[I7-1981] [Mantis 2017] Pausing a glulx game twice before the first command prompt freezes the IDE

curiousdannii-testing commented 2 years ago

Reported by : njsbrown

Description :

If a glulx game includes Basic Screen Effects and contains two "pause the game" phrases that are both used before the first command prompt is printed, the IDE freezes when the player presses SPACE for the first time.

Specifically, the on-screen text is not cleared, a frozen cursor appears in the top left corner of the Story panel, and neither panel can be interacted with. The tabs become unresponsive. The IDE doesn't crash entirely and can be quit normally, but most things stop working.

I've waited up to 30 minutes, but the IDE doesn't recover.

Steps to reproduce :

Include Basic Screen Effects by Emily Short.

Example Location is a room.

When play begins:
    pause the game;
    say "This text does not appear.";
    pause the game.

Additional information :

For this bug to occur, the story format must be glulx, not z8.

Despite the fact that the IDE freezes after the first game pause, there must be two "pause the game" phrases for the bug to occur. If either is commented out, the game works properly.

Both phrases must be used before the first command prompt is printed. If "when play begins" is replaced with "after jumping", the game works properly.

The two phrases can be part of any rules that are followed before printing the first command prompt, including "when play begins", "after printing the banner text", and "after looking" (which gets triggered as part of the first room description). The two phrases do not need to be in the same rule.

This bug will also occur if only one phrase is present, if the player restarts the game. In this case, the IDE freezes after the player types Y to confirm the restart.

imported from: [Mantis 2017] Pausing a glulx game twice before the first command prompt freezes the IDE
status: Reported
resolution: Open
resolved: 2022-04-07T05:02:19+10:00
imported: 2022/01/10

curiousdannii-testing commented 2 years ago

557058:4c095ffd-6d6f-47ce-9e73-77c613347b86:

Comment by zarf :
I can't reproduce this. (I'm using the version of I7 6M62 (Mac) which is currently on the inform7.com web site.)

curiousdannii-testing commented 2 years ago

557058:4c095ffd-6d6f-47ce-9e73-77c613347b86:

Comment by zarf :
Yeah, I retried it on my 2016 MBP, and saw the problem.

The IDE/interpreter processes are not busy-spinning; it's just dead somewhere.

I suspect this is another long-standing CocoaGlk bug which just happens to show up on newer systems.

curiousdannii-testing commented 2 years ago

557058:4c095ffd-6d6f-47ce-9e73-77c613347b86:

Comment by zarf :
Replicated on recent Mac laptops. (Retina displays? Hard to say what the inciting factor is. Probably nothing obvious.)

curiousdannii-testing commented 2 years ago

557058:4c095ffd-6d6f-47ce-9e73-77c613347b86:

Comment by dfremont :
To add a data point, this doesn't happen on my Early 2015 MacBook Pro (which has a 13-inch Retina display), running macOS 10.12.5, with I7 6M62 from inform7.com.

curiousdannii-testing commented 2 years ago

557058:4c095ffd-6d6f-47ce-9e73-77c613347b86:

Comment by njsbrown :
It appears to be caused by something computer-specific. When I tried it on my older Mid 2011 MacBook Air, the bug doesn't occur. But it always occurs on my newer 2016 MacBook Pro.

Both computers are running the current version of I7 6M62 (Mac) from the inform7.com website. They are also both running macOS Sierra 10.12.5.

If there are settings I should check, or specs I should report, please let me know.

curiousdannii-testing / inform7-imported-bugs