Jump to content




ComputerCraft Beta Versions - Bug Reports


72 replies to this topic

#41 jrockjake

  • New Members
  • 2 posts

Posted 27 September 2014 - 06:50 AM

VERSION:
1.6.4pr6

Description:
Entering any command (Dance or excavate) on a turtle crashes the game.

Expected results:
Dance makes the turtle boogie. Excavate enters Creative Mode Mining Mode.

Reproduction steps:
1. Place turtle down
2. Open turtle GUI
3. Enter command
4. Crash
5. Repeat

More Info:
http://pastebin.com/zvQZcPWQ


[This bug is a duplicate of these other reports. -L]

Edited by Lyqyd, 27 September 2014 - 06:55 AM.


#42 dan200

  • Administrators
  • 542 posts
  • LocationCambridge, England

Posted 31 January 2015 - 08:30 AM

Re-opening this topic for bug reports in CC 1.66.
As always, remember to stick to the format in the OP, and check if your bug has already been posted.

#43 Cycomantis

  • Members
  • 60 posts

Posted 31 January 2015 - 10:20 AM

VERSION:
1.66pr3
DESCRIPTION:
Ran the exec program on Basic and Advanced Computers
EXPECTED RESULT:
The program not to run as these are not Command Computers
REPRODUCTION STEPS:
Place a Basic or Advanced Computer and enter "/rom/programs/command/exec give yourname diamond 64


Whoops! Fixed for pr4. Computers created in pr3 may remain broken, but you shouldn't move save games between beta versions.

#44 Xukkorz

  • New Members
  • 2 posts

Posted 05 February 2015 - 02:31 PM

Upon updating to ComputerCraft Beta computers no longer start while ModTweaker3 is also installed.

VERSION:
ComputerCraft 1.66pr3

ADDITIONAL MODS:

MineTweaker 3.0.9c

DESCRIPTION:
Upon updating to Computercraft Beta computers no longer start.
EXPECTED RESULT:
Computers should boot when right clicking
REPRODUCTION STEPS:
Place the computer and right click, no errors occur. Included is http://paste.ee/p/EpW0w of minecraft log.
NOTES:
I am not sure if this is MineTweaker3 or if it is ComputerCraft. I am reporting on both sides. Note this appears to be a problem between these specific versions of the mods as if I remove MineTweaker ComputerCraft works and if I install the previous version of ComputerCraft. I tend to believe this is because ComputerCraft is now able to send Minecraft Commands directly rather then interfacing with a Command Block.


On investigation, this is a MineTweaker bug, i've reported it here: https://github.com/s...ker3/issues/154
However, i've updated ComputerCraft to handle the bug more gracefully.

#45 dan200

  • Administrators
  • 542 posts
  • LocationCambridge, England

Posted 31 March 2015 - 09:42 AM

Reopened for 1.74!

#46 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 31 March 2015 - 10:33 AM

VERSION:
ComputerCraft 1.74pr13, MC 1.7.10, Forge 10.13.1.1219, no other mods.

DESCRIPTION:
Crash when attempting launch. Updating to the latest Forge, 10.13.2.1343, has not helped, nor did going to Java 1.8.0_40.


whoops! that wasn't supposed to be in there. Fixed in pr14

Edited by dan200, 31 March 2015 - 12:32 PM.


#47 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 02 April 2015 - 12:29 PM

VERSION:

CC1.74pr16

DESCRIPTION:

Attempting to term.blit() a string off the edge of a window errors:

window:193: bios:7: bad argument: string expected, got nil

EXPECTED RESULT:

Auto-cropping of the line.

REPRODUCTION STEPS:

While redirecting to a window:

term.blit(string.rep("a",60), string.rep("b", 60), string.rep("c", 60))

Relevant code block:

					elseif nEnd > nWidth then
						sClippedText = string.sub( sText, 1, nWidth - nStart + 1 )
						sClippedTextColor = string.sub( sClippedTextColor, 1, nWidth - nStart + 1 )
						sClippedBackgroundColor = string.sub( sClippedBackgroundColor, 1, nWidth - nStart + 1 )

Presumably you wanted to substring "sTextColor" and "sBackgroundColor".


