I'm trying to define a graph by using a walker ability; Following is the code which I tried. I god this from the jaseci bible. But removed unnecessary lines for simplicity of the quection.
node person {
has name;
has byear;
}
walker init {
can setup {
person1 = spawn here --> node::person;
std.out(person1);
}
root {
::setup;
take --> ;
}
}
When I run this code I get following error
2022-11-10 18:48:24,238 - ERROR - rt_error: abilities_6.jac:blank - line 9, col 33 - rule node_ref - Spawn can
not occur on <class 'jaseci.actor.walker.Walker'>!
2022-11-10 18:48:24,238 - ERROR - rt_error: abilities_6.jac:init - line 9, col 29 - rule edge_ref - Internal Exception: 'NoneType' object has no attribute 'value'
I need assistance in fixing errors in this code.
Related
Expected Behavior:
Clicking "Run doctest" in vscode should execute one test from doctest snippets.
Terminal output SHOULD say ("1 passed;" or "1 failed;"), and "1 filtered out;".
Actual Behavior:
Clicking "Run doctest" in vscode executes 0 tests, and shows that 2 were filtered out.
Terminal output:
test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 2 filtered out; finished in 0.00s
Source code:
My misbehaving crate: https://github.com/darrow-olykos/rusty-gists/blob/main/crates/my_cache/src/lib.rs
Behaving crate (where this is not an issue): https://github.com/darrow-olykos/rusty-gists/blob/main/crates/math/src/lib.rs
My machine:
macOS 12
rustc 1.57.0
rust analyzer v0.3.954
What I have done to try to narrow down the scope of the problem:
Running the "same" command in the terminal demonstrates expected behavior. The terminal output shows test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 1 filtered out; finished in 0.40s when I run cargo test --doc --package my_cache -- "Cacher<T>::new" --nocapture, which is exactly what the terminal says is ran when I click on "Run Doctest".
Clicking "Run Doctest" in another crate I have, (called "math") in the same repo, demonstrates expected behavior.
Looking at the differences between my misbehaving crate and my working crate:
A. This misbehaving crate has it's doctest is inside of an impl where-as that other crate's doctest is at the root level of the file.
B. This misbehaving crate's doctest is for a generic struct that accepts a closure type.
C. Executing cargo test from crates/my_cache demonstrates expected behavior, with the following terminal output:
// ... some output omitted
Doc-tests my_cache
running 2 tests
test src/lib.rs - Cacher<T>::new (line 26) ... ok
test src/lib.rs - Cacher<T>::value (line 42) ... ok
test result: ok. 2 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.44s
Again, here is the source code:
Misbehaving crate: https://github.com/darrow-olykos/rusty-gists/blob/main/crates/my_cache/src/lib.rs
Behaving crate: https://github.com/darrow-olykos/rusty-gists/blob/main/crates/math/src/lib.rs
Maybe? notable details:
I have modeled my git repo after the structure of https://github.com/rust-analyzer/rust-analyzer, such that I'm using cargo workspaces and crates/* are members. I have a root cargo.toml which can specify local dependencies with something like my_cache = {path = "crates/my_cache}, however I cannot think of a reason why this would be a contributing factor because I've proven that my math crate can be in this structure without vscode getting confused and filtering out doctests in that crate unexpected.
My suspicions?:
Something is happening that is causing the doctest to be filtered out when it should not be filtered out.
Maybe the command that claims to be executing when I click Run Doctest isn't the ACTUAL command getting executed.
Maybe the (bug?) has something to do with the closure type. I forget where I read this, but I vaguely recall Rust closure types are "unnamed" in a way that makes referencing them strange. Unfortunately I cannot find the resource that I was reviewing that walked through this in detail. (It might have been a resource that was covering the Rust compiler and how the Rust compiler manifests data types in memory, but I do not recall details (maybe someone reading this will know what I'm referring to here)).
Here is the misbehaving crate's source code copied into this answer for the sake of longevity in case I make changes to my github repo:
// Credit to: https://doc.rust-lang.org/book/ch13-01-closures.html
// (I've modified their example to use a HashMap instead of a single value)
use std::collections::HashMap;
/// cacher for calculation with two u64's as input and u64 as output
/// can be generalized more
pub struct Cacher<T>
where
T: FnMut(u64, u64) -> u64,
{
calculation: T,
values: HashMap<String, u64>,
}
impl<T> Cacher<T>
where
T: FnMut(u64, u64) -> u64,
{
/// Returns a Cacher<T> which can cache results of calculations for the provided closure.
///
/// # Arguments
///
/// `T` - Closure that computes produces a value. Value is cached based on args. Cached value is returend on subsequent calls if args are the same.
