In all of my searching, I have not come across any trick, plug-in, or setting that would allow me to pre-populate a Watch Window for use during debugging.
The well-known steps to watch a variable in Visual Studio 2017 is to set a breakpoint (perhaps on the first instance of a variable being assigned a value), then, adding the variable to a Watch Window. On subsequent runs, that watched variable should remain in the Watch Window.
Is there a way to accomplish pre-populating the Watch Window before a debugging session? I have code that runs in a timed sequence. Taking the extended time during an execution break to pause and set up a watch causes the program to crash. Such timeout crashes while setting up variable watch objects makes for difficult interaction with the debugger. Pre-populating the watch list would help considerably.
I know that I could use a technique such as using Debug.Print(...) statements, which are printed to the Output Window during code execution. However, this doesn't allow me the control and visual feedback to my debugging efforts that comes from a Watch Window.
Watches can only be evaluated when you are paused in the debugger, say on a breakpoint. They can't be evaluated during normal execution. Your options to get around this limiation are:
Add Debug.Print(...) or something similar to your code that outputs a value.
Similar to #1 add a TracePoint which you'll find in the actions setting of a breakpoints setting (that's the gear icon when you hover over a breakpoint). When the TracePoints are hit they will then evaluate the expression that you specficied in the actions window and log it to the Output window. The advantage of TracePoints over adding your own logging is that you can turn it off and on without building your code. More info at: https://learn.microsoft.com/en-us/visualstudio/debugger/using-breakpoints?view=vs-2019#BKMK_Print_to_the_Output_window_with_tracepoints
Use the VS Enterprise only feature Snapshots which is part of IntelliTrace. Basically set your breakpoints but rather than stopping on them to look at the watch just continue. Each time you stop a snapshot of the process will be taken capturing the state of your application at that time. Then once you're finished use the Diagnostic Tools window to select each of the snasphots and activate them. For each snapshot you can use the debugger just as if you stopped the application. So you can use watches and inspect etc. Of course you can't step as the app has already ran but you can go to the next snapshot etc. More info at: https://learn.microsoft.com/en-us/visualstudio/debugger/view-snapshots-with-intellitrace?view=vs-2019 and https://devblogs.microsoft.com/visualstudio/step-back-while-debugging-with-intellitrace/
How Can Watch Variables be Configured in a Watch Window Before Running
a Debug Session in VS2017?
I'm afraid the answer is negative.For now,the Watch window can only be configured after the debug session start.It's like a runtime window only occurs during debugging.
So we can't pre-populate it before debug session for now. In other words, it's by design.
As alternative ways,you can follow Andy's detailed suggestions above.
And since your needs is meaningful in some specific debugging situation, you can also post your suggestion on development community like a user voice to add a new feature.
Related
I'm not talking about Disable All Breakpoints or Start without Debugging.
For example, I have my breakpoints set, including some in loops. I get to a point where I just want the code to run through, without having to disable or toggle off each breakpoint. I want the breakpoints to stay enabled for my next run, but I just want the current execution to complete.
If not, definitely nice to have.
You could use Debug -> Detach All which detaches the debugger but leaves the process running.
When i am debugging my program and an error occur, the debug session ends, but the program remains in memory. Using the activity manager of Windows to close it does not work. I need to close Visual Studio in order to kill the process.
Why is this happening?
When during debugging a program error occurs, the program usually does not "end". Instead, the debugger (VS2010) pauses execution, allowing you to inspect the code resulting in the error. Depending on the language used (e.g. C#) and the way you compiled your program, you may even be able to edit the program on the fly, move the execution cursor back a bit and continue the program from there.
If the Debug toolbar is visible (in my case it shows up automatically whenever I'm debugging), you should see a couple of "playback" buttons, allowing you to start/continue, pause, stop your program etc. If you stop your program, it will be gone from the task manager too.
As I mentioned in a comment on your question, you can also use the Debug menu to accomplish these tasks.
I have an app built in node.js and I use the node inspector in order to debug.
But it's quite hard because of this:
My breakpoints are never saved after I restart the server
I cannot put a breakpoint on a file that has not loaded yet; so I have to step into from the first script to the one I want; REALLY PAINFULL!
How do you really debug node.js with node inspector?
The videos on how to use node.js are quite misleading as everything is into a module...
http://www.youtube.com/watch?v=AOnK3NVnxL8
or this one the scripts appear are already loaded in the first script
http://www.youtube.com/watch?v=HJOH0-g8f6E&feature=mfu_in_order&list=UL
Edit:
Nobody can answer this question? :s
In javascript you can set breakpoints using the debugger; statement. However, they will only pause node if a debugger is actually attached.
So launch your node script using
node --debug-brk myfile.js
then launch node-inspector and press the play button to continue to the next breakpoint and it will hit your debugger; breakpoint (at least that works for me ATM)
(as noted in the comments: in recent versions of node you no longer have to separately install node-inspector. If you launch node using node --debug-brk --inspect myfile.js you get a url that launches the debugger in your browser).
you still need one extra click after restarting, but at least your breakpoints are saved.
if your breakpoint is not hit automatically, but only after some user action you don't need the --debug-brk of course.
