Address many more sign comparison warnings #1140

pull jgarzik wants to merge 6 commits into bitcoin:master from jgarzik:sign-compare changing 13 files +74 −64

jgarzik commented at 6:05 PM on April 22, 2012: contributor

This really quiets down -Wall. We are down to 3 key signed/unsigned warnings in main.h, and 2-3 other misc. warnings overall.

Prefer 'unsigned int' for loop index variables tested against ::size()

C++ STL ::size() generally returns unsigned, which implies that "int idx"
style of loop variable will generate a signed-vs-unsigned comparison warning
when testing the loop exit condition "idx < blah.size()"

Update areas of the bitcoin code where loop variables may be more properly and
correctly defined as unsigned.

faf705a42a

SigOp and orphan-tx constants and counts are always unsigned.
Fixes several sign-comparison warnings.
7bd9c3a3cf
Test ScriptSigArgsExpected() for error, before accumulating return value c0a0a93d02

Add casts for unavoidable signed/unsigned comparisons

At these code sites, it is preferable to cast rather than change
a variable's type.

1d8c7a9557

CBlock::WriteToDisk() properly checks ftell(3) for error return

Rather than storing ftell(3)'s return value -- a long -- in an
unsigned int, we store and check a properly typed temp.  Then, assured a
non-negative value, we store in nBlockPosRet.

5aa0b23825

Change signed->unsigned at 3 code sites

This resolves signed/unsigned comparison warnings.

735a60698c

in src/main.cpp:None in 3b1a096229 outdated

 194 | @@ -195,9 +195,9 @@ void static EraseOrphanTx(uint256 hash)
 195 |      mapOrphanTransactions.erase(hash);
 196 |  }
 197 |  
 198 | -int LimitOrphanTxSize(int nMaxOrphans)
 199 | +static unsigned int LimitOrphanTxSize(unsigned int nMaxOrphans)

gavinandresen commented at 9:12 PM on April 22, 2012:

Making this static breaks test/DoS_tests.cpp

jgarzik commented at 6:20 PM on April 23, 2012: contributor
RE @gavinandresen:
1. Updated and rebased, to fix LimitOrphanTxSize() issue.
2. RE "Test ScriptSigArgsExpected() for error" -- I agree ScriptSigArgsExpected() may not return <0 given current conditions, but there are two advantages to the new code, a) the code is less fragile, in case those conditions change in the future. checking for an error that a function is pretty harmless, and trivial in terms of CPU cycles. b) the compiler's optimizer is guaranteed a non-negative value, enabling certain value-range optimizations.
gavinandresen commented at 6:27 PM on April 23, 2012: contributor

ACK.
jgarzik referenced this in commit c2e8c8acd8 on Apr 24, 2012
jgarzik merged this on Apr 24, 2012
jgarzik closed this on Apr 24, 2012
coblee referenced this in commit ec93ad8369 on Jul 17, 2012
suprnurd referenced this in commit 714f9a4f0d on Dec 5, 2017
lateminer referenced this in commit 70fbbf7ae8 on Jan 22, 2019
xdustinface referenced this in commit 33c0e6641b on Oct 2, 2020
Bushstar referenced this in commit ba3c6519cd on Oct 21, 2020
DrahtBot locked this on Sep 8, 2021

Contributors

jgarzik

gavinandresen