///
/// # Examples
/// ```rust
/// use my_cache::Cacher;
/// let mut cacher = Cacher::new(|x,y|x+y);
/// ```
pub fn new(calculation: T) -> Self {
let values = HashMap::new();
Cacher {
calculation,
values,
}
}
/// Returns value of calculation `T`. Cached value is returned if unique `n`, `k` pair provided, otherwise calcuation runs and then value is cached.
///
/// # Examples
///
/// ```rust
/// use std::rc::Rc;
/// use std::cell::{RefCell, RefMut};
///
/// use my_cache::Cacher;
///
/// let mut count = Rc::new(RefCell::new(0));
/// let add = |x, y| {
/// let mut count_mut_ref = count.borrow_mut();
/// *count_mut_ref += 1; x + y
/// };
/// let mut cacher = Cacher::new(add);
///
/// assert_eq!(*count.borrow(), 0);
/// assert_eq!(cacher.value(2, 3), 5); // new calculation, count += 1
/// assert_eq!(*count.borrow(), 1);
/// assert_eq!(cacher.value(2, 3), 5); // repeat, use cache
/// assert_eq!(*count.borrow(), 1);
/// assert_eq!(cacher.value(2, 4), 6); // new calculation, count += 1
/// assert_eq!(*count.borrow(), 2);
/// ```
pub fn value(&mut self, n: u64, k: u64) -> u64 {
let key = n.to_string() + &k.to_string();
let cached_result = self.values.get(&key);
if let Some(value) = cached_result {
*value
} else {
let v = (self.calculation)(n, k);
self.values.insert(key, v);
v
}
}
}
This is a bug in a pre-release version of vscode's rust analyzer extension (v3.0.954) and can be mitigated by switching back to the latest release version of rust analyzer.
I realized to check this once I posted my question and confirmed that the unexpected behavior was only present in v3.0.954, however the latest release of rust analyzer vscode extension v0.2.948 works as expected (does not filter out the doctest unexpectedly).
It seems the exception stacktrace in iOS only contains the method name or there is a bug. Below is my code of handling exceptions in JSContext.
context.exceptionHandler = { (ctx: JSContext!, value: JSValue!) in
// type of String
let stacktrace = value.objectForKeyedSubscript("stack").toString()
// type of Number
let lineNumber = value.objectForKeyedSubscript("line")
// type of Number
let column = value.objectForKeyedSubscript("column")
let moreInfo = "in method \(stacktrace)Line number in file: \(lineNumber), column: \(column)"
Logger.error("JS ERROR: \(value) \(moreInfo)")
}
And I got logs like below
ERROR : JSContextRenderer.swift:308 : setupContext : JS ERROR: Error in method clearBackground
Line number in file: 162, column: 12"
Note there is a new line right after the "clearBackground" name, I think there probably more information there.
Can anybody having similar experience confirm? Any help is appreciated. Thanks.
Looks like it does show more information in the stack. Here is one of the log information I got:
JS ERROR: TypeError: undefined is not a function (evaluating 'msg.__assert__()') in method assert
syncShots
updateSync
initSync
setState
onLogicStateChanged
onLogicStateChanged
[native code]
updateMove
sendMove
shoot
onTouchUp
[native code]
_handleEvent
_dispatchEvent
. Line number in file: 183, column: 20
Long ago, #igrek asked where the keys in value come from. They're part of the Error object that was thrown by the JavaScript engine, and that now appears as value in the native error handler callback. What exactly Error includes depends on the implementation, but typically includes message, fileName, and lineNumber. It appears that stack is also supported. More details on the MDN page for Error .
My Meteor app runs fine on localhost, but when I deploy it to myApp.meteor.com, I get the error below. I struggle to understand these error logs. Any ideas?
[Thu Jun 25 2015 06:35:23 GMT+0000 (UTC)]
WARNING /meteor/dev_bundles/0.4.18/lib/node_modules/fibers/future.js:278
throw(ex);
^
[Thu Jun 25 2015 06:35:23 GMT+0000 (UTC)]
WARNING MongoError: ns name too long, max size is 128
at Object.Future.wait
(/meteor/dev_bundles/0.4.18/lib/node_modules/fibers/future.js:398:15)
at [object Object].MongoConnection._ensureIndex
(packages/mongo/mongo_driver.js:733:1)
at FileCollection.Mongo.Collection._ensureIndex
(packages/mongo/collection.js:620:1)
at new FileCollection (packages/vsivsi:file-
collection/gridFS_server.coffee:65:15)
at FileCollection (packages/vsivsi:file-
collection/gridFS_server.coffee:21:24)
at app/lib/collections.js:15:12
at /meteor/containers/55bd5013-6244-5f59-9f74-
c6798eb58003/bundle/programs/server/boot.js:229:5
- - - - -
And the code in my collections.js file is listed below with line 15 noted in a comment. It is based on the Meteor File Collection Sample App line 10. Note that the MD5 is actually quite a long identifier, but I'm not sure how/if it's used by Mongo. Regardless, it works on localhost.
