aboutsummaryrefslogtreecommitdiff
path: root/TODO
diff options
context:
space:
mode:
authorArmin Rigo <arigo@tunes.org>2013-10-14 18:23:30 +0200
committerArmin Rigo <arigo@tunes.org>2013-10-14 18:23:30 +0200
commit595a43dda79d6600554ee36a9136fbad23f159ff (patch)
treef00436f3dbd75eeb1210d4bbeb50e29e62204912 /TODO
parent(fijal, arigo) A branch that implements incminimark which is (diff)
downloadpypy-595a43dda79d6600554ee36a9136fbad23f159ff.tar.gz
pypy-595a43dda79d6600554ee36a9136fbad23f159ff.tar.bz2
pypy-595a43dda79d6600554ee36a9136fbad23f159ff.zip
Kill the TODO file from 'incremental-gc'. Most things have been fixed
already.
Diffstat (limited to 'TODO')
-rw-r--r--TODO24
1 files changed, 0 insertions, 24 deletions
diff --git a/TODO b/TODO
deleted file mode 100644
index cb0c1a57c6..0000000000
--- a/TODO
+++ /dev/null
@@ -1,24 +0,0 @@
-
-
-* fix increments in major_collection_step() in the phases
- STATE_MARKING, STATE_SWEEPING_RAWMALLOC, STATE_SWEEPING_ARENA,
- and probably STATE_FINALIZING
-
-* 'next_major_collection_*' need to be tweaked
-
-* check the 'reserving_size' argument here and there
-
-* maybe make incremental: dealing with old_objects_with_weakrefs
- and old_objects_with_light_finalizers and
- deal_with_objects_with_finalizers()
-
-* REDO external_malloc(): if somebody calls this function a lot, we must
- eventually force a full collection.
-
-* REDO card marking, starting with "card_page_indices": 128 in
- TRANSLATION_PARAMS
-
-* write barrier: avoid the case when during sweeping we have GCFLAG_VISITED
- on an object, so we call the slow path, but the slow path doesn't do
- anything, and we still have GCFLAG_VISITED so we will keep calling it
- on the same object