bootstrapworld / wescheme Goto Github PK

We need to retrofit android detection into browser-detect.js, and throw up a warning. Ask @shriram to help test.

Go to https://www.wescheme.org/run?publicId=uGd1iMzAFg

Following is the result at the end:

(world(posn250504.5)(list(cow(posn368473.5)"left")(cow(posn7473.5)"right")(cow(posn493473.5)"left")))

if there are multiple animate calls in the same file, only the first one is run and then the programs stops, in DrRacket the next one starts automatically (this is not a big issue)
http://www.wescheme.org/view?publicId=vGXt3Af9bf

As of Sept 30th, we switched to the concisely named Google Sign-In Platform Library for web in order to get WeScheme back up and running.

However, this is deprecated and will not work in the long run. We should switch over to the clearly-better-named Google Identity Services for Web instead.

October 16, 2023
WeScheme.org

Expected: following code gives an error because of the repeated to-draw clause.

Actual: no error, runs as if the only clause were the second one.

(define c (circle 10 "solid" "blue"))
(define b (square 10 "solid" "orange"))

(define (draw-1 img) 
  (overlay img (empty-scene 200 100)))
(define (draw-2 img) 
  (overlay (above img b) 
           (empty-scene 200 100)))
(define (th x)
  (beside x c))
(big-bang c
          (to-draw draw-1)
          (to-draw draw-2)
          (on-tick th 0.5))

WeScheme couldn't read a literal character #\{

I'm getting the behavior in this GIF when I try to insert images. Interestingly, this is the same behavior I get on CPO if I turn off the scopes for drive.photos and drive.photos.readonly (this is what prompted me to try!)

On my first clickthrough I got a permissions screen that looked like it was requesting the drive.file permission (the one for just files opened or created by the app).

Any thoughts?

Howdy!
I've just installed weScheme locally on my machine. I cloned wescheme2012 master branch, ran 'build-console-and-editor.rkt' and got the local server up and running. However, when serving the page http://localhost:8081/openEditor, the files support-min.js, undo.jpg and redo.png are 404. The installation did properly minimize the js files, but apparently war/js/mzscheme-vm/support.js was not cloned in the first place. I suspected some quirks in my installation, but after quickly looking in the repo, I couldn't find the three files in question. Where can I find those files?

Both flip-vertical and flip-horizontal screw up the bounding box handling. Apparently after a flip everything inside the bounding box is considered part of the image, not just the drawn part. Unlike DrRacket and non-flipped WeScheme images.

Sample code (first line for DrRacket only):

(require 2htdp/image)
(define good (triangle 100 "solid" "orange"))
(define bad (flip-horizontal (triangle 100 "solid" "blue")))
(beside good bad)

(beside/align "bottom"
              (frame (rotate 60 good))
              (frame (rotate 60 bad)))

Results from DrRacket 8.9 CS, as expected:

Results from WeScheme.org today:

The whitespace around the blue triangle ("bad") is expanded, presumably including the space outside of the triangle in the figure.

These two images are not reported equal, but should be. I think this is because scene equality checks the order of adding things to the scene, and doesn't have a fallback for false.

For example:

  (put-image (circle 50 'solid 'red)
             100 100
             (put-image (rectangle 50 60 'solid 'brown)
                        300 300
                        (rectangle 500 500 'solid 'blue))))

(define si2
  (put-image (rectangle 50 60 'solid 'brown)
                        300 300
             (put-image 
                        (circle 50 'solid 'red)
             100 100
                        (rectangle 500 500 'solid 'blue))))

(define si3
  (put-image (circle 50 'solid 'red)
             100 100
             (put-image (rectangle 50 60 'solid 'brown)
                        300 300
                        (rectangle 500 500 'solid 'blue))))

(equal? si si2)
(equal? si si3)
(equal? si2 si3)

reports

false
true
false

See if we can make REPL receive drop events from CMB

I followed the README and believe I have all the required dependencies that were listed, but I can't get the blocks-test branch to run.

From the root directory of the repo, I ran ./build-console-and-editor.rkt, which seemed to work fine. However, when I run ant runserver, I get the following error:

Buildfile: /home/pcardune/programming/bootstrapworld/wescheme/build.xml

copyjars:

BUILD FAILED
/home/pcardune/programming/bootstrapworld/wescheme/build.xml:28: /home/pcardune/programming/bootstrapworld/wescheme/lib/google-oauth-java-client/libs does not exist.

Total time: 0 seconds

I tried running git clean -f -x to remove any build files and start from a clean slate and got:

Buildfile: /home/pcardune/programming/bootstrapworld/wescheme/build.xml

copyjars:
     [copy] Copying 11 files to /home/pcardune/programming/bootstrapworld/wescheme/war/WEB-INF/lib

BUILD FAILED
/home/pcardune/programming/bootstrapworld/wescheme/build.xml:28: /home/pcardune/programming/bootstrapworld/wescheme/lib/google-oauth-java-client/libs does not exist.

Total time: 0 seconds

Which is essentially the same thing. Where am I supposed to get google-oauth-java-client from?

@schanzer this is the issue I was telling you about on the phone.

Cached Images aren't getting stored with the getAriaText result. See lines 484-488 of world.js.

Looks like it never worked before, but nobody ever ran into it!

Reported by @maueroats in the Racket Discord

(values a)
(define a 1)

evaluates to 1. In Racket/Scheme, a wouldn't be initialized yet.

What needs to happen to port @dyoo 's PLaneT package for "Simply Scheme" to WeScheme as a teachpack?

