I'm trying to copy a directory from one folder to another folder like so:
directory "C:\\test\\go" do
recursive true
action :create
end
cookbook_file "C:\\Automation" do
source "C:\\Automation"
path "C:\\test\\go"
action :create_if_missing
end
It creates the target folder C:\test\go but does not copy anything.
The documentation says that it should also handle directories so any ideas why it does not?
I've also tried a wildcard source "C:\Automation\*" and also tried forward slashed...
Here is the solution:
ruby_block "get the windows resources" do
block do
FileUtils.mkdir_p mod_path
FileUtils.cp_r(Dir["#{RESOURCE_DIR}/*.exe"], Chef::Config[:file_cache_path])
end
end
Related
I have a Jar (we'll call it a.jar) with a resource in it at path foo/bar.txt and a function as follows:
object FooBarLoader {
fun loadFooBarText() = javaClass.getResourceAsStream("foo//bar.txt")
?.bufferedReader()
?.readLines()
?.joinToString("\n")
}
When I test the function in a unit test (JUnit 4, running with Gradle 6), it loads the text from the resource file despite the obvious typo (the // in the middle of the resource path).
I also have a CLI application (in b.jar) that has a dependency on a.jar. When the CLI application calls loadFooBarText(), it got a null result due to the resource not being found. This was fixed by fixing the typo (// -> /) in the function in a.jar. No other changes were needed to fix it.
So, my question is why did the wrong path work in one situation (unit tests of a.jar) and not the other (call from b.jar)?
How do you run the unit test with a.jar ? Just run it in your IDE or use command java -jar a.jar ?
If you ran it just in IDE，I think difference is the search path between local files and zip files .
Your first application searches the file in your target directory and the second application searches it in the jar which is a compressed file.
When searching files in local path, command will be changed to right one by system.
The two commands below are the same in both Windows/Linux.
cd work//abc/ddd
cd work/abc/ddd
But when searching files in a jar file which is actually compressed zip file, path should be a restrict written or else the program will find nothing.
I am getting following error during deployment when there was a folder recursively removed from code. As a workaround i had to manually delete the folder from destination path on server.
Error: (12/26/2018 4:32:27 AM) An error occurred when the request was processed on the remote computer.
Error: An error was encountered when processing operation 'Delete Directory' on 'C:\Alam\Test\MSF\Config'.
Error: The error code was 0x80070091.
Error: The directory is not empty.
at Microsoft.Web.Deployment.NativeMethods.RaiseIOExceptionFromErrorCode(Win32ErrorCode errorCode, String maybeFullPat
h)
at Microsoft.Web.Deployment.DirectoryEx.Delete(String path)
at Microsoft.Web.Deployment.DirPathProviderBase.Delete(Boolean whatIf)
Error count: 1.
I can confirm this bug.
If a file in a directory that should be deleted (i.e. that is not excluded) is excluded then the file does not get deleted and the following rmdir crashes MsDeploy.
Work around
You must exactly ensure that no file exclution rule happens to match to a sub folder that is not excluded.
To satify this condition all file exclution rules must match from the deploy root path only. Otherwise a file in a sub directory may not be deleted and the following directory removal fails.
It can be quite difficult to ensure this, because the initial part of the deploy path ist typically not predictable if you do not always deploy to only one directory on only one server.
Fortunately MsDeploy tends to use different path delimiters for the target directory and the sub directories, e.g.:
Default Web Site/VirtualPath\Folder\File. Note the forward slash and the backslash.
So it is possible to use the forward slash as indicator of the root path.
Example
<MsDeploySkipRules Include="CustomSkipFile">
<ObjectName>filePath</ObjectName>
<AbsolutePath>/[^\\]+\\Custom</AbsolutePath>
</MsDeploySkipRules>
<MsDeploySkipRules Include="CustomSkipFile">
<ObjectName>dirPath</ObjectName>
<AbsolutePath>/[^\\]+\\Custom</AbsolutePath>
</MsDeploySkipRules>
This will exclude all files starting with Custom in the top level and all directories starting with Custom in the top level.
But this will not exclude a file that starts with Custom in a sub directory that does not start with Custom like notCustomDir\Custom.csv.
The /[^\\]+\\ prefix ensures that the following part cannot match to a deeper path because the latter is not delimited with a forward slash.
I would not consider this a sophisticated solution. It is just the only one I got working.
I am trying to skip a sub folder and certain file types while deploying a webapp. I tried the below
Test.deploy.cmd /T /M:test.noreply.com /
P:a /G:False "-skip:Directory="Config\\Testdata\\Resx" -skip:filePath,absolutePath="Config\\Testdata\\*.xml""
The above is not working for some reason, am i missing something?
i think below worked for me
-skip:file="Config\\Testdata\\.*\.xml"
I am trying to access static resource (eg. first.html) packed inside the same .jar file (testJetty.jar), which also has a class which starts the jetty (v.8) server (MainTest.java). I am unable to set the resource base correctly.
The structure of my jar file (testJetty.jar):
testJetty.jar
first.html
MainTest.java
==
Works fine on local machine, but when I wrap it in jar file and then run it, it doesn't work, giving "404: File not found" error.
