| Age | Commit message (Collapse) | Author |
|
Add some javadoc comments
|
|
Several packages were moved from BJC-Utils to esodata. This makes the
necessary changes
|
|
Clean up some warnings that were hanging around
|
|
|
|
|
|
|
|
|
|
This refactors the front-end to use a tree for capturing errors, instead
of throwing exceptions.
This has the benefit that you will receive notifications about all of
the error messages you have, instead of only the first.
I'm a bit fuzzy on the details, since it's been a while since I wrote
these changes.
|
|
|
|
|
|
|
|
|
|
GenerationState now uses ReportWriter instead of StringBuilder. This
will make certain upcoming features (namely, formatting strings) work
much better
|
|
This is to prepare for replacing contents with a ReportWriter
|
|
It now works, but is called find-replace-rule instead. There is also
reject-rule, to say to reject any rule that doesn't match a provided
regex.
NOTE: For providing the regexes, if they contain spaces, they must be
quoted.
|
|
|
|
This simplifies the internal way affixes are applied, as well as adding a new circumfix-with pragma
|
|
Enable autovivifying variables. These will have their definition
automatically ran when they are first referenced.
|
|
This makes the two maps in GenerationState private. Use the
defineVar/findVar or defineRuleVar/findRuleVar instead.
This is another change on the way to doing autovivify
|
|
This makes variables scoped by grammar. Previously, you could set a
variable in one grammar, and access it from any grammar, as long as the
rule case that defined it was executed before hand. Now, only cases in
that grammar can use it.
There are still issues to do with the fact that variables can be used at
any time after they've been defined, even after their introductory rule
returned. I'm not certain that this is enough of an issue to bother
trying to fix it though.
The next step is an autovivify pragma for rules to solve several corner
cases, with the main one being wanting to use exhaustible variables from
prefix/suffix pragmas.
|
|
|
|
|
|
|
|
|
|
This enables a directory-scanning behavior for files.
Now, we just have to add internal names to almost all of the grammars
|
|
The overall grammar parser now uses levelSplit where possible.
|
|
|
|
Logging now uses the RGrammarLogging system so as to allow for better
customizability of how the logging works
|
|
|
|
|
|
|
|
A major refactoring to the internals of how things work. I believe that
nothing is broken
|
|
The main refactoring here is removing the type field from the various
classes, but there are a few other smaller ones.
This also contains the grounds for a refactoring on variable use
|
|
Templates should now work, though there is no syntax to reference them
from rules yet
In addition, several internal things have been changed so as to improve
code quality
|
|
More work that leads towards getting templates working
|
|
More work for getting going on templates, as well as just some
refactoring for future changes
|
|
Move config loading to a new class, in preparation for adding template
support
|
|
|
|
There is now syntax for rule references. Use @ instead of $, and use :=
to indicate you want exhaustion enabled
Use @ to refer to it
|
|
Add the capability to store regular/exhaustible rule variables
|
|
Simple/expanding variables are now marked with a $, as they are when
used.
Additionally, some debugging help was added
|
|
|
|
|
|
Blocks should now properly give absolute numbers as to where they are
from
|
|
Inline rules are now enclosed in [...], not {...} to better match other
reference types
|
|
Parser meta-tokens now use <...> syntax instead of {...} so as not to
get them confused with reference tokens.
|
|
This adds a few new syntax features, as well as allowing certain ones to
be used in pragmas.
The next syntax feature to be added will be some sort of quote feature,
to allow the inclusion of spaces where they would otherwise not be
permitted
|
|
This controls the number of times a rule can be entered into during a
single generations, and is set to 5 by default
|
|
Weighting now has a syntax. Include a {^<weight>} token to set the rule
weight
|
|
This re-adds the capability to weight rules.
However, no syntax has yet been add to perform that rule weighting
|