Fixed in pr17

-------------------------------------

DESCRIPTION:

term.blit() allows use of non-shades on non-advanced computers.

EXPECTED RESULT:

An error indicating the colours are invalid.

REPRODUCTION STEPS:

On a non-advanced computer;

term.blit("a", "b", "c")


Fixed in pr17. I don't emit an error (checking for this might be slow, and blit is meant to be fast, like a memcpy), the invalid colours simple fallback to black when displayed (the same as what happens if you try to blit with a character out of the hexadecimal range).

Edited by dan200, 02 April 2015 - 02:25 PM.


#48 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 02 April 2015 - 11:52 PM

VERSION:

CC1.74pr16

DESCRIPTION:

Turtles fail to boot after exiting a world, until such time as MineCraft is restarted. Exiting to the main menu, then entering the same or a different world without closing and re-opening the whole client, renders all turtles (newly placed or pre-existing) unusable.

Furthermore, in one instance I was able to get all my turtles to boot without access to the turtle API. I've been unable to re-produce this, however, and suspect it has something to do with one of the other mods I had installed at the time.

Edit: Scratch that, the turtle-table thing is still happening with CC alone. Still trying to figure out the requirements to trigger it.

Edit 2: It seems that both issues can be triggered by placing and breaking turtles - place a bunch of them (some normal, some advanced), then boot a random one, break a random one, boot a random one, maybe boot another random one, etc. The boot issue is far more likely to occur first, and is presumably triggered when a certain turtle shuts down.

I suspect this has something to do with mismatched IDs. At one stage, I had a normal turtle boot up as if it were advanced, with full colour support. After restarting the whole game, that turtle had a different ID.


Fixed in pr17

Edited by dan200, 03 April 2015 - 10:39 AM.


#49 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 03 April 2015 - 02:38 AM

VERSION:

CC 1.74pr16

DESCRIPTION:

Auto-complete may interfere with command history in shell / lua console.

EXPECTED RESULT:

History to disable auto-complete until such time as a character is typed.

REPRODUCTION STEPS:

In shell, try entering:

clear
help

You'll then be unable to use the up arrow to access the "clear" command, as the auto-complete functionality takes over before you get there.

In the Lua console:

a = fs.open("somefile.txt","w")
a

Again, you'll be unable to reach the first command via history.


Fixed in pr18

Edited by dan200, 03 April 2015 - 11:42 AM.


#50 Wojbie

  • Members
  • 630 posts
  • LocationKinda lost

Posted 03 April 2015 - 10:59 PM

1)
VERSION:

CC 1.74pr18

DESCRIPTION:

Duplicate line in /rom/startup file.

EXPECTED RESULT:

No duplicated lines.

REPRODUCTION STEPS:

Look at lines 188 and 189 in /rom/startup file.

2)

VERSION:

CC 1.74pr18

DESCRIPTION:

I believe wrong variable is used in error message in program edit line 383

EXPECTED RESULT:

Message "Error saving to "..sTempPath

REPRODUCTION STEPS:

Look at line383 in edit program. It states
sStatus="Error saving to "..sPath
but its error to saving to file placed under
sTempPath


Both fxied in next version

Edited by dan200, 09 April 2015 - 11:36 AM.


#51 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 04 April 2015 - 12:04 AM

VERSION:

CC 1.74pr18

DESCRIPTION:

Edit's "Run" function generates a new tab displaying "No such program" when attempting to use is anywhere other than the root of a system's drive.

EXPECTED RESULT:

Script to run.

REPRODUCTION STEPS:

cd <some folder>
edit <some file>
Ctrl => Run

Fault appears to be in shell.openTab; suggest changing shell line 299 to:

return multishell.launch( tEnv, "rom/programs/shell", sCommand, unpack( tWords, 2 ) )

... or skipping the use of shell.resolveProgram() in that function, as it's called later along the stack in run().


Fixed in next version

#52 Anavrins

  • Members
  • 775 posts

Posted 06 April 2015 - 10:51 AM

VERSION:
1.74pr18

DESCRIPTION:
Any two byte characters (é, è, ê, ë, ç, ö) etc... cause any program to crash with "window:78: Arguments must be the same lenght" or actually line 157.