I tried to set the resourcebase with the following values, all of which failed:
a) Tried setting it to .
resource_handler.setResourceBase("."); // Results in directory containing the jar file, D:\Work\eclipseworkspace\testJettyResult
b) Tried getting it from getResource
ClassLoader loader = this.getClass().getClassLoader();
File indexLoc = new File(loader.getResource("first.html").getFile());
String htmlLoc = indexLoc.getAbsolutePath();
resource_handler.setResourceBase(htmloc); // Results in D:\Work\eclipseworkspace\testJettyResult\file:\D:\Work\eclipseworkspace\testJettyResult\testJetty1.jar!\first.html
c) Tried getting the webdir
String webDir = this.getClass().getProtectionDomain()
.getCodeSource().getLocation().toExternalForm();
resource_handler.setResourceBase(webdir); // Results in D:/Work/eclipseworkspace/testJettyResult/testJetty1.jar
None of these 3 approaches worked.
Any help or alternative would be appreciated
Thanks
abbas
The solutions provided in this thread work but I think some clarity to the solution could be useful.
If you are building a fat jar and use the ProtectionDomain way you may hit some issues because you are loading the whole jar!
class.getProtectionDomain().getCodeSource().getLocation().toExternalForm();
So the better solution is the other provided solution
contextHandler.setResourceBase(
YourClass.class
.getClassLoader()
.getResource("WEB-INF")
.toExternalForm());
The problem here is if you are building a fat jar you are not really dumping your webapp resources into WEB-INF but are probably going into the root of the jar, so a simple workaround is to create a folder XXX and use the second approach as follows:
contextHandler.setResourceBase(
YourClass.class
.getClassLoader()
.getResource("XXX")
.toExternalForm());
Or change your build tool to export the webapp files into that given directory. Maybe Maven does this on a Jar for you but gradle does not.
Not unusually, I found a solution to my problem. The 3rd approach mentioned by Stephen in Embedded Jetty : how to use a .war that is included in the .jar from which Jetty starts? worked!
So, I changed from Resource_handler to WebAppContext, where WebAppContext is pointing to the same jar (testJetty.jar) and it worked!
String webDir = MainTest.class.getProtectionDomain()
.getCodeSource().getLocation().toExternalForm(); ; // Results in D:/Work/eclipseworkspace/testJettyResult/testJetty.jar
WebAppContext webappContext = new WebAppContext(webDir, "/");
It looks like ClassLoader.getResource does not understand an empty string or . or / as an argument. In my jar file I had to move all stuf to WEB-INF(any other wrapping dir will do). So the code looks like
contextHandler.setResourceBase(EmbeddedJetty.class.getClassLoader().getResource("WEB-INF").toExternalForm());
so the context looks like this then:
ContextHandler:744 - Started o.e.j.w.WebAppContext#48b3806{/,jar:file:/Users/xxx/projects/dropbox/ui/target/ui-1.0-SNAPSHOT.jar!/WEB-INF,AVAILABLE}
So I have a bunch of data that I want to load into database from CSV. I've hacked together a solution that works in local development, but when I deploy to meteor.com, it no longer works.
I'm loading the csv file in the folder /server/data/:
function readData(name){
var fs = __meteor_bootstrap__.require('fs');
var path = __meteor_bootstrap__.require('path');
var base = path.resolve('.');
var data = fs.readFileSync(path.join(base, '/server/data/', name));
return CSVToArray(data);
}
After I deploy to meteor.com, i got:
INFO Error: ENOENT, no such file or directory '/meteor/containers/98eb1286-120b-ee84-8e98-ce673fa2eab7/public/data/categories.csv'
at Object.openSync (fs.js:240:18)
at Object.readFileSync (fs.js:128:15)
at readData (app/server/models.js:10:16)
at app/server/categories.js:6:7
at /meteor/containers/98eb1286-120b-ee84-8e98-ce673fa2eab7/bundle/server/server.js:132:63
at Array.forEach (native)
at Function.<anonymous> (/meteor/containers/98eb1286-120b-ee84-8e98-ce673fa2eab7/bundle/server/underscore.js:76:11)
at /meteor/containers/98eb1286-120b-ee84-8e98-ce673fa2eab7/bundle/server/server.js:132:7
Any idea how I can get meteor to see the csv file after deployment?
I realize this question is old, but it still ranks high on certain keyword searches. So, if you're using Meteor 0.6.5+, you can use the new Assets API.
The issue is that meteor only bundles files that it knows about (ie. JS/CSS/HTML/+more depending on which packages you use) up when it deploys.
Try putting the file you need in the public directory (this directory is exempt from the above rule).
Thanks to SamuelDavis and Tom Coleman's tips. I ended up figuring out what the problem is. Turns out the bundled app is no longer formated as client, public, and server. I ended up debugging it by running meteor bundle to create a tarball. extract the tarball and took a look inside to find where the data folder is. Tom was also right that the data folder needed to be in the public folder in order to get bundled in.
It appears that the base directory is not in the same location that contains the file '/server/data/xxx.csv'.
Before you try anything else, log the base path after calling "var base = path.resolve('.'). If that value is what you expected, log the files that appear in that directory. Again if the files are what you expected, navigate into the /server folder and print out those directories and so forth.
This should pinpoint you to which folder and/or directory is missing and should indicate where you should place the CSV file in future.