with Fortify's plugin for vs2015. The new error states it cannot find the plugin. I assume this is because pathways are hard coded? Seeing as TeamCity doesn't use the System environment variables and I have to pass them in for TeamCity to find these directories. Is there an easier way to use the batch file to load the system environment variables to avoid hard coding pathways? Would setlocal in the batch file help load these system environment variables, so I can just call sourceanalyzer with out creating environment variables or hard coding pathways?
IIRC, Teamcity will ask you if you want to install the build agent on the System account or the user account. By default it selects the System account and as long as you aren't running any GUI apps, you won't notice the difference... until something like this happens. If Fortify is GUI based, then reinstall your build agent on the user account and ignore the following. Otherwise...
When you set your PATH variables using the System(Control Panel) advanced settings, there are two panes, one for user and one for system. Here you can inspect the System variables to make sure they are correct.
What I will generally do is create a new key, say FORTIFY_PATHand prepend %FORTIFY_PATH% to the System PATH variable. THEN RESTART YOUR MACHINE. The path won't get updated correctly until you do.
Next, login to the system account using PSExec: https://superuser.com/a/596395 and try to run your tools from that command prompt to verify that they are working in the build agent's environment. I once had trouble getting an SVN script to upload until I logged into the System account and provided my SVN password. Some settings are stored in %APPDATA% which is different from the user account.
If you can't get Fortify to run from the System command prompt, then you should probably reinstall your build agent to your user account. Or install Fortify to the System account (if possible).
When configuring TeamCity build agents, check the agent system and environment variables By going to Agents->Agent->Agent Parameters, or the /agentDetails.html?id=1&tab=agentParameters&kind=envpath on your server.
After changing the parameters, restart the agent or restart the agent's machine.
How do i do this?
i've tried all i can think of.
Browsing for my .exe file i want to run.
run a bat file
writing forcedos.exe in program textbox and path to my bat file in command line parameters textbox
Why cant it just work with an normal exe?
and i know it should run a program because it can restart the service correctly.
EDIT
Application: test.exe
Framework Version: v4.0.30319
Description: The process was terminated due to an unhandled exception.
Exception Info: System.Exception
This is my latest try to make it work.
the script has the code
Dim objShell
Set objShell = WScript.CreateObject( "WScript.Shell" )
objShell.Run("iexplore")
Set objShell = Nothing
Now the script opens internet explorer if i run the cmd command
wscript.exe "C:\asfh.vbs"
The "run a program" recovery option for when a service crashes runs the specified executable in the same way as the service, i.e., it runs in session 0 (and so is affected by session 0 isolation, see also related questions) and it runs with the same security context as the service.
This means that it can't interact directly with the user (you can display a GUI, but nobody will see it) but it also restricts what the executable or script can do. For example, some shell API functions will not work properly unless the user account has been interactively logged in at some point. In the example script you posted, the script itself is probably running, but is unable to launch Internet Explorer because IE is only designed to run in an interactive session.
Provided you restrict yourself to basic functionality, it should all work as expected. (There is no master list that I know of describing what functionality is safe to use in a service context, but it is usually easy to guess. You can resort to trial and error if necessary!)
Also note that as far as I know forcedos.exe is no longer present in modern versions of Windows. If you want to run a batch file, you can specify cmd.exe as the application and /c myscript.bat as the command line parameters.
Create a new file test.bat, the content is as follows:
echo aa;
and execute it in the dos command line,
test.bat
there are no output in the window. But when I tested it in another computer, it outputs
aa
When I double click the bat file, it always disappeared suddenly even I add a "pause" clause line in the test.bat file. Both the systems are Window XP. Can sb tell me where is the problem, please?
I have checked all the things Mofi suggested, but the problem remains.
Now, I divscover when reboot to the system using another Administrator account. All is fine.
So, whether the problem comes from the registry record of the Windows XP?
See my answer on Iterating over files in directory prepends extra character to filename where I have already explained what to check if batch files are not working as expected.
The environment variables ComSpec, PATHEXT and PATH should be verified by you first. And check also the values of registry key Command Processor as explained in referenced answer. Also Autorun value of key Command Processor is very critical.
Further, start Regedit and check standard value of registry key HKEY_CLASSES_ROOT\.bat displayed on right side. It should be batfile. Then go to HKEY_CLASSES_ROOT\batfile, open it and navigate to subkey open - command. The standard value is "%1" %*. It could be that your Windows registry is partly corrupt which makes it not possible to run a batch file by double click.
A check for a problem with registry would be opening a command prompt window by running
C:\Windows\System32\cmd.exe
via Run entry in start menu of Windows. Then run your batch file from within the command prompt window.
You should also run from a command prompt window
C:\Windows\System32\chkdsk.exe C: /F
and confirm running the check disk tool of Windows on next boot of Windows. Then restart Windows to let this tool check the file system for errors resulting in 1 or more corrupt files. If an error was really detected in file system, run after Windows finished starting from a command prompt window
C:\Windows\System32\chkdsk.exe C: /F /R
Confirm running this tool once again on next boot of Windows, but this time with read/write tests on all sectors of your hard disk and repairing sectors if that is possible at all. It can take several hours depending on size of drive C: to finish this task.