EXPECTED RESULT:
The character to show up.

REPRODUCTION STEPS:
Enter any two byte characters in the shell or in any program that would cause those characters to appear,
if your keyboard doesn't allow you, copy and pasting "é" works.

Edited by Anavrins, 06 April 2015 - 02:03 PM.


#53 CrazedProgrammer

  • Members
  • 494 posts
  • LocationWageningen, The Netherlands

Posted 06 April 2015 - 11:07 AM

VERSION:
ComputerCraft 1.74pr18
DESCRIPTION:
It seems very strange, but term.blit("test", "0000", "eeee") draws test with black text and an orange background.
The colors API tells that 0 is white, and e is red.
This means that the colors are inverted. (0 -> f - 0 = f, e -> f - e = 1)
This seems to happen on all computers (command computer and advanced computer/turtle/pocket computer and normal computer/turtle/pocket computer)
and all monitors (advanced and normal).
On the black/white displays I tested it with term.blit("test", "0000", "7777") and that draws test with black text and a light gray background. (which again is inverted)
EXPECTED RESULT:
It should draw test with white text and a red background.
REPRODUCTION STEPS:
Go to the lua prompt and simply type:
term.blit("test", "0000", "eeee")
SCREENSHOT/VIDEO:
Posted Image



Fixed in next version

#54 Wojbie

  • Members
  • 630 posts
  • LocationKinda lost

Posted 09 April 2015 - 02:35 PM

VERSION:
1.74pr20
DESCRIPTION:
Mistake in default auto-completion for chat program.
EXPECTED RESULT:
Options {"host ", "join "} for chat
REPRODUCTION STEPS:
in line 160 in startup file there is 3th option "run" that is not a valid argument for chat program.

(sorry i am reporting lua side bugs only but i can't get stuff to crash on me.)


Fixed

Edited by dan200, 10 April 2015 - 12:14 AM.


#55 SeaLife

  • Members
  • 23 posts

Posted 11 April 2015 - 08:25 PM

VERSION:
1.74pr14 to 1.74pr20 (Tested)
DESCRIPTION:
Mistake in http api - Cant get responseCode from all headers (Not the 200 OK)
EXPECTED RESULT:
h.getResponseCode() on 404 Pages should return 404

(Same with 401, 403, ...)

REPRODUCTION STEPS:

h = http.get("http://hdjawhdjwadw.de/file.txt")
print(h.getResponseCode())


>> lua:2: attempt to index ? (a nil value)


Code Demonstration
local function downloadFile( url )
	local h = http.get( url )
	local httpCode = h.getResponseCode()
	local content = ""
	if h then
		content = h.readAll()
	else
		error("Server Returns HTTP STATUS: " .. httpCode )
	end

	return content
end

downloadFile( "http://google.de/doesnotexistsfile.lua" ) -- will crash
downloadFile( "http://computercraft.info" ) -- returns the sourcecode of the front page of cc.info (if previous line is commented)

Reason of Report
Im trying to make an "GitLab" Installer...
This Installer should base on GitLab's API http://doc.gitlab.co...api/README.html

I cant get messages like 401 and 404 and this is bad.. i cant tell the User that the requested Project does not exists or that the User is not permitted to download the Projekt..


Not a bug. A missing feature maybe, but not a bug

Edited by dan200, 13 April 2015 - 09:53 AM.


#56 _removed

  • Members
  • 262 posts

Posted 12 April 2015 - 12:17 PM

VERSION:
1.7.4pr20
DESCRIPTION:
Placing a door with command computers place an invisible door.
EXPECTED RESULT:
A normal door to appear.
REPRODUCTION STEPS:
commands.setBlock(x, y, z, "minecraft:wooden_door") -- Errors on any type of door, using wood for this example.
SCREENSHOT/VIDEO


Not a bug. THis behaves exactly the same as typing the same command manually or using a Command Block.

Edited by dan200, 13 April 2015 - 06:17 AM.


