more liberal scanning of potentially malformed symbols;

vacuous execution after first malformed command;

clarified Command.range vs. Command.proper_range according to Scala version, which is potentially relevant for command status markup;

tuned markup;

tuned message;

clarified "bad" markup: proper body text, invent missing serial on Scala side (counting backwards);

reports with body text, not just markup;

fixed "double rev" bug that arose in situations where a % comment arose on the last line of a file without \n at the end

special code with lists no longer necessary, use sets

proper error prefixes;

more precise recover_quoted, recover_verbatim, recover_comment (cf. ML version) -- NB: context parsers expect explicit termination;

tuned message;

clarified Linear_Set.next/prev: check definedness;

merged

Improved complete lattice formalisation - no more index set.

merged

tuned proofs;

sneak message into "bad" markup as property -- to be displayed after YXML parsing;

apply all text edits to each node, before determining the resulting doc_edits -- allow several iterations to consolidate spans etc.; expand Clear edit before sending to prover; at most one full reparse of each node;

clarified undefined, unparsed, unfinished command spans; common reparse_spans, diff_commands; some support for consolidate_spans after change of perspective;

tuned;

discontinued mostly unused markup for command spans;

more visible markup of malformed input as "bad";

tuned proofs

some attempts to keep malformed syntax errors focussed, without too much red spilled onto the document view;

refined recover_spans: take visible range into account, reparse and trim results -- to improve editing experience wrt. unbalanced quotations etc.; tuned signature;

tuned signature;

more direct Linear_Set.reverse, swapping orientation of the graph; tuned;

tuned signature;

tuned;

refined recovery of scan errors: longest prefix of delimited token after failure, otherwise just one symbol;

explicit FIXME;

tuned signature;

corrected header

refined isabelle mkroot; allow empty 'theories' to facilitate generated templates;

simplified session specifications: names are taken verbatim and current directory is default;

added build option -D: include session directory and select its sessions;

discontinued obsolete "isabelle makeall";

even more permissive approximation of master_dir, which is only required to access external resources ('uses' etc.);

back to implicit commit via isabelle-process -- save image only once (cf. 181b91e1d1c1);

proper axiomatization of "mem" -- do not leave it formally unspecified;

simplified Pure bootstrap -- separate pure_setup.ML was required for Alice/ML at some point;

more casual exit back to ML toplevel, to accomodate commit in SML/NJ which continues at the saved point;

obsolete;

eliminated obsolete ISABELLE_USEDIR_OPTIONS;

updated ML settings;

add legacy target for HOL-Probability, needed by AFP/Markov_Models

proper isabelle update_keywords tool;

configure Admin as component, with its own lib/Tools;

SOMEthing went utterly wrong in 5b51ccdc8623;

more legacy targets;

discontinued obsolete IsaMakefile and ROOT.ML files from the Isabelle distribution;

tuned;

proper quoting

merged

more structural parsing for minor modes; tuned signatures;

clarified Sidekick configuration, including minor modes;

stop distinguishing between complete and incomplete slices, since this is very fragile and has hardly any useful semantics to users

specify full path to clausifier

documentation