You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Original bug ID: 6448 Reporter: gerd Assigned to:@damiendoligez Status: resolved (set by @damiendoligez on 2016-04-05T14:03:56Z) Resolution: suspended Severity: minor Version: 4.00.1 Category: back end (clambda to assembly) Monitored by:@gasche
Bug description
The attached program bug.ml outputs different numbers on different runs (for 3.12.1, 4.00.1 but not 4.01.0):
RUN 1
r=0
r=1
r=2
RUN 2
r=0
r=3
r=4
However, there is no mutable state outside RES, and both runs should produce identical numbers.
It looks like as if the old instance of the module RES is cached while building the class. The double definition of U (in Genutil) is also crucial for triggering the bug.
In 4.01.0 the program works as expected; nevertheless I'm reporting the issue because it is so strange, and there is the chance that the bug is now only better hidden. At least I could not find anything in the Changes file or in Mantis that could explain the misbehavior.
Original bug ID: 6448
Reporter: gerd
Assigned to: @damiendoligez
Status: resolved (set by @damiendoligez on 2016-04-05T14:03:56Z)
Resolution: suspended
Severity: minor
Version: 4.00.1
Category: back end (clambda to assembly)
Monitored by: @gasche
Bug description
The attached program bug.ml outputs different numbers on different runs (for 3.12.1, 4.00.1 but not 4.01.0):
RUN 1
r=0
r=1
r=2
RUN 2
r=0
r=3
r=4
However, there is no mutable state outside RES, and both runs should produce identical numbers.
It looks like as if the old instance of the module RES is cached while building the class. The double definition of U (in Genutil) is also crucial for triggering the bug.
In 4.01.0 the program works as expected; nevertheless I'm reporting the issue because it is so strange, and there is the chance that the bug is now only better hidden. At least I could not find anything in the Changes file or in Mantis that could explain the misbehavior.
File attachments
The text was updated successfully, but these errors were encountered: