Commit Graph

33 Commits

Author SHA1 Message Date
20dc99ba1f Re-indent some code to take up less horizontal space
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-09 16:57:03 -07:00
f0b0d51b48 Add unproven lemma about fold implication
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 23:38:23 -07:00
8ff88f9f9e Move helper code into separate function
I'll need to reuse it.

Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 23:35:02 -07:00
838aaf9c58 Start end-to-end proof of correctness
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 23:30:03 -07:00
4ac9dffa9b Prove that the var->lattice maps respect equality
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 22:53:21 -07:00
3f5551d70c Add a lemma about the effect of joinAll
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 22:34:02 -07:00
5837fdf19b Prove that 'updateAll' has preservation
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 22:29:36 -07:00
4350919871 Add proof about setVariablesForState
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 22:09:56 -07:00
7fe46b014c Slightly simplify evaluation code
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 22:05:50 -07:00
66d229c493 Prove that multi-statement evaluation "preserves" the validity of the analysis
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 21:51:53 -07:00
1b8bea8957 Use foldl in multi-statement evaluation
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 21:50:38 -07:00
ad4592d4d2 Switch to using implicit arguments where needed
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 21:34:17 -07:00
8d0d87d2d9 Start on proofs of correctness
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-05-08 20:50:21 -07:00
3859826293 Define interpretation of the sign lattice
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-04-30 21:58:41 -07:00
de956cdc6a Split the Language file into modules
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-04-13 18:39:38 -07:00
44f04e4020 Get forward analysis working again
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-04-13 14:08:40 -07:00
5d56a7ce2d Fix comments in Forward.agda
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-23 12:09:14 -07:00
2e096bd64e Extract common parts of forward analyses into Forward.agda
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-22 17:50:29 -07:00
1a7b2a1736 Adjust behavior of eval to not require constant 'k in vars' threading
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-22 17:15:40 -07:00
f0da9a9020 Move more code out of Sign and into Main
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-11 13:07:42 -07:00
040c13caba Use instances to simplify printing code
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-11 12:50:05 -07:00
3e88a64ed9 Add some debugging code to sign analysis to print the results
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 22:23:45 -07:00
8a85c4497c Prove that evaluation is monotonic and complete sign analysis
Other than monotonicity of plus and minus, god damn it.

Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 21:25:46 -07:00
8964ba59a1 Prove monotonicity of eval
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 20:29:05 -07:00
96f3ceaeb2 Use the previous join function directly in GeneralizedUpdate
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 19:41:02 -07:00
237250cf72 Stop using modules in 'Sign' analysis
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 19:23:48 -07:00
8515491327 Simplify AboveBelow a bit to avoid nested modules
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 18:43:10 -07:00
f21ebdcf46 Start working on the evaluation operation.
Proving monotonicity is the main hurdle here.

Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 18:13:01 -07:00
51accb6438 Define 'minus', too -- with no monotonicity proof.
I'm still thinking about how this should be achieved most easily.

Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 16:40:49 -07:00
afe5bac2dc Commit result of (unsuccessfully) trying to prove monotonicity of plus.
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-10 13:54:19 -07:00
f84a1c923c Prove that the 'join' transformation is monotonic
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-09 23:06:47 -08:00
1b1b80465c Use named modules to avoid having to pass redundant parameters
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-09 21:46:15 -08:00
0c30f8be48 Start on sign analysis (mostly just imports)
Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
2024-03-09 14:00:10 -08:00