Meteor.startup(function () {
console.log('collections are loading');
myData = FileCollection('myFiles', { // This is line 15 that triggers error
resumable: true,
http: [
{
method: 'get',
// This GET routes path like /gridfs/myFiles/md5/9f4cd4e1d9e7cb1273ad2860fa909398.jpeg
path: '/md5/:md5',
// note that my :md5 also contains suffix with media type. Suffix was appended in link helper
// I added suffix so that other programs can determine MIME type from URL's suffix, without reading file header
lookup: function (params, query) {
// params.md5 contains more than the MD5 and hence needs to be split to get true MD5 portion
var tempArray = params.md5.split('.');
var firstHalf = tempArray[0]; // heres the MD5 part to ensure that Vaughn's code isn't affected
return {
md5: firstHalf // a query mapping url to myFiles
};
}
}
]
});
});
Looks like your problem is related to the closed issue here: https://github.com/vsivsi/meteor-file-collection/issues/55
Your local db name is probably much shorter than the one assigned to you in meteor.com environment.
Try adding something like
resumableIndexName: 'fci',
after line
resumable: true,
could you show me app/lib/collections.js:15 ?
Do you have a collection with really long name defined?
Trying to learn the Meteor framework as well as coffeescript/node all at once. Been working on a simple file upload program that utilizes onloadend. When the FileReader onloadend event function is called I try to determine if the file already exists and if so I update with the new file data and version.
The code works for insert but not update. Can someone help? I've posted to meteor-talk w/o an answer as I suspect its the weekend (when I do most of my experimentation).
Code snippet...
file_reader.onloadend = ((file_event) ->
(event) ->
f_filename = escape file_event.name
version = 0
f_record = null
f_record = doc_repo.findOne { name: f_filename }
if f_record && f_record.name
doc_repo.update
name: f_filename
,
$set:
version: 10
else
doc_repo.insert
name: f_filename
data: event.target.result
version: 0
)(file_obj)
Error
Exception while invoking method '/documents/update' TypeError: Cannot read property 'toBSON' of undefined
at Function.calculateObjectSize (/usr/local/meteor/lib/node_modules/mongodb/node_modules/bson/lib/bson/bson.js:210:12)
at BSON.calculateObjectSize (/usr/local/meteor/lib/node_modules/mongodb/node_modules/bson/lib/bson/bson.js:1463:15)
at UpdateCommand.toBinary (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/commands/update_command.js:67:20)
at Connection.write (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/connection/connection.js:138:40)
at __executeInsertCommand (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/db.js:1837:14)
at Db._executeInsertCommand (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/db.js:1912:7)
at Collection.update (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/collection.js:445:13)
at app/packages/mongo-livedata/mongo_driver.js:178:16
at Db.collection (/usr/local/meteor/lib/node_modules/mongodb/lib/mongodb/db.js:507:44)
at _Mongo._withCollection (app/packages/mongo-livedata/mongo_driver.js:51:13)
It looks like Mongo is not getting the second parameter which it needs to do the update. So in regular JavaScript it's expecting this:
collection.update({..selector..}, { .. modifier });
So I'd try putting some curly brackets around the modifier object, like this:
doc_repo.update
name: f_filename,
{
$set:
version: 10
}
Is there an easy way to tell Selenium IDE that any action that results in a http 500 response means the test failed?
I have tests that are 75 page requests long. Sometimes, I get a crash and burn somewhere in the middle, but the tests come back green.
Taking a look at selenium-api.js, I saw that there is a parameter ignoreResponseCode in the signature of the doOpen method in selenium-api.js :
Selenium.prototype.doOpen = function(url, ignoreResponseCode) {
This parameter is used by the browserbot object :
if (!((self.xhrResponseCode >= 200 && self.xhrResponseCode <= 399) || self.xhrResponseCode == 0)) {
// TODO: for IE status like: 12002, 12007, ... provide corresponding statusText messages also.
LOG.error("XHR failed with message " + self.xhrStatusText);
e = "XHR ERROR: URL = " + self.xhrOpenLocation + " Response_Code = " + self.xhrResponseCode + " Error_Message = " + self.xhrStatusText;
self.abortXhr = false;
self.isXhrSent = false;
self.isXhrDone = false;
self.xhrResponseCode = null;
self.xhrStatusText = null;
throw new SeleniumError(e);
}
I've tried calling the open function from selenium IDE with value = false and this results in an error (test failed).
My PHP test page was :
<?php
header('HTTP/1.1 500 Simulated 500 error');
?>
And this results in :
For me, this solves the problem of checking HTTP response status.
Make a JavaScript file called "user-extensions.js" and add it to the Selenium-IDE under Options > Options. If you are running Selenium RC, pass it into the parameter when starting up your server in the jar command. There should be a user extensions javascript file attribute.