Added on 2014-09-28:
Anti-virus application
Do you have checked already the configuration of the anti-virus application?
It could be that it blocks the execution of all *.bat files.
Windows System Restore
Do you have already tried to use Windows System Restore for the account on which batch file execution fails and go back to a restore point of a date on which you are sure that batch file execution worked?
If ntuser.dat or ntuser.dat.log (= registry of current user) is partly corrupt or something is misconfigured in Windows registry for this user account, this would be the easiest method to get back a working environment for the account.
Process Monitor Analysis
As execution of batch files does not work only with one account, the reason must be a corrupt or wrong set key or value in registry of this account.
I suggest to download and run free tool Process Monitor from Sysinternals (Microsoft) using first the account on which batch file execution does not work as expected.
The first displayed on start of this tool is a dialog to configure a filter. Select from first list Process Name, let is selected on second list, enter cmd.exe in edit field, and let Include selected in last list. Click next on button Include and then on OK.
Now in main application window of Process Monitor click on last 3 symbols in toolbar to toggle off those features resulting in monitoring only registry and file system accesses. Clear current list by clicking on fifth symbol in toolbar or by pressing Ctrl+X.
Now double click on your sample batch file in Windows Explorer. Then switch batck to Process Monitor, scroll up to beginning and look line by line what happens on executing of a batch file. Perhaps you can see something which you think is not correct.
The column Result should show only SUCCESS and NAME NOT FOUND and rarely NO MORE ENTRIES for registry entries in log. INVALID in column Result is not good. The log can be saved as CSV file with Ctrl+S or by clicking on disk symbol in toolbar.
Exit Process Monitor, logout from this account, logon on other account, start here also Process Monitor, configure the same filter, execute same batch file and save also this log into a CSV file. Compare the 2 CSV files and look for differences (other than time and line number).
Windows Registry Compare
Another method to find out what is different between the 2 accounts resulting in batch file execution not working is comparing the registry hives of those 2 accounts. In both accounts start Regedit and export entire key HKEY_CURRENT_USER to a *.reg file. Compare the two registry exports with a text comparison tool. The new administrator account has definitely less keys and values. Interesting are those blocks with just similar, but not identical lines.
Most interesting are the keys:
HKEY_CURRENT_USER\Console
HKEY_CURRENT_USER\Environment
HKEY_CURRENT_USER\Software\Classes\.bat
HKEY_CURRENT_USER\Software\Classes\batfile
HKEY_CURRENT_USER\Software\Microsoft\Command Processor
Some keys from this list can be missing completely. That would be okay as in this case the same key under HKEY_LOCAL_MACHINE is taken into account by Windows. But when a key is missing in account on which batch file execution fails, it should be also missing in the other account, or something is wrong here.
What could be a big and hard to find problem is a registry key which is displayed on viewing the registry tree with Regedit, but when clicking on the key it cannot be read.
I have seen such partly corrupt registry keys already several times and it is very hard to fix such a registry error. This happens for example if an application removes a registry key without removing it also from index table of the registry.
Totally unexpected things happen if a registry key is present in index, but does not exist anymore in registry. For example I fixed once an issue where HKLM\Software\Classes\mailto was in index list, but did not exist anymore in registry. When the user clicked on a mailto link, Windows started creating new emails in Outlook in a loop until Windows run out of resources and user restarted Windows.
Please add # before echo, like this:
#echo aa
#pause
Its working in my Windows7 machine.
I am writing a terminal-based application, but I want the user to be able to edit certain text data in a separate editor. For example, if the user chooses to edit the list of current usernames, the list should open as a text file in the user's favorite editor (vim, gedit, etc.). This will probably be an environment variable such as $MYAPPEDITOR. This is similar to the way commit messages work in svn.
Is the best way to do this to create a temporary file in /tmp, and read it in when the editor process is terminated? Or is there a better way to approach this problem?
There's already a $EDITOR variable, which is extremely standard and I have seen it working on a wide variety of unixes. Also, vi is always an option on any flavor of unix.
Debian has a sensible-editor command that invokes $EDITOR if it can, or falls back to some standard ones otherwise. Freedesktop.org has an xdg-open command that will detect which desktop environment is running and open the file with the associated application. As far as I know, sensible-editor doesn't exist on other distributions, and of course xdg-open will fail in a text-only environment, but it couldn't hurt to try as many options as possible, if you think it's important that a desktop user can see their happy shiny gedit or kate instead of scary old vi or nano. ;)
The way crontab and sudoedit work is also by making a file in /tmp. git puts it under .git, and svn actually puts it in the current directory (not /tmp).
The way svn and mercurial do it is by making a file in /tmp.
BTW, you don't need a MYAPPEDITOR, on nix there's EDITOR already present.
Since you mention svn in your post, why not just follow the same methodology? svn opens a file with a particular name with whatever $EDITOR (or $SVN_EDITOR) contains - this might actually require some work on your part; determining the parameters to each supported editor. In either case, you have the name of the file that was saved (or the error code of the application if something failed) and you can just use that.