Fix for #12768: -blockmaxsize has been removed, but some tests were using this feature, so update with -blockmaxweight
[Tests] Use blockmaxweight where tests previously had blockmaxsize #12790
pull conscott wants to merge 1 commits into bitcoin:master from conscott:12768_remove_blockmaxsize changing 2 files +6 −6-
conscott commented at 1:23 PM on March 26, 2018: contributor
-
in test/functional/feature_fee_estimation.py:141 in 321c448163 outdated
139 | + ["-blockmaxweight=32000", "-maxorphantx=1000"]]) 140 | # Use node0 to mine blocks for input splitting 141 | # Node1 mines small blocks but that are bigger than the expected transaction rate. 142 | # NOTE: the CreateNewBlock code starts counting block size at 1,000 bytes, 143 | - # (17k is room enough for 110 or so transactions) 144 | + # (68k weight is room enough for 120 or so transactions)
conscott commented at 1:24 PM on March 26, 2018:I have verified these numbers are still correct.
jnewbery commented at 3:04 PM on March 26, 2018:Good! Is this test using segwit or legacy transactions? I assume that would have an impact on these numbers?
fanquake added the label Tests on Mar 26, 2018conscott force-pushed on Mar 26, 2018in test/functional/feature_maxuploadtarget.py:37 in 5e1709628c outdated
34 | + 35 | def set_test_params(self): 36 | self.setup_clean_chain = True 37 | self.num_nodes = 1 38 | - self.extra_args = [["-maxuploadtarget=800"]] 39 | + self.extra_args = [["-maxuploadtarget=800", "-blockmaxweight=3996000"]]
conscott commented at 1:45 PM on March 26, 2018:May just remove these, since it's the default, but the limit was explicitly added in the original test.
jnewbery commented at 2:55 PM on March 26, 2018:I agree that we should just remove this - there's no point in explicitly setting config to the default value (in fact it just distracts from the purpose of the test).
conscott commented at 1:48 PM on March 26, 2018: contributorRebased.
in test/functional/feature_fee_estimation.py:140 in 5e1709628c outdated
137 | - ["-maxorphantx=1000"]]) 138 | + ["-blockmaxweight=68000", "-maxorphantx=1000"], 139 | + ["-blockmaxweight=32000", "-maxorphantx=1000"]]) 140 | # Use node0 to mine blocks for input splitting 141 | # Node1 mines small blocks but that are bigger than the expected transaction rate. 142 | # NOTE: the CreateNewBlock code starts counting block size at 1,000 bytes,
jnewbery commented at 3:04 PM on March 26, 2018:This comment can be updated to say "CreateNewBlock code starts counting block weight at 4,000 weight"
jnewbery commented at 3:05 PM on March 26, 2018: memberutACK. I agree that
-blockmaxweight=3996000should be removed. One request to update comments inline.promag commented at 3:13 PM on March 26, 2018: memberPR title and commit message are outdated, there is no such thing as
blockmaxsize😄conscott force-pushed on Mar 26, 2018conscott force-pushed on Mar 26, 2018conscott renamed this:[Tests] Replace deprecated blockmaxsize with blockmaxweight
[Tests] Use blockmaxweight where tests previously had blockmaxsize
on Mar 26, 2018in test/functional/feature_fee_estimation.py:143 in 0868bd1592 outdated
142 | - # NOTE: the CreateNewBlock code starts counting block size at 1,000 bytes, 143 | - # (17k is room enough for 110 or so transactions) 144 | + # NOTE: the CreateNewBlock code starts counting block weight at 4,000 weight, 145 | + # (68k weight is room enough for 120 or so transactions) 146 | # Node2 is a stingy miner, that 147 | # produces too small blocks (room for only 55 or so transactions)
jnewbery commented at 3:43 PM on March 26, 2018:Is this number still correct?
conscott commented at 3:52 PM on March 26, 2018:Yes. I tested this as well.
in test/functional/feature_pruning.py:34 in 0868bd1592 outdated
30 | @@ -31,14 +31,14 @@ def set_test_params(self): 31 | 32 | # Create nodes 0 and 1 to mine. 33 | # Create node 2 to test pruning. 34 | - self.full_node_default_args = ["-maxreceivebuffer=20000", "-checkblocks=5", "-limitdescendantcount=100", "-limitdescendantsize=5000", "-limitancestorcount=100", "-limitancestorsize=5000" ] 35 | + self.full_node_default_args = ["-maxreceivebuffer=20000", "-blockmaxweight=3996000", "-checkblocks=5", "-limitdescendantcount=100", "-limitdescendantsize=5000", "-limitancestorcount=100", "-limitancestorsize=5000"]
jnewbery commented at 3:44 PM on March 26, 2018:I think this change can be removed
in test/functional/feature_pruning.py:41 in 0868bd1592 outdated
39 | self.full_node_default_args, 40 | ["-maxreceivebuffer=20000", "-prune=550"], 41 | - ["-maxreceivebuffer=20000"], 42 | - ["-maxreceivebuffer=20000"], 43 | + ["-maxreceivebuffer=20000", "-blockmaxweight=3996000"], 44 | + ["-maxreceivebuffer=20000", "-blockmaxweight=3996000"],
jnewbery commented at 3:44 PM on March 26, 2018:I think this change can be removed
[Tests] Use blockmaxweight where tests previously had blockmaxsize b466f6be95conscott force-pushed on Mar 26, 2018MarcoFalke commented at 4:03 PM on March 26, 2018: memberutACK b466f6be959b25f3a07cede9f03563ed0bbda0f
jnewbery commented at 4:12 PM on March 26, 2018: memberDon't we persist the mempool to disk?
Yes, unless
-persistmempool=0is set.jnewbery commented at 4:29 PM on March 26, 2018: memberTested ACK b466f6be959b25f3a07cede9f03563ed0bbda0fa
promag commented at 5:20 PM on March 26, 2018: memberutACK b466f6b.
ajtowns commented at 10:32 PM on March 26, 2018: contributorutACK b466f6be959b25f3a07cede9f03563ed0bbda0fa
laanwj merged this on Mar 29, 2018laanwj closed this on Mar 29, 2018laanwj referenced this in commit 490644d29e on Mar 29, 2018fanquake cross-referenced this on Mar 29, 2018 from issue [tests] Fix block size config in feature_fee_estimation.py, feature_maxuploadtarget.py and feature_pruning.py by jnewberyconscott deleted the branch on Jul 31, 2018bitcoin locked this on Sep 8, 2021
github-metadata-mirror
This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2026-05-20 06:55 UTC