Mantis Bug Tracker

View Issue Details Jump to Notes ] Issue History ] Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0004900OCamlOCaml generalpublic2009-10-26 08:382009-10-26 14:36
Reportershinwell 
Assigned To 
PrioritynormalSeverityminorReproducibilityalways
StatusclosedResolutionopen 
PlatformOSOS Version
Product Version3.11.1 
Target VersionFixed in Version3.11.2+dev 
Summary0004900: max_overhead vs. caml_gc_major / caml_gc_full_major
DescriptionI noticed this morning that gc_ctrl.c:test_and_compact, called from caml_gc_major and caml_gc_full_major, appears to be able to call caml_compact_heap (which doesn't test caml_percent_max) in the situation where the first calculation of fp yields >=1_000_000, caml_percent_max is set to 1_000_000, and there is more than one heap chunk.

Is the setting of caml_percent_max (aka max_overhead) to 1_000_000 not supposed to prevent automatic compaction in all cases? I wonder if the test in that function should be "fp > caml_percent_max", or whether 1_000_000 - 1 should be used for the saturating calculation.
TagsNo tags attached.
Attached Files

- Relationships

-  Notes
(0005142)
doligez (administrator)
2009-10-26 14:36

well spotted

- Issue History
Date Modified Username Field Change
2009-10-26 08:38 shinwell New Issue
2009-10-26 14:36 doligez Note Added: 0005142
2009-10-26 14:36 doligez Status new => closed
2009-10-26 14:36 doligez Fixed in Version => 3.11.2+dev


Copyright © 2000 - 2011 MantisBT Group
Powered by Mantis Bugtracker