ghp reports-
On a script element, rightclick mouse to get the ‘help/duplicate/delete-Popup.
In win-scratch-1.4, the popup appears.
In beta-scratch, in addition to the popup there is a bubble popup ‘Enter text to narrow selection down to matching items’.

In some situations, sporadic, not well reproduceable, the programs exits.
'Exit' means: window closed, process is away, no error popups, no dialogs or alike.
Found this when create variable (global), create same name again (local), you get a dialog: variable already exists, when pressing ok then scratch exits.
Had this also on startup, when remote sensor network is enabled, you get a dialog like 'sensor network enabled', when pressing ok then scratch dies.
This is happening not each time. Looked for logfiles or dumps, but could not find any.

This is a bit of code that illustrates so many bits of bad programming practice in Smalltalk that it's seriously embarrassing.

Need refactoring someday

It seems that input events are either missed occasionally or get very delayed. Users often have to click twice or more to get a button to respond. D-Clicks frequently un-noticed.

Trying to open the sound recorder fails in the initialise method.

Choose a sprite, select the Sounds tab, click on 'Record'.

ghp reports -
Let a sprite ‘say ‘hello”.
In raspbian-scratch-1.4, the text is prox double the size as in beta-1.4.

ghp reports -
I have a scratchClient (remote sensor protocol), sending infrared control signals as string like ’11000000000011′. I convert this to a numeric value in scratch. To doublecheck, the following sample uses prefix AB to ensure the value is a string type. The quotes in the folling sample are only for clarification, not enterd in scratch.

create two variables a, b
set a=’AB11111111111111111111111111111111111′
set b=0
if letter(14 of a) = ’1′
change b by 1

in raspbian-scratch-1.4, the result is b==1
in beta-scratch, result is b=0. When I replace the if-statement by ‘if 1=1′, the increment is performed. Seems to be a char to int autoconversion problem.

Intermittently - how I hate that - the morph 'readout pane' is drawn over by the #clearLibrary method (typically when opening a new project, though occasionally when changing view mode) and subsequently they x/y value displays appear to be floating on their own instead of attached to the curvy frame.

The problem seems like it must be something to do with damage reporting. Dragging a menu (say, the language list) over the are will correctly redraw damaged areas. Altering the window size will correct it. Opeing the halo will correct it. It doesn't happen in a development situation where the Scratch pane is not filling the screen. It doesn't always happen; in fact it sometimes feels like there is a random small number being secretly generated that is well sized to create maximum debugging frustration.

I can't find a solution right now. It's not a huge issue functionally and may not even appear for some users. Grrr.

szenario: script on stage, Duplicate last block of script and add to end of same script. The block is added, but very low height, no text visible. It can be dragged away and added somewhere else.

Some users find the default font sizes a poor choice. IT might be practical to add a new keyword to the .ini file to set a scale factor that would have to be composted with the language file RenderScale value.

bhp reports -
'in some places I have seen some squeak error popup.
Scenario:
have a list with data, export to file system (works)
export again to same file name, there is a error popup
http://www.raspberrypi.org/forums/download/file.php?id=6978

The painter wants to grid the cursor location to handle the scaling whilst painting. OldScratch did nasty things to HandMorph. Should try to do nicer.
Old notes from Evernote stack -
PaintCanvas wants to keep the pointer position on a grid if the scale is > 1.
See PaintCanvas>mouseEnter:, mouseLeave: & exitPaintEditor, plus old Scratch code in HandMorph.
See also HandMorph>gridTo:origin:
How do we do it now? Maybe in PaintCanvas>step instead of gridPointRaw we can put the grid on the canvas and extend screenToCanvas: to grid things. That would remove any need for grid related methods in HandMorph> ‘Scratch-extensions'

ghp points out the the pingpong script attached should make the watcher for 't' visible once 'i' reaches 200 but that in presentation mode it doesn't.
The really odd bits is that if the watcher for 'i' is left visible before starting, the 't' watcher does indeed become visible!

create a multiline text content like
"Treppenhausautomat
Lichtschalter
Automat
Kontrollzentrum"

Problem 1: editing the text is difficult. Cursor does not move down to second, third line.
Problem 2: only Line 1 is displayed on stage.