It looks like more recent versions of the Closure Compiler are available as a node.js module, which could be called without ever using Python at all! This feels worlds easier than trying to get the old Python2 codebase working on Python3, and it’s always a good thing to get rid of dependencies…especially when they involve loading an entire programming language! Python is used in two places in the build script:
1) To calculate dependencies for the compiler.
2) To calculate dependencies for the closure library.

Both of these should be replaced by the newer and better-maintained JS version of the same script. If you can pull this off, we can remove Python entirely! Don’t forget to edit the README file to reflect this.

CM-block entries should plug into a pre-existing announcements node.

From bootstrapworld/curriculum#1626:

Just noticing that Pyret can take in either "dark-green" or "darkgreen" and WeScheme only accepts "darkgreen". This means that we as authors (me!) sometimes write WeScheme examples that wouldn't actually work. If it would be easy to make WeScheme also accept hyphenated versions of those strings it would make authoring less error prone.

Following two tests fail:

(check-expect (image-width (rectangle 0 10 "solid" "purple"))
              0)

(check-expect (image-height (rectangle 10 0 "solid" "purple"))
              0)

The result is:

check-expect: actual value 1 differs from 0, the expected value.
check-expect: actual value 1 differs from 0, the expected value.

Modify the java (found in src/) so that it works on at least Java 11. If you want to really knock this out of the park, Java 17 would be even better. Given that you’re got plenty of Java experience, my hope is that this one isn’t going to be much trouble either.

EXAMPLE failures are not sufficiently visible. Need CSS love to make them pop in the REPL, but also some more serious work to have it visible in running bing-bang

Using #; results in incorrect highlighting, which causes students to believe something is wrong when it is not.

Examples from using wescheme.org.

1. (Confusing.) Example with #; - my first line is purpleish (comment color). Actually both lines should be hilighted as commennts.
   

2. Block comments have the same highlighting as code. (This might not be fine.)

Mariana Hernandez would love to have this for her kids.

The documentation link in WeScheme has a reference to a color database that points to http://www.wescheme.org/draw/color-database___.html, which is a dead link.

The link is in this section:

http://www.wescheme.org/doc/wescheme.html#%28def._%28%28lib._2htdp%2Fimage..rkt%29._image-color~3f%29%29

• hyphenated color names
• pinhole and bounding box calculation

Animate function seems to "jump" over the beginning of the animation, e.g. it is not showing the first 200 (?) steps. This flower should start from 1 pixel size and grow, now it starts in the middle...??
http://www.wescheme.org/view?publicId=PLX4vs7DcR

Back in 2013, I used toDataURL for comparing image tiles. But toDataURL has no guarantees of being lossless, causing the equality function to occasionally produce false negatives.

getImageData is now available, and supported in all browsers. We should switch to that instead, and check to see if tiling is even needed anymore.

Error message from Google indicates that users in a "Google Workspace for Education" (e.g., school org) under 18 are blocked, as well as children under 13.

As of 2021-10-04, changed from reported to work on 2021-10-03.

From Tiina:
Missing Scandinavian letter support in interaction window is still an issue (ä, ö). This comes up in every time I'm giving a Racket course for Finnish teachers... (last time last Tuesday…).

Ctrl-' should read the previous one

Requests that use HTTPS (https://www.wescheme.org) should work.

Technical reason: it aids composability in the long run. Other pages that may embed wescheme.org will not be able to do so at the moment if they themselves are HTTPS, as browsers such as Chrome prohibit a "downgrade" from HTTPS to HTTP. For example, let's say that https://hashcollision.org has a IFRAME that tries to embed wescheme.org. It can't at the moment because wescheme.org doesn't currently support HTTPS.

From Shriram:

• Do we really need to be using window.location for every button?
• Can we remove the Session constructors that we aren't using?
• Can we do all the session-setting via AJAX, and minimize page redirects?

Also see bootstrapworld/curriculum#1747

Finish backporting from brownplt/code.pyret.org@0cd2d5c

This will kill off IE9 support, but it's worth doing

"If you don't see output above, it may be just definitions."

Try to access: http://www.wescheme.org/view?publicid=peppy-amend-money-grimy-owner

results in 500 server error

The correct url is http://www.wescheme.org/view?publicId=peppy-amend-money-grimy-owner

Either the error should be more helpful, or we create a redirect

I follow README and try to run wescheme locally but can not succeed.

I check dependencies, seems that all are ok.
I installed it by respect to defined 3 steps in Readme with succeed

Then
$ ant runserver
finished with succeed and google.appengine related exceptions.

But when i request to localhost there is no response browser can not connect to server.

When running ./build-console-and-editor.rkt on the npm branch, the console prints warnings related to unnecessary escape characters in regex expressions.

For example:

WARNING in .../wescheme/war-src/js/js-runtime/js-numbers.js at 2416, 24: Unnecessary escape: '\.' is equivalent to just '.'
WARNING in.../wescheme/war-src/js/js-runtime/js-numbers.js at 2416, 31: Unnecessary escape: '\.' is equivalent to just '.'
WARNING in .../wescheme/war-src/js/js-runtime/js-numbers.js at 2416, 38: Unnecessary escape: '\.' is equivalent to just '.'
WARNING in .../wescheme/war-src/js/js-runtime/js-numbers.js at 2416, 45: Unnecessary escape: '\.' is equivalent to just '.'

Given my limited experience w/ Regexes, I am having trouble discerning whether or not the escape characters are in fact necessary. To the best of my knowledge, the . character must be escaped but the warning suggests otherwise.

Warnings are produced by the following files:

• .../wescheme/war-src/js/js-runtime/js-numbers.js
• .../wescheme/war-src/js/openEditor/tokenizer.js EDIT: Addressed in 3de5b6b (needs testing)
• ...wescheme/war-src/js/compiler/lex.js EDIT: Addressed in 57087d4