I'm experimenting with Dart/Angular/WebStorm for the first time. One thing which I've found a little jarring has been the build->error cycle. In Visual Studio, I am used to this work flow:
Write some code
Running a build
Having a fresh list of errors being created
Fixing a subset of them (some or all)
Go to 1.
I'm wondering what is the workflow with Dart?
I have the following issues:
I can't figure out a way to just run pub/transformer/whatever-it-is-that-roughly-equates-to-a-build. The only way I can do this is by attempting to run a configuration
When the transformer is run, it just dumps a gigantic error output to the Pub Serve window. It does not clear the existing output, so I end up with duplicate error or errors I've already fixed. So I'm left manually scrolling through the list but taking care not to So I must manually right-click and clear the output window and rerun it.
The transformer only runs when it detects a file change. This makes sense, but when coupled with 1 and 2, I've often cleared the output and I am running the transformer just to see a fresh list of errors. Which I don't get.
So my workflow becomes:
Write some code.
Run
Close dartium browser window (I'm not actually interested in running it, just seeing my errors)
See a bunch of errors. Realise that I didn't clear the errors from the previous run.
Right click and clear the pub serve output window.
Run again
Close dartium browser window again
Realise that the transformer has not run because it already ran in steps 1-3 and I haven't changed a file.
Change a file
Run again
Close dartium browser window again
Scroll through error list to find errors to fix
I find this a little cumbersome. Perhaps there is a philosophical point here on relying too much on my tooling to identify and fix errors (although I thought that was the entire point) but I'm just wondering what other people do to simplify this - I'm faintly surprised I appear to be alone in this.
You may run 'Pub Build' (available in the right-click menu of pubspec.yaml file and also right in the editor when pubspec is open). It is not incremental, so it runs longer (i.e. runs from scratch each time) but it gives you the list of errors just as if you've cleared Pub Serve output, edited each file in the project, started run configuration and closed a browser.
Sometimes errors are only shown when pub serve generates output the first time. For reloads some errors aren't shown anymore.
I'm not sure if this is a limitation of pub serve or a bug in the transformers.
pub serve is going to be replaced a new build system that builds to disk instead of in-memory only.
DDC isn't perfect yet either, but it's the future and I'd suggest to try this instead. There are known performance problems with Angular, but they are working on it.
See also
- https://webdev.dartlang.org/tools/dartdevc
- https://github.com/dart-lang/build
Related
I have a solution written in VB with some C# components. The solution uses some libraries from 2 outside sources. I have been working on this project for several months without issue. I cannot identify anything specific that I did to change my system or configuration. I was just working through the code, transitioning from an old set of library calls to the new library calls. The new library calls require complete rewrite so I change sections of the code and test to that point. Visual Studio 2013 debugger as of Friday morning will no longer recognize or show my local variables in this solution. The only things that appear in the Locals window are under Me. The code does work and I have it writing out to a text log file to confirm the values of variables at certain points, but the debugger has gone blind. When I add any of these local variables to Watch the response is " is not declared. It may be inaccessible due to its protection level."
Steps I've taken so far with no permanent success:
looked online and tried the few matches I found with no success
deleted the bin and obj folders and had the solution rebuild with no
success
recreated solution from scratch, copied over base files and rebuilt
solution and project (which worked for a few hours), until I did a
rebuild project and problem appeared again
updated to pack 5 and no success
I have opened my older projects and checked them. The debugger runs just fine and shows the variables. It is obviously something that happens during the rebuild process.
Any assistance would be greatly appreciated.
Thanks
Update:
Let me try to explain a little more clearly the situation.
I have an application I have built and am selling to some customers. Version 1 is installed and running at several locations. It is written in VB and uses some older COM libraries for a particular integration process.
The vendor is retiring the COM libraries. Their new libraries are in C#.
I created a new copy of my entire application (solution) and imported the new C# libraries. I have been going through and replacing the old code with the code for the new calls.I recompiled the solution and everything ran fine in debug.
The objects used with the new calls are completely different and there is limited documentation so I update a section of code and test to that point. Each time I "Save all Files", rebuild the project and test the changes. Everything worked fine for a few days. On Friday morning I started working on more changes and got an odd error. The system was not getting a proper value for a certain variable. When I went to check it in the WATCH window, debugger said it could not evaluate it. I figured something was hung up so I shut everything down and rebooted my machine. When I tried it again later, the same problem.
After several hours of no success I exited VS, renamed the folder to "OLD" and recreated the solution from the older version. Immediately everything was looking fine. I started making the changes and testing. Each time I did a rebuild, everything looked fine until the last change. Here I am again.
The code works fine up to the point I have updated. The only issue is that the debugger windows are not working correctly. If a variable is declared at the Class level outside the Sub, they can be seen. The only variables the debugger is blind to are the local variables within the running Sub.
I looked for anyone else with this issue and only found a few items. I tried the suggestions but no joy. I am left with having to temporarily define the variables outside the sub so I can see them while debugging.
I am on VS 2013 Update 5.
Do I need to move to VS 2015 to get around this?
Thanks again for your time and assistance.
