Unable to visualize a large table when read is preceded by a long sleep

[hubertp]

Consider a simple program

from Standard.Base import all
from Standard.Visualization import all
polyglot java import java.lang.Thread

main =
    text1 = 'Hello world! Works?'
    integer2 = text1.length
    Thread.sleep 10000
    node1 = enso_project.data/'customers_2024.csv' . read
    42

where the data file is irrelevant apart from the fact that it is ~80MB. The order of expressions matters.
This results in a failure to visualize read data:

[JaroslavTulach]

• Thread.sleep(...) is blocking operation from a point of view of safe points
• No safe points can happen while Thread.sleep is sleeping
• Rely on ThreadManager when interrupting Enso execution #12655 - cannot interrupt Thread.sleep

[JaroslavTulach]

• there is Runtime.sleep since Add Data.fetch_many and Runtime.sleep #13999
• it however suffers from the same problem as the example in the description of this issue
• proper sleep needs to be done as in https://github.com/enso-org/enso/pull/13907/files#diff-ae8a122ef71fea9237cd5644f7db76aa6c7d139d8104acdc859a47d1b01ab919R204
• e.g. with setBlockedThreadInterruptible

[JaroslavTulach]

where the data file is irrelevant apart

@hubertp, what is then relevant? Why Runtime.sleep 10000? If I use just 1000, then everything seems to work. What gets broken with 10s? Is there some 10s timeout in visualizations? Where is it? Will it be gone if we run visualizations in parallel?

[JaroslavTulach]

• The problem with Runtime.sleep 10000 is going to be solved by Progress of Runtime.sleep visualized #14275
• but there are many other ways to block JVM (Object.wait(10000)?)
• addressing those would require other changes