]>
matita.cs.unibo.it Git - helm.git/log
Ferruccio Guidi [Tue, 24 Jul 2007 11:52:05 +0000 (11:52 +0000)]
New developement LOGIC about the cut elimination of implication for Sambin's basic logic; one of the rules must be corrected (work is in progress)
Enrico Tassi [Tue, 24 Jul 2007 08:58:16 +0000 (08:58 +0000)]
added test about dependent coercions
Claudio Sacerdoti Coen [Mon, 23 Jul 2007 15:14:29 +0000 (15:14 +0000)]
Prototype of min_aux changed.
Now (min_aux off n f) find the smallest i such that
* f i = true or i = n+off
* \forall j, n <= j <= n+off, f j = false
The new function does no longer compute with off. Thus we obtain
for free a great computational speed-up in every function defined
in terms of it.
Ferruccio Guidi [Mon, 23 Jul 2007 13:22:22 +0000 (13:22 +0000)]
auto vs autobatch fixed
Ferruccio Guidi [Mon, 23 Jul 2007 13:12:40 +0000 (13:12 +0000)]
autobatch parameters reajusted
Enrico Tassi [Mon, 23 Jul 2007 09:14:53 +0000 (09:14 +0000)]
fixed makefiles to make it compile cleanly again
Ferruccio Guidi [Sun, 22 Jul 2007 18:33:38 +0000 (18:33 +0000)]
Procedural: the statement body and it inner types must satisfy the Barendregt
convenction on bound variables
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 16:59:27 +0000 (16:59 +0000)]
Another nicer version.
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 16:53:33 +0000 (16:53 +0000)]
Even nicer script.
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 16:20:30 +0000 (16:20 +0000)]
The nicest script obtained so far.
What is left, is a bunch of change/normalize/whd/simplify that are difficult
to control precisely.
Ferruccio Guidi [Fri, 20 Jul 2007 14:48:36 +0000 (14:48 +0000)]
acic_procedural: bug fix:
sometimes whd is not enough, we need normalize ...
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 14:28:55 +0000 (14:28 +0000)]
More simplification due to the new conversion strategy.
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 14:26:01 +0000 (14:26 +0000)]
Script simplification due to the new efficient conversion strategy.
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 09:58:30 +0000 (09:58 +0000)]
Tentative bug fix for diverging pretty-printing function.
Claudio Sacerdoti Coen [Fri, 20 Jul 2007 09:32:39 +0000 (09:32 +0000)]
Initialization of matita.map_unicode_to_tex moved from matitaGui to matitaInit
(so that ocamlc now works again).
Ferruccio Guidi [Thu, 19 Jul 2007 18:00:55 +0000 (18:00 +0000)]
cicUtil : new test function "is_sober" to test the integrity of a term
now fails when fake applications are detected
acic_procedural: some bug fix
Claudio Sacerdoti Coen [Thu, 19 Jul 2007 17:54:40 +0000 (17:54 +0000)]
Convertibility now converts machines in place of terms.
Actually, the heads of the machines (i.e. everything but the stack) are
still unwinded and converted as terms. Thus there is much space for
improvement.
Claudio Sacerdoti Coen [Thu, 19 Jul 2007 12:07:44 +0000 (12:07 +0000)]
map_unicode_to_tex is no longer optional and it always refers to the current
user choice.
Claudio Sacerdoti Coen [Thu, 19 Jul 2007 10:19:28 +0000 (10:19 +0000)]
paste_unicode_as_tex is now false by default; moreover the flag is used
everywhere (but in patterns, I do not know why)
Enrico Tassi [Thu, 19 Jul 2007 10:03:42 +0000 (10:03 +0000)]
COERCIONS: tentative addition of an equivalence relation over coercion source
carriers (convertibility) for the moment used only in the FunClass case.
