mirror of
https://github.com/circify/circ.git
synced 2026-04-21 03:00:54 -04:00
50b97f6acb
This PR does two things: - Fixes some scoping issues in Z# having to do with the context in which function and struct definitions are interpreted. - Improves speed and reduces TermTable memory consumption by 2x (by keying on Arc<TermData> rather than TermData) - Improves speed for Z# array handling with a BTreeMap -> HashMap swap - Adds intrinsic widening casts in Z#. These were previously handled by bit-splitting, which is wasteful. On array-heavy Z# programs, this PR improves speed by about 5x and memory consumption by roughly the same. --- * zsharp::term::ir_array should use HashMap (faster, less memory) * perf improvements constant folding: get rid of redundant clones bitvector: impl arith for refs TermTable: don't store redundant term as key * EMBED widening casts * generic inf: need to push fn defn context onto filestack * canonicalize path in err msg * need struct defn on filestack, too * no need to warn about U32 indexing in zsharp * fix warning
17 lines
169 B
Plaintext
17 lines
169 B
Plaintext
/target
|
|
/pf
|
|
/assignment.txt
|
|
/params
|
|
__pycache__
|
|
/P
|
|
/V
|
|
/pi
|
|
/x
|
|
/perf.data*
|
|
/heaptrack.*
|
|
/flamegraph.*
|
|
/.gdb_history
|
|
.features.txt
|
|
scripts/aby_tests/tests
|
|
/flamegraph*.svg
|