

@1509

10 years 
mulligan 
i hate subtraction over the nats



@1506

10 years 
mulligan 
changes to costs proof over weekend



@1503

10 years 
mulligan 
inductive type complete



@1502

10 years 
mulligan 
changes to inductive defn



@1501

10 years 
sacerdot 
We must take in account the labelled_p predicate.



@1500

10 years 
sacerdot 
Proof sketch for one of the two main proofs.



@1499

10 years 
mulligan 
part way through main statement transcription



@1498

10 years 
mulligan 
added new file for proof that costs are preserved at asm level



@1497

10 years 
mulligan 
a bit of tidying up, removing dead code, etc.



@1496

10 years 
mulligan 
problem fixed with tactics missing a semicolon to stop greedy parsing



@1495

10 years 
mulligan 
proper calculation of costs



@1494

10 years 
mulligan 
changes to get everything compiling again



@1493

10 years 
mulligan 
finished well labeled check, up to injectivity of the label map



@1487

10 years 
mulligan 
committing some code for well labelling



@1486

10 years 
mulligan 
finished asm costs



@1485

10 years 
sacerdot 
Less nice definitiion of add_with_carries that avoids a quadratic …



@1484

10 years 
sacerdot 
…



@1482

10 years 
sacerdot 
1. very long standing conflict committed (but don't ask me what the …



@1479

10 years 
boender 
 added insert_lookup_opt
 assembly compiles now



@1478

10 years 
sacerdot 
Almost completed (up to is_finals).



@1476

10 years 
sacerdot 
…



@1475

10 years 
sacerdot 
Towards the two fullexec transition systems that represent …



@1474

10 years 
mulligan 
adding missing asmcosts file for computing the costs of an assembly …



@1463

10 years 
mulligan 
added erasure for lin



@1461

10 years 
mulligan 
rewrote erasure for assembly programs



@1460

10 years 
mulligan 
most of cost label erasure for assembly language complete, with one …



@1459

10 years 
boender 
 moved stronger occurs_exactly_once lemma to its proper place in …



@1426

10 years 
boender 
removed axiom



@1424

10 years 
sacerdot 
1. fold function over BitVectorTries? moved from ERTLToLTL to …



@1417

10 years 
boender 
 proved that jumps always increase  this should make termination easy



@1416

10 years 
sacerdot 
Maps from hardware registers to beval now implemented in ASM/I8051 (in …



@1415

10 years 
sacerdot 
1. hwreg_store/retrieve no longer returns a res (but it is still …



@1404

10 years 
boender 
 reworked + added
 added an axiom to arithmetic, but should be provable



@1393

10 years 
boender 
 added invariant for policy trie to assembly
 change (syntax only) …



@1363

10 years 
boender 
 done stuff with create_label_trie



@1335

10 years 
sacerdot 
Ported to new Matita stdlib.



@1333

10 years 
sacerdot 
Avoid using the name of the construction of jmeq.



@1330

10 years 
campbell 
Evict obsolete file.



@1323

10 years 
campbell 
Reduce number of notations for destructive let on pairs to one.



@1316

10 years 
campbell 
Merge in idlookupbranch to trunk.



@1309

10 years 
boender 
 refounded JEP



@1279

10 years 
sacerdot 
Bug fixed in definition of ltb.



@1207

10 years 
campbell 
Second part of fixing temporaries in Clight to Cminor stage.



@1193

10 years 
mulligan 
work on colouring algorithm halted as it can be axiomatised. now …



@1187

10 years 
mulligan 
fixed build.ma



@1161

10 years 
mulligan 
changes from today: merged ertl, ltl and lin into one datatype to …



@1159

10 years 
boender 
 added 'nth' theorems
 moved up \bot a bit



@1145

10 years 
mulligan 
changed naming in i8051 of classes of registers to make them consistent



@1119

10 years 
sacerdot 
Type for evaluation of opaccs fixed (maybe wrongly: should it return …



@1112

10 years 
mulligan 
got lin > asm stuff working



@1103

10 years 
boender 
 reverted to old policy



@1094

10 years 
mulligan 
some changes from today to do with liveness analyses



@1089

10 years 
mulligan 
more changes from earlier in the week



@1075

10 years 
mulligan 
nearly completed rtl > ertl pass removing all option types with dep. types



@1074

10 years 
boender 
 added lookup lemma



@1071

10 years 
mulligan 
changes the specific form that the added proofs take to use None, not …



@1070

10 years 
campbell 
Show that entry and exit labels are in the RTLabs graph.



@1069

10 years 
campbell 
Change odd proof obligation problem back.



@1066

10 years 
mulligan 
changes from today



@1064

10 years 
mulligan 
changes from today, nearly complete rtlabs translation pass



@1063

10 years 
mulligan 
changes from today



@1062

10 years 
mulligan 
separated jmeq and coercions from foldstuff.ma in order to fix the …



@1061

10 years 
mulligan 
more work, bug found, ridiculous map3 function with dep. types added



@1060

10 years 
mulligan 
work from this morning and yesterday



@1059

10 years 
mulligan 
work from today, bit of a mess at the moment



@1057

10 years 
mulligan 
changes from today



@1054

10 years 
boender 
 proven policy safe



@1052

10 years 
mulligan 
removed offsets after reading cerco mailing list



@1045

10 years 
mulligan 
resolved conflict in rtlabs



@1044

10 years 
boender 
 more fold/forall stuff



@1043

10 years 
sacerdot 
Axiom commented out.



@1042

10 years 
sacerdot 
Dead code removed.
Slow code uncommented.



@1041

10 years 
sacerdot 
fetch_assembly is still working after bug fix



@1040

10 years 
sacerdot 
Bug fixed in assembly.



@1039

10 years 
sacerdot 
fetch_assembly_pseudo2 repaired from dependent type madness



@1038

10 years 
boender 
 some more BVT improvements



@1037

10 years 
sacerdot 
Main theorem: comments are working again.



@1036

10 years 
sacerdot 
…



@1035

10 years 
sacerdot 
Main theorem (broken because of dependent types) almost restored.



@1034

10 years 
boender 
various & sundry fold/forall lemmas



@1033

10 years 
sacerdot 
ispelled & submitted



@1032

10 years 
sacerdot 
Final version.



@1031

10 years 
sacerdot 
..



@1030

10 years 
sacerdot 
…



@1029

10 years 
sacerdot 
…



@1028

10 years 
sacerdot 
One more sentence restored and fitted in.



@1027

10 years 
sacerdot 
Bug fixed in figure.



@1026

10 years 
mulligan 
final version? under 16 pages



@1025

10 years 
mulligan 
removing stray single words to reduce page usage



@1024

10 years 
mulligan 
tidied explanation of proof



@1023

10 years 
mulligan 
changes to english in matita section, shrunk diagrams in introduction …



@1022

10 years 
sacerdot 
…



@1021

10 years 
mulligan 
tidied english in sect 3



@1020

10 years 
sacerdot 
More on Matita.



@1019

10 years 
sacerdot 
Finished rewriting of Section 3.



@1018

10 years 
mulligan 
tidying



@1017

10 years 
mulligan 
complete, just under 16 pages



@1016

10 years 
sacerdot 
Many fixes to the code snippets.



@1015

10 years 
sacerdot 
One intermediate version of main_thm0 close to be repaired.



@1014

10 years 
sacerdot 
The main theorem is completely broken (again).