Enrico Tassi [Thu, 19 Jul 2007 10:02:11 +0000 (10:02 +0000)]
the cade was escaping the table name and not the uri
Enrico Tassi [Thu, 19 Jul 2007 10:01:40 +0000 (10:01 +0000)]
fixed escaping for sqlite
Enrico Tassi [Thu, 19 Jul 2007 10:01:13 +0000 (10:01 +0000)]
fixed an escaping error, added more infos to the generic error, callback catches and prints exceptions
Claudio Sacerdoti Coen [Thu, 19 Jul 2007 09:17:32 +0000 (09:17 +0000)]
Typo fixed.
Claudio Sacerdoti Coen [Thu, 19 Jul 2007 09:15:31 +0000 (09:15 +0000)]
Super-nice notation for the assembly stuff.
Ferruccio Guidi [Wed, 18 Jul 2007 16:07:11 +0000 (16:07 +0000)]
the predicate for elim was not built correctly when more than one right parameter was found
Enrico Tassi [Wed, 18 Jul 2007 11:17:46 +0000 (11:17 +0000)]
recursive argument in let rec is not printed explicitly.
I let to ferruccio optimize the case in which there is only one argument or the recno is 0 and thus can be omitted
Enrico Tassi [Wed, 18 Jul 2007 10:46:11 +0000 (10:46 +0000)]
fixed coercion graph print, moved coercion graph and auto gui to the view menu, added a mini test about coercions
Claudio Sacerdoti Coen [Tue, 17 Jul 2007 11:41:07 +0000 (11:41 +0000)]
added missing parenthesis
Enrico Tassi [Tue, 17 Jul 2007 10:00:30 +0000 (10:00 +0000)]
fixed includes and added notation for bytes
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 22:20:08 +0000 (22:20 +0000)]
More lemmas.
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 21:39:13 +0000 (21:39 +0000)]
More daemons closed. A couple left in byte and many in extras.
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 20:48:20 +0000 (20:48 +0000)]
More daemons/axioms closed.
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 17:51:02 +0000 (17:51 +0000)]
One daemon less.
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 17:35:38 +0000 (17:35 +0000)]
More daemons got rid of (and more extra axioms to be proved).
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 16:10:40 +0000 (16:10 +0000)]
Last daemon killed :-)
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 16:07:42 +0000 (16:07 +0000)]
One less daemon (about "update"s).
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 16:04:05 +0000 (16:04 +0000)]
All sub-proofs about "update" closed.
Claudio Sacerdoti Coen [Mon, 16 Jul 2007 14:51:18 +0000 (14:51 +0000)]
assembly.ma splitted into many files
Stefano Zacchiroli [Mon, 16 Jul 2007 14:34:55 +0000 (14:34 +0000)]
* NOT RELEASED YET
Stefano Zacchiroli [Mon, 16 Jul 2007 14:33:30 +0000 (14:33 +0000)]
* NOT RELEASED YET
Stefano Zacchiroli [Mon, 16 Jul 2007 14:30:27 +0000 (14:30 +0000)]
proper path for ps.gz doc
Stefano Zacchiroli [Mon, 16 Jul 2007 14:21:44 +0000 (14:21 +0000)]
texlive-base-bin, texlive-latex-extra
Stefano Zacchiroli [Mon, 16 Jul 2007 14:17:45 +0000 (14:17 +0000)]
- add build-dep for doc generation: graphviz, texlive-latex-recommended,
Stefano Zacchiroli [Mon, 16 Jul 2007 14:12:20 +0000 (14:12 +0000)]
- add build-dep for doc generation: graphviz, texlive-latex-base,