#57 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 14 April 2015 - 01:32 PM

VERSION:

1.74pr20

DESCRIPTION:

The placing of more than two consecutive periods into a file-system path is mishandled.

EXPECTED RESULT:

"...." and so forth to be treated as ".".

REPRODUCTION STEPS:

Eg, fs.list(".....") produces:

Java Exception Thrown:
java.lang.NullPointerException

This has been an issue for many builds, but seems to matter a bit more with auto-complete in place as entering "cd ..." into the shell (then trying to type anything further) leads to the whole system shutting down as a result of it.


Dan: This produces "not a directory" here

BB: You're just poking fun at me for running Windows, aren't you. :P

But just on the off chance that you really don't see the problem, in the land of MS all file/directory names composed of dots-only are treated as "special" - not just "." and ".." - and the fs API's behaviour mirrors that of the filesystem the MineCraft server is running on. Under NTFS, shell.resolve("...") should return the same as shell.resolve(".") does.

Edited by Bomb Bloke, 21 May 2015 - 11:51 AM.


#58 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 17 April 2015 - 01:15 PM

VERSION:

1.74pr20

DESCRIPTION:

Going up / down through read()'s auto-complete options, then tapping right to select one, may stop short and switch to a different suggestion.

EXPECTED RESULT:

If a particular suggestion was specifically chosen, it should be used.

REPRODUCTION STEPS:

In the Lua prompt, try typing:

os.pullEv

If you tap up to the "os.pullEventRaw(" suggestion then tap right, you'll end up with "os.pullEvent" and the suggestion will change to just "(".

Suggest rigging up a boolean variable to keep track as to whether the user has actually used up/down to pick a particular suggestion, and skipping the prefix-checking code in acceptCompletion() if so (simply adding tCompletions[ nCompletion ] to the line instead).

#59 Goof

  • Members
  • 751 posts

Posted 27 April 2015 - 09:17 PM

Version 1.7.4pr20 SMP and SSP

Note: This was tested on both a multiplayer server, AND in singleplayer

Description:

Im working on a server and played around with a "boot computer" (computer which turns on all computers on the wired network)

After some hours, i left ( chunks unloaded ), and joined later.
When i ran the boot computer, the other computers on the network didnt respond to the
peripheral.call(v, 'turnOn')
An easy fix is reconnecting the cables to the computer.

Expected "result"
All cabled computers should turn on

Reproduction:

Place 2-5 computers with startup code (print something, to clarify it started)
Connect everything up, and make a "booter" with the following code:
local Cable = peripheral.wrap('SIDE')
for k, v in pairs(Cable.getNamesRemote('SIDE')) do
  if Cable.getTypeRemote(v) == 'computer' then
	print('Booting ' .. v)
	--# For testing purposes, i tried "rebooting" the computers instead, but no difference
	--#peripheral.call(v, 'shutdown')
	--#sleep(.05)
	peripheral.call(v, 'turnOn')
  end
end
Close minecraft/unload the chunks.
Reenter the chunks, and run the boot computer.
Done. The cabled computers did NOT turn on.

Edited by Mikk809h, 27 April 2015 - 09:25 PM.


#60 Bomb Bloke

    Hobbyist Coder

  • Moderators
  • 7,067 posts
  • LocationTasmania (AU)

Posted 29 April 2015 - 03:36 AM

Re the above, under 1.74pr20 I've been encountering instances where systems won't shut down. When they attempt to restart (due to a bad crash, or even Ctrl + R), they instead hang on whatever they were last displaying; Ctrl + S/R has no effect at this point, nor does a remote peripheral call ordering them to turn off. Breaking / replacing the computer, or otherwise unloading the chunk they reside in, seems to sort it. I assume it boils down to another symptom of the exact same bug.

Anyway;

VERSION:

1.74pr20

DESCRIPTION:

Crafting an advanced monitor, using eight gold nuggets and one glass pane, returns one advanced monitor.

EXPECTED RESULT:

Four advanced monitors.

Not too sure when this change was introduced. I'm seeing it in every CC build that's available for MC 1.7.10. It may've been intentional; but if so, it makes advanced monitors very, very expensive.


Fixed. Not sure how this regression was introduced





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users