Closed boschmitt closed 2 years ago
Hi Bruno,
Could we get failing regression test(s) for the fix(es)?
Best
Merging #523 (ef95cf8) into master (4ceb7e0) will decrease coverage by
0.00%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #523 +/- ##
==========================================
- Coverage 82.61% 82.60% -0.01%
==========================================
Files 146 146
Lines 18080 18099 +19
==========================================
+ Hits 14936 14950 +14
- Misses 3144 3149 +5
Impacted Files | Coverage Δ | |
---|---|---|
include/mockturtle/networks/mig.hpp | 91.13% <100.00%> (+0.08%) |
:arrow_up: |
include/mockturtle/networks/xmg.hpp | 88.22% <100.00%> (+0.10%) |
:arrow_up: |
...nclude/mockturtle/algorithms/dsd_decomposition.hpp | 87.32% <0.00%> (-4.23%) |
:arrow_down: |
.../mockturtle/algorithms/node_resynthesis/direct.hpp | 54.12% <0.00%> (-1.02%) |
:arrow_down: |
include/mockturtle/networks/aig.hpp | 93.91% <0.00%> (+0.05%) |
:arrow_up: |
include/mockturtle/networks/xag.hpp | 93.67% <0.00%> (+0.06%) |
:arrow_up: |
include/mockturtle/algorithms/refactoring.hpp | 72.61% <0.00%> (+0.32%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 4ceb7e0...ef95cf8. Read the comment docs.
Thanks!
The comment in the test cases(s) has a minor typo: /constate/constant/
Someone must check the other networks to see if the problem also exists, e.g.,
aqfp
seems to suffer from it too.
@mdsudara FYI
Both
mig
andxmg
uses the same logic asaig
to retrieve a primary input index from its node. However, they create primary inputs differently: instead of making their children equal to thepi
index, as in aaig
, the make them equal to the numerical limit ofuint64_t
. Effectively this means that thepi_index
method onmig
andxmg
in broken.This commit solves the issue. Someone must check the other networks to see if the problem also exists, e.g.,
aqfp
seems to suffer from it too.