texlive-base-bin
Stefano Zacchiroli [Mon, 16 Jul 2007 14:09:23 +0000 (14:09 +0000)]
* debian/rules
- build ocamldoc documentation at package build time
- add build-dep on graphviz
Stefano Zacchiroli [Mon, 16 Jul 2007 14:05:57 +0000 (14:05 +0000)]
invoke make doc after build to create ocamldoc docs
Stefano Zacchiroli [Mon, 16 Jul 2007 13:59:36 +0000 (13:59 +0000)]
* debian/svn-deblayout
- add repository layout information
Stefano Zacchiroli [Mon, 16 Jul 2007 13:56:27 +0000 (13:56 +0000)]
* rebuild against OCaml 3.10 and ocamlnet 2.2
* add preliminary support for cookies (new "cookies" method added to an
- use ocaml.mk CDBS class
- add build-dep on camlp4, which is now in a separate package
- bump debhelper dep and compatibility level to 5
Stefano Zacchiroli [Mon, 16 Jul 2007 13:49:22 +0000 (13:49 +0000)]
* debian/*.install.in
- do not attempt to instal *.ml files in target lib directory since they
are no longer installed by Makefile
Stefano Zacchiroli [Mon, 16 Jul 2007 13:31:55 +0000 (13:31 +0000)]
* rebuild with OCaml 3.10
* bump debhelper deps and compatibility level to 5
- s/Source-Version/binary:Version
Wilmer Ricciotti [Mon, 16 Jul 2007 13:07:53 +0000 (13:07 +0000)]
corrected axiom mod_plus
Ferruccio Guidi [Sat, 14 Jul 2007 14:37:11 +0000 (14:37 +0000)]
new definitions and new theorems
Claudio Sacerdoti Coen [Fri, 13 Jul 2007 22:04:05 +0000 (22:04 +0000)]
More conjectures closed.
Claudio Sacerdoti Coen [Fri, 13 Jul 2007 18:08:28 +0000 (18:08 +0000)]
1. requires the new pretty printer for natural numbers
2. important properties of plusbyte proved
3. all important conjectures almost proved
Claudio Sacerdoti Coen [Fri, 13 Jul 2007 18:04:35 +0000 (18:04 +0000)]
Dirty patch by Zack: natural numbers of Matita are now pretty-printed as numbers. Eureka.
Claudio Sacerdoti Coen [Fri, 13 Jul 2007 09:56:42 +0000 (09:56 +0000)]
Last crazy commit reverted.
Claudio Sacerdoti Coen [Fri, 13 Jul 2007 08:35:47 +0000 (08:35 +0000)]
Final theorem proved. Many many conjectures left. I am unsure of one of them.
Claudio Sacerdoti Coen [Thu, 12 Jul 2007 18:38:04 +0000 (18:38 +0000)]
The loop invariant I chosed was not correct!
1. statement to be fixed
2. is the new loop invariant strong enough?
3. lot of daemons here and there
4. two conjectures I am not sure of:
plusbytenc (x*n) x = byte_of_nat (x*S n)
plusbytec (x*n) x = plusbytec (x*S n)
Claudio Sacerdoti Coen [Wed, 11 Jul 2007 17:10:26 +0000 (17:10 +0000)]
Getting close to the final result.
The showstopper now is the GtkMathView/Pango bug!
Moreover, unification has become mostly useless (and rewriting with it).
Claudio Sacerdoti Coen [Wed, 11 Jul 2007 14:54:04 +0000 (14:54 +0000)]
1. loop invariant stated, but not proved
2. final theorem stated and proved
Ferruccio Guidi [Wed, 11 Jul 2007 12:23:14 +0000 (12:23 +0000)]
native dependences fixed
Claudio Sacerdoti Coen [Wed, 11 Jul 2007 10:52:49 +0000 (10:52 +0000)]
1. status factorized out in tick
2. 0*0=0, 0*2=0, x*1=x, x*2=x+x proved
3. still way too slow!
