moved code around

give Z3 TPTP proofs a chance

fixed postprocessing of 'coinduct' formula to obtain right property format (without needless hypotheses)

add more derivative and continuity rules for complex-values functions

Moving the remote prefix deleting on Sledgehammer's side

Correcting the type parser

imported patch leo2_skolem_simplication

add support for Isar reconstruction for thf1 ATP provers like Leo-II.

lemmas about the moments of the normal distribution

NEWS

properties of normal distributed random variables (by Sudeep Kanav)

announce Tree

new theory of binary trees

formal variable name: IVar NONE is strictly spoken not supported on lhs of function definitions, e.g. in Scala

merged

added [simp]

tuning

renamed Sledgehammer options

removed dead code

reintroduced vital 'Thm.transfer'

tuned dependencies

updated docs

added support for CVC4 in SMT2

don't ask proof-disabled solvers to do proofs

tuning

took out broken support for Yices from SMT2 stack -- see 'NEWS' for rationale

made CVC3 work with SMT2 stack

properties of Erlang and exponentially distributed random variables (by Sudeep Kanav)

clean up ContNotDenum; add lemmas by Jeremy Avigad and Luke Serafin

uniform treatment of trivial unit instances: simplify by default, unfold in code preprocessor

adapted examples to changes in SMT triggers

reduces Sledgehammer dependencies

eliminate dependency of SMT2 module on 'list'

tuning

removed subsumed record-related code, now that records are registered as 'ctr_sugar'

made lookup more robust in the face of missing (dummy) case constant

use 'ctr_sugar' abstraction in SMT(2)

register record extensions as freely generated types

mixin definitions are within scope of "for"s of import expression

proper proof context transfer wrt. background theory avoids ad-hoc restore operation

refactoring

moved generic code further up

tuned code

factor out SMT-LIB 2 type/term parsing from Z3-specific code

simplify slightly ATP proof generated for Z3

tuned whitespaces

updated contributors to include students

moved new highly experimental Waldmeister-specific code (authored by Albert Steckermeier) into Isabelle

adapted SMT examples to new, corrected handling of 'typedef'

fixed unsoundness in SMT(2) as oracle: don't register typedef Abs_x as constructor unless it is known to be injective

updated NEWS slightly

updated docs w.r.t. Z3

rationalized CVC3 and Yices environment variable -- no need (unlike for Z3) to distinguish between old and new versions

removed '_new' sufffix in SMT2 solver names (in some cases)

removed old SMT module from Sledgehammer

got rid of 'listF' example, which is now subsumed by the real 'list' type

changed syntax of map: and rel: arguments to BNF-based datatypes

tuning

updated Z3 certificates

generate ':named' attributes in SMT-LIB 2 problems, to facilitate reconstruction

tuning

add type class instances for unit

use 'where' clause for selector default value syntax

tuning

added List.union

Sup/Inf on functions decoupled from complete_lattice.

tuned data structure

recovered level-free fishing for locale, accidentally lost in dce365931721

tuned terminology: emphasize stack-like nature of nested local theories

self-contained locale_declaration operation

yet another attempt for terminology: foo_target_bar denotes an operation bar operating solely on the target context of target foo, foo_bar denotes a whole stack of operations to accomplish bar for target foo

tuned

recovered structure of module, which got somehow convoluted due to incremental modifications

re-unified approach towards class and locale consts, with refined terminology: foo_const_declaration denotes declaration for a particular logical layer, foo_const the full stack for a particular target

revert effect of 63c7b29d29ac -- existing pervasive interpretation of class order for dvd etc. is too obstrusive at the moment

less ad-hoc verision of educated guess: guess identity of declaration morphism wrt. canonical morphism rather than observing particular effects of declaration morphisms only; n.b. this fullfills promise given in 63c7b29d29ac

treat non-canonical interpretations of classes the same way as ordinary locale interpretations

tuned

avoid odd Named_Target.reinit altogether

clarified terminology: toplevel is interwined with named targets in particular, not with local theories in general

less baroque interface

dropped obscure and unused ad-hoc before_exit hook for named targets

added lemma

mira: USER_HOME must exist for building JEdit documentation

added lemmas

sharpened criterion: bare named target is only at the bottom level

tuned

extended stream library

be more explicit: made sml/nj happy

always refine interpretation morphism using canonical constant's definition theorem

set USER_HOME to affect also ISABELLE_PATH et al

merge

updated SMT2 certificates

tune

disable hard-to-reconstruct Z3 feature

new Z3 4.3.2 component, based on more recent repository version, and whose Mac binary was built on Mac OS X 10.7

use 0 as integral-value for non-integrable functions, simplify a couple of rewrite rules

removed SMT weights -- their impact was very inconclusive anyway

make SMT code less dependent on Z3 proofs

tuning

avoid division by 0

formal treatment of dangling parameters for class abbrevs analogously to class consts

explicit passing of params

refactored Z3 to Isar proof construction code

simplified counterexample handling

split replay and proof parsing for Z3

removed counterexample parser (obsolete and useless in practice)

remove superfluous assumption

basic setup for zipperposition prover

document property 'sel_set'

generate 'sel_set' theorem for (co)datatypes

removed some spurious warnings in new (co)datatype package

add option to keep duplicates, for more precise evaluation of relevance filters

tuned whitespace

definition in class: provide explicit auxiliary abbreviation carrying potential mixfix syntax in presence of dangling parameters