Then close and restart Selenium-IDE. The User-Extensions file is cached when the IDE starts up.
Add this code to your Selenium user-extensions.js file to make a custom command called "AssertLocationPart". As you know "assertLocation" and "storeLocation" are standard commands. I tried to reduce the extra line of code to storeLocation just by getting the href in the custom function. I wasn't able to get the doAssertValue command to work. I'll have to post my own question for that. That's why it's commented out. For now, just use "this.doStore" instead. And add an extra line to your script after your custom AssertLocationPart command. Since we're not actually doing an assertion in the custom function/command, we should call it "storeLocationPart" (function would be named "doStoreLocationPart"), not "assertLocationPart" (function would be named "doAssertLocationPart"), and just pass in the first parameter. But if you can get the doAssert* to work, please let me know. I'll mess with it another day since I need to do this same thing for work.
Selenium.prototype.doAssertLocationPart = function(partName,assertTo) {
var uri = selenium.browserbot.getCurrentWindow().document.location.href;
//alert("URI = " + uri);
var partValue = parseUri(uri,partName);
//alert("Part '" + partName + "' = " + partValue);
//this.doAssertValue(partValue,assertTo);
this.doStore(partValue,"var_"+partName);
};
// Slightly modified function based on author's original:
// http://badassery.blogspot.com/2007/02/parseuri-split-urls-in-javascript.html
//
// parseUri JS v0.1, by Steven Levithan (http://badassery.blogspot.com)
// Splits any well-formed URI into the following parts (all are optional):
//
// - source (since the exec() method returns backreference 0 [i.e., the entire match] as key 0, we might as well use it)
// - protocol (scheme)
// - authority (includes both the domain and port)
// - domain (part of the authority; can be an IP address)
// - port (part of the authority)
// - path (includes both the directory path and filename)
// - directoryPath (part of the path; supports directories with periods, and without a trailing backslash)
// - fileName (part of the path)
// - query (does not include the leading question mark)
// - anchor (fragment)
//
function parseUri(sourceUri,partName){
var uriPartNames = ["source","protocol","authority","domain","port","path","directoryPath","fileName","query","anchor"];
var uriParts = new RegExp("^(?:([^:/?#.]+):)?(?://)?(([^:/?#]*)(?::(\\d*))?)?((/(?:[^?#](?![^?#/]*\\.[^?#/.]+(?:[\\?#]|$)))*/?)?([^?#/]*))?(?:\\?([^#]*))?(?:#(.*))?").exec(sourceUri);
var uri = {};
for(var i = 0; i < 10; i++){
uri[uriPartNames[i]] = (uriParts[i] ? uriParts[i] : "");
if (uriPartNames[i] == partName) {
return uri[uriPartNames[i]]; // line added by MacGyver
}
}
// Always end directoryPath with a trailing backslash if a path was present in the source URI
// Note that a trailing backslash is NOT automatically inserted within or appended to the "path" key
if(uri.directoryPath.length > 0){
uri.directoryPath = uri.directoryPath.replace(/\/?$/, "/");
if (partName == "directoryPath") {
return uri.directoryPath; // line added by MacGyver
}
}
return uri;
}
Then add this to your web.config file and make sure customErrors is turned off. Since you have a 500 error, it will redirect the user to the default page. Feel free to add a custom page for a 500 HTTP status code if you want to be specific in your Selenium scripts.
<customErrors mode="On" defaultRedirect="/ErrorHandler.aspx">
<error statusCode="401" redirect="/AccessDenied.aspx" />
<error statusCode="403" redirect="/AccessDenied.aspx" />
<error statusCode="404" redirect="/PageNotFound.aspx" />
</customErrors>
This is what your commands will look like in the IDE:
Make sure you're on this page (or something similar) before running the script:
https://localhost/ErrorHandler.aspx?aspxerrorpath=/path/pathyouweretryingtoviewinwebapp.aspx
Log shows that it passed!
[info] Executing: |storeLocation | var_URI | |
[info] Executing: |echo | ${var_URI} | |
[info] echo: https://localhost/ErrorHandler.aspx?aspxerrorpath=//path/pathyouweretryingtoviewinwebapp.aspx
[info] Executing: |assertLocationPart | fileName | ErrorHandler.aspx |
[info] Executing: |assertExpression | ${var_fileName} | ErrorHandler.aspx |
Using the error handler from my previous answer:
Command: assertLocation
Target: regexp:^(https://localhost/ErrorHandler.aspx).*$
Or (per your comment) it's inverse if you don't have error handling turned on, use AssertNotLocation. This may require more work on the person writing the scripts. You'd have to keep track of all pages.
More on pattern matching:
http://seleniumhq.org/docs/02_selenium_ide.html#matching-text-patterns
http://www.codediesel.com/testing/selenium-ide-pattern-matching/