When loading a sprite, size in px and kB is displayed correctly. When editing the sprite, the display gets scrambled.
There wasa similiar problem in earlier releases, but there the display was scrambled always.

Starting NuScratch beta 1 is slow

open costume editor
expected: the zoom level is second level; sliders are prox 2/3 of screen size, centered
current: sliders are prox 1/10 of screen size, positioned left/top

Problem: slider size is not adjusting with zoom level and initial viewport.

Old Scratch used some nasty hacking of HandMorph to scale the dragged sprite within the HandMorph code. Better method needed

The ScratchNoteSelectorMorph is very out of date wrt events. Some pretty serious reworking will be needed to clean it up.

See ScratchNoteSelector, particularly #startUp and compare with the PianoKeyboardMorph.
One problem is that the selector is expected to behave like a popup menu; be opened with a click on the ExpressionArgWithMenuMoprh, only start working after the mouse button is released, then act modally. It simply isn't doing that right now.

Simple fudge to get it functioning at all is to remove the 'Sensor waitNoButton' from #startUp so that it at least doesn't bugger the event queue.

Strings and numbers are casually thrown around and expensive conversions made in too many places. Improvements needed. In Asteroids we see about 15% of time spent on this.

Had a nice project with some sprites. Imported this to an almost empty project. Import completed, but sprites in 'object inventory' had only small white rectangles displayed. When I pressed 'file open' (wanted to load project directly to validate the sprites), the images got refreshed and displayed correctly.

have a variable displayed on stage; switch to presentation mode; rightclick on variable --> a popup appears. "watchermorph-->does not understand."
Also Mouse middle click brings up the 'fancy bubbles'.

bhp reports -
'import sounds, Music Loops, Eggs gives a MPEG-Error.
http://www.raspberrypi.org/forums/download/file.php?id=6985
Sound is 'playable' then, when I choose proceed.

ghp reports -
Menu help/help page does not work.
In raspbian-scratch-1.4, dillo appears with help content.
In beta-scratch, nothing happens.

Unpleasantly convoluted code; extra mess from apparently relying upon whether results from #optionList are 'unicode'.

create a project with a variable, display variable on stage, choose 'large readout'.
save project
open project again
the variable display is not 'large', but the default name/value display.

 Richard Hayler says: 26th May 2014 at 11:01 pm

There’s a minor bug when you use a slider to set a variable: when set to a non-zero value and you click on the slider knob, it resets the variable straight back to 0.

...
Nope, the current version of Scratch 1.4 for Raspbian, the Windows and Mac versions do not behave in this way. What I’m talking about is when you’ve already set a variable to a value using the slider (say 15) and then you click on the slider knob to change it. With other versions of Scratch you can adjust the value in either direction (e.g to 14 or 16). With this new beta, the value jumps back to zero.

tim noted -
Yeah, I see what you mean. You can drag the slider ok but next time you drag it, it jumps back to 0. Never noticed that before. Good catch.

The menu listing available languages is far too long and currently doesn't scroll, split or otherwise accommodate itself to the screen,
Derived from
#2

by ghp » Thu May 01, 2014 12:08 pm
bug or feature ?

Hello,
starting from a thread viewtopic.php?f=77&t=76056 I did some examinations on the handling of code pages in remote sensor protocol.

I send a 'broadcast "test_äöüÄÖÜß"' from remote to scratch which uses popular german umlaut characters. Basic is the encoding nneded which is utf-8 for these chars. In scratch, these chars are not decoded correctly, but same in windows scratch 1.4.
I double checked the bytes sent on the socket by a simple socket server implementation.

The scratch wiki page says:
Words and strings are encoded in UTF-8. ...
Here are the two most useful message types:
broadcast

The other way round from scratch to scratchClient, the special chars are encoded correctly.

SimpleSi reports -
'simplesi wrote:
File open asked me if I wanted to save current project when I hadn’t done anything at all

Also I had to click File then Open then Open again to get it work

Tracking the 'dirty' flag is fairly tricky to get exactly right and it's probably a bit enthusiastic right now. Probably better than missing and throwing away a project?'

ghp reports-
drag an operator (keep it simple, take (1+1)) to script area.
When doubleclicking aka executing this operator (without a ‘hat’), the result is displayed in a bubble popup.
In win-1.4-scratch, move the operator and the popup disappears.
In beta-scratch, move the operator and the popup does not disappear.

sensor values are displayed differently on stage, they have different color and label.
The pattern is NN.N in legacy scratch, one decimal place always.
In nuscratch8, the pattern falls back to NN when values have only zeroes after digit, e.g. '19.0'

Can't see any particularly good reasons why the HandMorph toolType stuff is there. Surely the main functionality belongs in ScratchFrameMorph those days?