I am assuming that you haven't changed versions of visual studio since the last time a rebuild worked for you.
recreated solution from scratch, copied over base files and rebuilt solution and project (which worked for a few hours), until I did a rebuild project and problem appeared again
Based on this, you create it from scratch and everything works until you do the rebuild right? But you are copying the base files still and you have new library calls since the last time a rebuild didn't mess up the locals window. So one of those is almost assuredly the culprit.
Since the library calls seem to be the thing that changed based on your post start there. If you go back to the old code and do a rebuild does it fix it? Assuming so, put the library calls back one at a time until it breaks.
If going back to the old code doesn't fix it, create from scratch with the old code and copy over the base files and rebuild. If that fixes it, add new library calls one at a time and rebuild after each until it breaks.
If that doesn't fix it either, then you will need to dig deeper on what else might have changed.
You are copying base files so eliminate those as the problem if you can:
Are you able to use placeholders instead of the base files or something that won't necessarily work as a finished product but that will allow you to debug, rebuild, debug again to see if the problem is related to one of them? Check the dates on the base files and ensure that they haven't changed since the last time a rebuild worked.
Something you could do concurrently could be to have have a colleague do a rebuild on their machine and see if the same issue comes up for them. It would (almost) completely eliminate the possibility that it is a configuration / program corruption issue on yours. Alternately, there are some free vb.net compilers online that you can upload files and code to. I'm not sure if that would be practical for you (due to the components of your program and/or sensitivity of the data) or not and haven't ever tried any where there is C# code in there but I wouldn't think that would be an issue.
We are trying to use grunt-lesslint in our project, as our UI developer is comfortable fix errors in less file. grunt-recess seems more powerful but not sure if it can point errors in less file itself. I am unable to comprehend enough from lesslint page, and there do not seem to be many examples. Does anyone know the following:
How to prevent lesslint from displaying on the console. I use formatters and the report file is generated, but it also prints on console, which I do not want to.
How to make lesslint fail only in the case of errors (not warnings). Also csslint seems to report errors also, while lesslint mostly gives warnings only, why is that so? Does lesslint throw errors as well? How to make it fail only in case of errors?
I tried using 'checkstyle-xml' formatter, but it does not seem to use it (I have used in jshint and it gives a properly formatted xml, which it does not give for lesslint).
Is it possible to compile less (many files or directories) in conjunction with lesslint? Any example?
Thanks,
Paddy
I'd say it's more of a common practice to display stdout for this kind of thing; the JSHint plugin does it, as does any other linting plugin that I've used. If you get in another developer that uses Grunt they'll probably expect stdout too. If you really want to override this, use grunt-verbosity: https://npmjs.org/package/grunt-verbosity
Again, this is a convention in Grunt; if a task has any warnings then it fails. The reason being if you lint a file and the linter flags something up it should be dealt with straight away, rather than delay it; six months time you have 500 errors that you haven't fixed and you're less likely to fix them then. Most linting plugins allow you to specify custom options (I've used CSS Lint and that is very customisable), so if you don't like a rule you can always disable it.
This should work. If there's a bug with this feature you should report it on the issue tracker, where it will be noticed by the developers of the plugin. https://github.com/kevinsawicki/grunt-lesslint/issues
Yes. You can set up a custom task that runs both your linter and compile in one step: something like grunt.registerTask('buildless', 'Lint and compile LESS files.', ['lesslint', 'less']); note that you'll have to install https://github.com/gruntjs/grunt-contrib-less to get that to work. Also note that, failing linting will not compile your LESS files; mandate that your code always passes the lint check; you'll help everyone involved in the project.
Currently, I define the following function in the REPL at the start of a coding session:
(defn rt []
(let [tns 'my.namespace-test]
(use tns :reload-all)
(cojure.test/test-ns tns)))
And everytime I make a change I rerun the tests:
user=>(rt)
That been working moderately well for me. When I remove a test, I have to restart the REPL and redefine the method which is a little annoying. Also I've heard bad rumblings about using the use function like this. So my questions are:
Is using use this way going to cause me a problem down the line?
Is there a more idiomatic workflow than what I'm currently doing?
most people run
lein test
form a different terminal. Which guarantees that what is in the files is what is tested not what is in your memory. Using reload-all can lead to false passes if you have changed a function name and are still calling the old name somewhere.
calling use like that is not a problem in it's self, it just constrains you to not have any name conflicts if you use more namespaces in your tests. So long as you have one, it's ok.
using lein lets you specify unit and integration tests and easily run them in groups using the test-selectors feature.
I also run tests in my REPL. I like doing this because I have more control over the tests and it's faster due to the JVM already running. However, like you said, it's easy to get in trouble. In order to clean things up, I suggest taking a look at tools.namespace.
In particular, you can use clojure.tools.namespace.repl/refresh to reload files that have changed in your live REPL. There's alsorefresh-all to reload all the files on the classpath.
I add tools.namespace to my :dev profile in my ~/.lein/profiles.clj so that I have it there for every project. Then when you run lein repl, it will be included on the classpath, but it wont leak into your project's proper dependencies.
Another thing I'll do when I'm working on a test is to require it into my REPL and run it manually. A test is just a no-argument function, so you can invoke them as such.