Claudio Sacerdoti Coen [Wed, 11 Jul 2007 09:16:43 +0000 (09:16 +0000)]
auto new => autobatch
Claudio Sacerdoti Coen [Wed, 11 Jul 2007 09:11:18 +0000 (09:11 +0000)]
auto new => autobatch
Claudio Sacerdoti Coen [Tue, 10 Jul 2007 22:06:50 +0000 (22:06 +0000)]
0. less nice solution by Enrico reverted
1. some bug fixes in assembly code
2. test program simplified
3. reduction now works as expected, thanks to the new reduction strategy
4. BIG BUG SOMEWHERE: (270 \mod 256) reduces to 14 instead of 4!!!
5. requires primitive addition over exadecimal numbers and bytes to
avoid representing 0+x as byte_of_nat (nat_of_byte x) that does not
reduce
Claudio Sacerdoti Coen [Tue, 10 Jul 2007 18:48:14 +0000 (18:48 +0000)]
New reduction strategy: the new reduction strategy behaves as the previous
one (performing both CVN and CBV in parallel), but it is much more lazy.
In particular, unwinding is never performed twice on the same term.
On assembly/assembly.ma this seems to solve every performance problem.
Let's see what happens this night with the benchmarks.
Enrico Tassi [Tue, 10 Jul 2007 17:11:56 +0000 (17:11 +0000)]
La programmazione funzionale e' come TeX, funziona meglio se la prendi a calci.
Ferruccio Guidi [Tue, 10 Jul 2007 15:42:40 +0000 (15:42 +0000)]
persistent inner types are now generated in publishing mode
Enrico Tassi [Tue, 10 Jul 2007 15:15:33 +0000 (15:15 +0000)]
fixed a bug in the cleanup ofsedir that was not properly catching #xpointer
Enrico Tassi [Tue, 10 Jul 2007 08:15:09 +0000 (08:15 +0000)]
...
Enrico Tassi [Mon, 9 Jul 2007 17:47:31 +0000 (17:47 +0000)]
1. bug fixed in tick
2. added notation for some big number
3. the whd CBN is responsible for the huge timing;
reduce works since it is CVB
Enrico Tassi [Mon, 9 Jul 2007 14:40:00 +0000 (14:40 +0000)]
signal hadler restored after runnig external 'make'
Claudio Sacerdoti Coen [Mon, 9 Jul 2007 14:34:57 +0000 (14:34 +0000)]
Interesting theorem added (but still to be proved).
Enrico Tassi [Mon, 9 Jul 2007 13:32:17 +0000 (13:32 +0000)]
added few more fun to this test
Ferruccio Guidi [Mon, 9 Jul 2007 13:04:18 +0000 (13:04 +0000)]
new configuration file sample
Ferruccio Guidi [Mon, 9 Jul 2007 13:01:25 +0000 (13:01 +0000)]
tassi: ported to the new DB architecture.
since the 'user' db is mandatory, a fake one should be provided
Enrico Tassi [Mon, 9 Jul 2007 09:58:45 +0000 (09:58 +0000)]
auto->autobatch
Enrico Tassi [Sat, 7 Jul 2007 11:34:18 +0000 (11:34 +0000)]
inclusion of div_and_mod
Enrico Tassi [Fri, 6 Jul 2007 14:49:47 +0000 (14:49 +0000)]
maxipatch for support of multiple DBs.
new 3 different kind of BD can coexist:
user) the user db, where his stuff is stored
tables are named foo_user
library) the library db, a read only database where the standard library should be
installed and that is the target of the publish act
legacy) a legacy database (not mandatory) where really read only data is stored
every db can be implement with an sqlite file or a mysql database.
the default configuration file sets 1 and 2 to be the mysql matita database on mowgli,
thus nothing should change for mowgli users. database 3 is not used if you are on mowgli.
the way it should be used by matita users out of the unibo network is commented in the config file:
1) file://~/.matita/user.db for his development
2) file:///usr/share/matita/metadata.db for the matita precompiled standard library
3) mysql://mowgli.cs.unibo.it for the legacy coq stuff, both queries and xml are fetched trough
the net
this allows us to really distribute matita in a rather sane way, with a
precompiled library visible systemwide and having all per user data and
metadata in ~/.matita. There is no need for the user to install and configure
mysql, and can use mowgli just decommenting few lines in the config file.