The problem with client-side breakpoints is that it's hard to keep track of the breakpoint position when the file changes. Unlike in an editor, it cannot keep track of lines being changed, etc.
#RyanOlds suggestion of using debugger; statements is also a good one, but you have to make sure the debugger is connected before the statement is evaluated, because it is ignored otherwise. Starting with --debug-brk is a good way to force this, because the execution is paused on the first line allowing you to attach the debugger and then continue the execution.
You could try debugging with node's internal debugger.
Edit: However, according to the v8 DebuggerProtocol it's possible to set breakpoints on script that hasn't been loaded yet AND you can set breakpoints by function, script and more. It should therefore be possible for node-inspector to keep track of your breakpoints (in a session, or whatever). It doesn't do so right now, though.
Maybe if v8 allows a certain piece of code to trigger a breakpoint, similar to nodes debugger?
Edit: It does, you should be able to trigger a break by throwing any old exception (caught or uncaught).
The new version (0.3.x) of node inspector saves breakpoints in browser's local storage and restores them automatically.
https://github.com/node-inspector/node-inspector/pull/116
Try using IntelliJ WebStorm - there's a free trial and licenses aren't outrageously expensive. It lets you save breakpoints in all your files prior to starting up its own internal node process and remembers them across process restarts.
I agree - node-inspector looks brilliant, but is quite useless unless your app has a clear place to set a breakpoint in the top level script just after your source files have loaded, but before you hit the area you want to debug. You can structure your own code this way, but you won't be so lucky with other helpful libraries you want to include. Also... why should a debugging tool dictate your project structure!
Forgetting breakpoints is extremely unhelpful... most of my debug runs take more than one walkthrough, as in other people's code it's easy to step past where you want to be.
You can use node-codein for inspection. It won't do runtime breakpoints but it should ease the inspection process.
https://github.com/ketamynx/node-codein/
Also worth noting.. vscode has a great debugger for node.
https://code.visualstudio.com/
Available on Mac, Linux, & Windows.
It does runtime breakpoints (without the need of writing debugger; statements),
supports variable watches, and even has a call stack window (very nice).
Everything is so automated, it is now my goto over sublime text when using nodejs (and I LOVE sublime).
This is built in now including saving breakpoints. I just tested it in node 7.3.0.
node --inspect --debug-brk app.js
This prints a url like this
To start debugging, open the following URL in Chrome:
chrome-devtools://devtools/bundled/inspector.html?experiments=true&v8only=true&ws=127.0.0.1:9229/c3d5d93e-9d27-41b9-a4da-607e43c9d4f8
Put that in Chrome and you're good to go.
If you want to skip copy/pasting the url, do this:
npm install -g inspect-process
inspect --debug-brk app.js
Unfortunately the inspect-process method doesn't retain the breakpoints :-(.
Here's a video I made: https://youtu.be/rtZKUnks6jI
I like how you can capture breakpoints as Intellitrace events. Breakpoints are easy to add/enable/disable, and can be added/removed at runtime (as opposed to say Console events).
However, of course breakpoints stop the flow of execution, and I have to press F5 F5 F5 F5 F5 F5 constantly while I debug the running application. I'd love to be able to capture breakpoint events in intellitrace and just let the code execute without actually breaking on the breakpoints.
I've also used Console statements as captured intellitrace events, but breakpoints are much easier to add (F9 as opposed to pasting a Console.WriteLine("...") everywhere), and I can add and disable new breakpoints at runtime, whereas if I want to add or remove a Console based event at runtime, I have to stop the program, add/delete, and restart.
Are there any good ways to capture intellitrace events in this way?
I think this is what you are looking for:
http://geekswithblogs.net/sdorman/archive/2009/02/14/visual-studio-2008-debugging-tricks-ndash-advanced-breakpoints.aspx
Basically, you can use an advanced breakpoint to do a number of different things. In your case, you would probably want to use the tracepoint that simply gives a hit count. In other cases, you might want to specify that the breakpoint only occurs after a certain number of hits (maybe you want to be sure a loop isn't infinite).
Visual Studio has a ton of great features inside of the breakpoint option. I'm sure that even if the above option didn't solve your problem that there is still something that is included in the above article that will solve your problem.
Normally in Visual Studio, a watch cannot be evaluated unless the debugger is stopped at a breakpoint. Is there a trick or add-on to make Visual Studio evaluate a watch while the application is still running? For example, evaluate the watch every time execution passes a point in the code while it's still running and without changing the code to insert statements like Debug.WriteLine.
Not sure this is possible, but I thought I'd ask.
Yes, this is possible. Set a breakpoint at the location where you'd want to see the value. Right-click the breakpoint and choose "When Hit...". Tick "Print a message" and write an expression like { value }. The message is displayed in the Output window while your program runs.
I would do that using compiler directives.
#if DEBUG
Debug.WriteLine
#end if
No this is not possible to do. The evaluation feature in Visual Studio is a stack frame based mechanism. That is that every evaluation is done in the context of a given stack frame (as viewed through the stack window). When the program is running the set of stack frames is currently changing and hence it's not possible to do a stable evaluation.
Additionally there are other limitations in the CLR which prevent this from managed code. It's not possible for instance to execute a function unless the debugee process is in a very specific state.