I am so far impressed with lein-midje
$ lein midje :autotest
Starts a clojure process watching src and test files, reloads the associated namespaces and runs the tests relevant to the changed file (tracking dependencies). I use it with VimShell to open a split buffer in vim and have both the source and the test file open as well. I write a change to either one and the (relevant) tests are executed in the split pane.
I have a rather large solution of about 20 projects, that I have worked with for about three years now. In recent weeks, I seemed to have lost the ability to just press "Run and Debug" on my toolbar. If I do this, I lose all ability to debug and use breakpoints. My breakpoints all get a little yellow exclamation point on them and never run. Below is an example.
To fix this so I can get debugging again I have to fully clean the project. If I try to run after cleaning, i get an error saying the file does not exist. So I have to then fully re-build the project before I can run. It works just once, and then I lose my ability to debug again. Below is the error if I don't manually build first:
This seems to be effecting several (maybe all) projects in my solution and is causing extremely large problems for me. Some of my projects I need to breakpoints to verify data before it runs, and if I forget to do this, it runs before verifying.
Any help in fixing this would be amazing. Under solution settings, my start up project i set to "Current Selection" if that helps.
UPDATE:
So far under solution settings -> configuration a bunch of my items had "Build" not selected for Debug. Not sure how that happened, but updating that for now seems to have solved it. If it goes to not building each time again I will re-update, but for now I think that solved the issue.
I'm seeing a strange build bug a lot. Sometimes after typing some code we receive the following build error.
Class 'clsX' must implement 'Event PropertyChanged(sender As Object, e As PropertyChangedEventArgs)' for interface System.ComponentModel.INotifyPropertyChanged'.
And
'PropertyChanged' cannot implement 'PropertyChanged' because there is no matching event on interface 'System.ComponentModel.INotifyPropertyChanged'.
Those error should never go together! Usually we can just ignore the exception and build the solution but often enough this bug stops our build. (this happens a lot using Edit and Continue which is annoying)
We're using Vb.net and c# mixed in one big solution.
Removing the PropertyChanged event and retyping the same code! sometimes fixes this.
Question:
Has anyone else seen this problem and has some suggestions how to prevent his?
We're using a code generator that causes this error to surface but just editing some files manually triggers this exception too. This error occur's on multiple machines using various setups.
Someone had the same exact issue discussed here. It sounds like there is an issue with this build picking up an old version of a binary. I would try the following in order:
Verify all assembly references use project references where possible within the Visual Studio solution.
Disable build parallelization in case there is some weird file locking issue with concurrent project builds. Go to Tools -> Options, Projects and Solutions -> Build and Run, then set "maximum number of parrellel project builds" to 1. Not the best solution but it may help narrow down the problem.
Disable the Hosting Process in case it's locking some file causing an assembly to not get rebuilt correctly. For C# project go to Project Properties, Debug tab, and uncheck "Enable the Visual Studio hosting process". For VB.NET project you'll need to Unload Project, Edit the project file, and add <UseVSHostingProcess>false</UseVSHostingProcess> to the PropertyGroup of each configuration. Again, not the best solution but you probably won't notice a difference.
Lastly, try doing a Clean + Build to try and resolve the issue when it occurs (I know this is not a fix but it's easy enough to do), also Rebuild may be slightly different than Clean + Build so try the latter if the former doesn't work.
As I can not comment due to lack of appropriate points.
But I would like to share one of my experience:
In an aspx.cs page I was working, used to compile fine and some time gave mysterious error of a variable not defined or function not defined or sometime variable or the function defined two times. I changed possibly each and every variable and function name but there seemed no effect , but after entering a simple space or a new line at any place in the file used to solve the compile error. At one time I tried to save the file (in a different encoding as i am used to experiments) and found that the file was not saving in the correct encoding (i.e. the ansi encoding because the file had a unicode character ), I removed the unicode character and that compile error didn't bothered me again.
This unicode character problem could be (not a hard and fast rule) there so you could check it.
Nuke & restore using source control (TFS instructions here):
Make sure you have everything checked in
Exit Visual Studio
Rename the project directory to .Bak (effectively deleting it)
Reopen Visual Studio and in source control:
Get Specific Version
check 'Overwrite... not checked out' and 'Overwrite ... even if local version matches'
Re-open project
Another problem: Make sure some source files are not newer than the current date (or your date is set back). Often this happens in apps where you are doing logic that requires certain things to happen differently on certain dates. You change your clock to test it, make a revision to the source with the date advanced, set the date back, and viola, rebuild does not rebuild that file.
You say 'typing it in again' - can you try just saving? After 40 years since MULTIX the .net build still decides what has changed by checking the file timestamp.
good luck!
When you get the error, is it always on the VB calling C# side, or vice-versa, or does it work both ways?
If the answer is either of the first two situations, try building the "callee" project within the solution before building the "caller" project to see if it stops the situation.
Also, just in case it may jog something for you to think about, does this error crop up when you change a VB file or a C# file, or is there no correllation?
Oh, and sorry this looks like an answer instead of a comment, I cannot post comments yet (need 50 rep).