Claudio Sacerdoti Coen [Thu, 5 Jul 2007 21:47:27 +0000 (21:47 +0000)]
Exadecimal numbers are now used. This is a great speed-up.
Moreover, byte_of_nat is now modulo 256. Thus ADDd is now implemented
correctly.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 22:17:13 +0000 (22:17 +0000)]
Example program executed for x,y=0.
It works (slowly, much slowly)!
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 21:36:49 +0000 (21:36 +0000)]
Quick hack: matita natural numbers are now accepted by the parser/disambiguator.
If nat/nat.ma is not compiled yet, I do not know what happens.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:19:24 +0000 (20:19 +0000)]
Bug fixed: a ~with_cast is now inserted when appropriate to avoid translation
of a bottom-up conversion into a top-down conversion.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:17:47 +0000 (20:17 +0000)]
In place of conclude, obtain FIXMEXXX is now generated when required.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:16:47 +0000 (20:16 +0000)]
1. Code simplification
2. Bug fixed: do not anticipate justification for rewritingLR or rewriting RL
if justification is "exact"
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:15:18 +0000 (20:15 +0000)]
is_top_down was not propageted correctly under a bottom-up conversion.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:13:49 +0000 (20:13 +0000)]
1. "by proof" now allowed as a justification in equality chains.
It opens a side-proof to be proved immediately.
2. Side proofs restored in equality chains in content2pres.ml
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:12:41 +0000 (20:12 +0000)]
Added ~with_cast to the change tactic.
When with_cast is true, a cast is recorded in the lambda term.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 20:05:55 +0000 (20:05 +0000)]
Bug fixed: name in letin was printed as "previous" even if given.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 17:58:08 +0000 (17:58 +0000)]
More opcodes (badly) implemented.
Claudio Sacerdoti Coen [Wed, 4 Jul 2007 17:12:22 +0000 (17:12 +0000)]
List.ma: added function nth (with default value in case of failure)
Assembly: a toy example on proving properties of assembly programs.
It should be moved somewhere else. It shows HUGE problems with
simplification and display of (very very large) terms.
Cristian Armentano [Sat, 30 Jun 2007 17:15:15 +0000 (17:15 +0000)]
New definition of Euler's totient function.
New theorems about sigma_p (invoked with an always true predicate).
Theorem about the sum of totient invocations on the divisors of a natural number.
Claudio Sacerdoti Coen [Sat, 30 Jun 2007 12:33:24 +0000 (12:33 +0000)]
In order to generate executable declarative scripts, we are now splitting
the "Rewrite" content-level method into "RewriteRL" and "RewriteLR".
Claudio Sacerdoti Coen [Sat, 30 Jun 2007 12:20:10 +0000 (12:20 +0000)]
Bug fixed in bottom-up conversion.
Since the bottom-up conversion body is put in the applicative context and since
applicative contexts get flattened the test for being in a bottom-up conversion
must be rather complex.
Claudio Sacerdoti Coen [Sat, 30 Jun 2007 12:16:20 +0000 (12:16 +0000)]
BU Conversion was not generated for Rels fixed. I wonder why...
Claudio Sacerdoti Coen [Sat, 30 Jun 2007 12:11:01 +0000 (12:11 +0000)]
Nicer layout but possibly more bugged.
We need continuations here to do a good job.
Claudio Sacerdoti Coen [Sat, 30 Jun 2007 12:10:09 +0000 (12:10 +0000)]
Spurious "we need to prove" at the beginning of Intros+LetTac are no longer
generated.
Cristian Armentano [Fri, 29 Jun 2007 16:30:31 +0000 (16:30 +0000)]
generic version, specializing generic_sigma_p.ma
Cristian Armentano [Fri, 29 Jun 2007 14:33:37 +0000 (14:33 +0000)]
generic version