tuned

merged

tuned

more complete proof replay for Z3: support universally quantified rewrite steps

postpone const declarations for nested context after canonical const declarations: keep const declarations stemming from interpretation together

tuned

explicit is better than implicit

tuned names

dropped accidental duplicate application of morphism

generalizd measurability on restricted space; rule for integrability on compact sets

better support for restrict_space

must not cancel common factors on both sides of (in)equations in linear arithmetic decicision procedure

merged

updated cygwin -- include perl_vendor for libwww-perl;

made 'Kuehlwein-style' be really like Python code, we now think

make SML code closer to Python code when 'nb_kuehlwein_style' is true

merge

added sleep to give time for the server to shut down -- this is a hack, but it's only in experimental code that will hopefully soon go away

introduce more powerful reindexing rules for big operators

merged

make double-sure that old popup is dismissed, before replacing it;

more robust bold_style, e.g. relevant for accidental \<^bold> before keyword;

added another way of invoking Python code, for experiments

make SML naive Bayes closer to Python version

tuned whitespace, to make datatype definitions slightly less intimidating

more work on exporter

got rid of 'linearize' option

extend exporter with new versions of MaSh

tuned

tuned signature

terminating code equations

more direct passing of right-hand side

even more uniform treatment of result after 95e5a633a18f

Merge

New theorems to enable the simplification of certain functions when applied to specific natural number constants (such as 4)

removed Kleene_Algebra because of superior AFP entry; authors agreed

typo

uniform treatmen of result

tuned variable names

more generous max number of suggestions, for more safety

changed MaSh to use SML version instead of Python version of naive Bayes by default (i.e. if MASH=yes in the settings, or 'fact_filter=mash' with no other explicit setting)

export more ML functions, for experimentation

tuned signature;

disabled IDF for now -- empirical evidence points the wrong way (as usual)

tuning

tuning

optimized computation

enabled IDF for naive Bayes ML

tuning

repaired subscript problem in SML kNN

tuning

always remove duplicates in meshing + use weights for Naive Bayes

updated naive Bayes

don't generate discriminators and selectors for 'datatype_new' unless the user asked for it

don't conceal (co)datatypes

changed '-:' to 'dead' in BNF

got rid of '=:' squiggly

use '%x. x = C' as default discriminator for nullary constructor C, instead of relying on odd '=:' syntax

renamed 'MaSh' option

document '=:' syntax better

capitalize even more carefully (see 5ac67041ccf8), e.g. relevant for option "z3_non_commercial" and prospective "MaSh";

tuned;

clarified copy_file: allow to rename base name, e.g. relevant for 'display_drafts';

support for regular Windows TeX installation;

more portable file names;

more portable -- accomodate MiKTeX on Windows;

receovered alternative abbrevs for \<open> \<close> from 8e8243975860, to accommodate national keyboard layouts where "`" might be hard to produce;

strip trailing white space, to avoid notorious problems of jEdit with last line;

added fifth member to BNF team

removed noise

fixed semantics of 'linearize'

automatically reload state file when it changes on disk

tuned

tuned names

tuned signature

moved const declaration further down in bootstrap hierarchy: keep Named_Target free of low-level stuff

unused

tuned

more uniform order of operations; tuned names

common background_abbrev operation

tuned signature

tuned: prefer separate function trails for locales and classes rather than ad-hoc case distinction

compactified

include Nominal2 keywords -- Proof General legacy;

another attempt to revive isatest -- reverting 801c01004a21;

avoid slow inspection of proof terms now that dependencies are stored in 'state'

properly mark relearns as dirty

disable weights that cause more harm than they help in kNN

add self dependency to naive Bayes

make MaSh Python the default when passing 'fact_filter = mash' without enabling the 'maSh' Isabelle system option

compactified level discriminator

properly reconstruct helpers in Z3 proofs

reverted '|' features in MaSh -- these sounded like a good idea but never really worked

tuning

shorten Sledgehammer output, as suggested by Andrei Popescu

until naive Bayes supports weights, don't incorporate 'extra' low-weight features

spell-checker completion is restricted to explicit mode, to avoid odd effects with immediate edits vs. delayed language context markup, and occasional delays due to dictionary lookup of many variants;

merged

updated to scala-2.11.1, with full uncensored classpath;

updated cygwin more thoroughly;

document property 'sel_map'

generate 'sel_map[simp]' theorem for (co)datatypes and tuning 'disc_map_iff'

merged

more uniform Font_Info.Zoom_Box; misc tuning and clarification;

added zoom box, like for outer output windows;

tuned signature;

consolidate "break_thm" and "break_term" attributes into "simp_break";

docs

added comment

move exhaust first, for technical reasons

avoid markup-generating @{make_string}

generalized Bochner integral over infinite sums

unused;

obsolete;

approximative update of versions;

incorporate isabelle.graphview into Pure.jar, which saves 20..30s build time; discontinued pointless "isabelle graphview" command-line tool (Proof General legacy);

remove stray println;

CONTRIBUTORS

added naive Bayes ML implementation, due to Cezary Kaliszyk (like k-NN)

added Isabelle system option 'mash'

updated cygwin;

afford strict check (see also AFP/a8e08d947f0a);

add various lemmas

merged

merged