davidnemeti/sarcasm

Sarcasm is an SDK for creating typesafe, domain-bound grammars. It also has a grammar driven general unparser.

Sarcasm

Summary

Sarcasm is an SDK for creating typesafe, domain-bound grammars. It also has a grammar driven general unparser.

Sarcasm provides the following features:

• Index-free AST building using domain-grammar bindings.
• Typesafe AST building using typesafe grammar.
• Automatic unparsing provided by the general unparser based on the grammar.

In a Sarcasm's domain-bound grammar you can write this e.g.:

binaryExpression.Rule =
    expression.BindTo(binaryExpression, t => t.Term1)
    + binaryOperator.BindTo(binaryExpression, t => t.Op)
    + expression.BindTo(binaryExpression, t => t.Term2)
    ;

For more information see the Documentation. You can start with Introduction.

---

License information

Sarcasm has been released under the GNU Lesser General Public License (LGPL). It means that you can use it freely as a library even in your prorietary (closed source) software. However, if you copy or modify the whole source code or part of it, that derivate work should also be released under the LGPL.

Hint: if you want to use Sarcasm in your proprietary software, but you need to modify Sarcasm's source code in order to get some extra features, modified behavior or bugfixes, you can simply clone/fork Sarcasm repository, do the changes in the clone repository, release it under LGPL, and use that modified library in your proprietary software, while keeping your proprietary software's source code closed. You can even ask for a pull request to the original Sarcasm repository if your changes are bugfixes, or you find that your extra features, modified behavior should be in the original Sarcasm SDK.

For legal license see License.

For precise, legal license information see License descriptions.

For usable, comprehensible information see LGPL on Wikipedia.

As of Jan 31, 2023, Irony is released under the MIT License, which is even more permissive than LGPL. For more and up to date information see Irony's License. About the MIT License see MIT License on Wikipedia.

sarcasm's Issues

Handle StringOptions.AllowsDoubledQuote in unparser

Speed parsing up by making the structure of the grammar rules simpler

Contract the rule-chained bnfiterms together if there is only one bnfiterm on the right side of the rule; for BnfiTermCopy and BnfiTermMember (just like it is done for BnfiTermCollection and BnfiTermValue).

MultiParser for parsing every snippets

Proxies and partial invalidation for unparser

Store obj in every utokens

Table format in unparse

Handle embedded tables by designating the "owner" of the table.

Store parentUnparsableObject, prevUnparsableObject and nextUnparsableObject in utoken's unparsableObject

Note that the unparser should have the ability to turn off this feature (actually turned off by default), since storing these references means that we keep the whole unparsed tree in memory during unparse.

Maybe we should store first/last childUnparsableObject as well.

Should use state in unparser instead of params

Parse utokens

No formatting for parallel unparse in certain situations

UnparseHint in grammar

Image can be given in utoken in addition to text

Parallel unparse using Task Parallel Library (TPL)

Parse/unparse of comments and store them in AST

Check inverseValueConverterForUnparse and UtokenizerForUnparse in BnfiTermValue whether any of them is not null

Check for typesafety in unit tests

Utoken revamp

UtokenWhitespace.Text (non parsable, non editable)

Simple rule for "if-else" in grammar (?)

Create NuGet package

Create NuGet package for modified Irony as well.

Eliminate "recursive yield" during unparse

This will speed up unparse.

http://stackoverflow.com/questions/3969963/when-not-to-use-yield-return

Change library to portable

Changing it to PCL needs Irony to be changed to PCL as well.

Introduce and use InsertUnder for IndentBlock and remove yin-yang handling

Unparse directly to utokens by delegate

UtokenizerForUnparse should be a list of bnfterm/utoken mix