Complex mechanism for firing up scripts.
Events with string names are passed around and strings compared (sigh)
Lots of stupid isKindOf: nonsense to find relevant tiles to check for scripts.

See ScratchStageMorph>broadcastEventNamed:with:

ScratchSpriteMorph>eventReceived: :-(

ghp reports-
Language selection display.
In Win-scratch-1.4, the language selection menu displays complete language names like ‘Deutsch’ or ‘Francais’
In beta-1.4, it is ‘de’ and ‘fr’. Only English is complete.

ghp wrote:
The character set problem (german umlaut), the error popups now and then, the 'letter i from text' conversion and not alpha backgrounds should be fixed till then, all the others I have seen here are lower severity.

Character set issues are probably some problems with unicode handling, which is always such fun.
alpha backgrounds? That doesn't seem familiar?

Load a project
Import a project

The inventory has sliders bottom and right, and the imported sprites is displayed as a white rectangle. (not correct)
The costumes of this 'fancy thing' are correct.
When clicking into the window, the sliders disappear (correct)
When e.g. editing this costume, leaving editor, the sprite inventorty is updated correctly.

Still wants to save an empty script before loading new one

Can't doubleclick to open a folder

(I do all my work remoting in using VNC but not noticed this behaviour with standard Scratch)

SimpleSi reports -
'Tried using scissors tool on a wait block to remove it from the middle of set of blocks - nothing happened - CPU load went up to 100% had to restart NuScratch'

ghp reports -
'Minor problem:
In 'control' there are the 'when I receive' blocks.
Select one and drag to script area.
open the popup, usually populated with 'new' and available signal names. If you move the mouse over a name or 'new', the entry is highlighted. There is a small gap between the highlighted ares. Drag the popup in the gaps (also possible in the frame around it), and you can move it away. In scratch 1.4, clicking in the gap areas closes the popup.
'

Scenario: in a loop, I calculate float values in a global variable and send them to scratchClient.

Instead of receiving decimal values, I receive strings like "(121856/5)".
Works in 1.4-scratch.
Tested with nuscratch5.

There was a similiar problem with the display of variables on stage in early days of nuscratch.

Seems to not detect a lot of mouse-clicks (missing at least 1 in 3) which might explain the double-click failure as well
(I do all my work remoting in using VNC but not noticed this behaviour with standard Scratch)

ghp reports -
Import sprite, things, banana.
win-scratch-1.4, the background is alpha.
beta-scratch-1.4, the background is not ‘alpha’, but white.
Clearly visible if you have a non-white stage.
Also in object palette, the background is displayed white.

Since jan 14 the v4linux support has been there (don't forget
sudo apt-get update && sudo apt-get install v4l-utils
sudo apt-get update && sudo apt-get upgrade
modprobe bcm2835 v4l2
)
but the colour map seems to swap green and blue or whatever and everyone looks like a Smurf. Not a happy sight

We need the Unicode plugin set up and installed and used

Somewhere the rendering of commandmorphs got broken wrt ScratchTranslator>>isRTL. It's present in the current Pi release as well as NuScratch so it almost certainly came from work to speedup morph drawing & the cacheing of the results.

following up from the other event menu report - adding a new event name works but adding a second one seem to delete the previous.

ghp reports-
weird behaviour in both scratch versions.

create variable, display on stage. Expected is to have at most 1 decimal place displayed.
So if you set a=1, and then repeat 5 times, a = (a / 2) the display is ’0′.
But if you run a block set a = 0.0312, it is displayed as 0.0312 on stage. Cool.

The NuScratch beta VM has been changed from the alpha VM to use a thread based ticker instead of the timer interrupt ticker. The interrupts interact with some kernel problem and with the ASLA (mis)use of interrupts to cause total lock-ups when trying to play a sound.

One use reported a problem when testing -
'pthread_setschedparam failed; consider using ITIMER_HEARTBEAT: Operation not permitted'