I've in mind something akin to the F# mechanism described here.
In looking into it, I've found nothing explicitly saying it does nor that it doesn't.
Neither the Definition of Standard ML from 1990 nor the Definition of Standard ML (Revised) from 1997 defines pattern guards.
There have been some proposals to support them — see https://github.com/SMLFamily/Successor-ML/wiki/Summary-of-proposed-changes — but I think the only implementation that does so is 'HaMLeT S'. In particular, neither MLton nor SMLNJ seems to support them (see http://mlton.org/SuccessorML and https://smlnj.org/doc/features.html, respectively).
Related
I'm looking for a definition of the syntax for the Cypher query language. I tried the docs but they're very vague.
Ideally, I'd like a BNF (or any variant) definition, or one of those "graph" definitions like this or this. Really, anything resembling a formal definition.
What you are looking for will be available in openCypher. Several items will be released as part of the project, one of the first of which is the BNF grammar.
Update 2016-01-30: A first draft of the grammar is now avialable at \https://github.com/opencypher/openCypher/blob/master/grammar.ebnf.
Update: 2016-10-17: EBNF and Antlr grammars, TCK, railroad diagrams, and a list of community projects are available at http://www.opencypher.org/#resources
Take a look at the recently announced (Oct 2015) openCypher project. It involves releasing the language specification, among other things.
From the announcement:
1. Cypher reference documentation:
A comprehensive user documentation describing use of the Cypher query language with examples and tutorials.
2. Technology compatibility kit (TCK):
The TCK consists of a number of tests that a software supplier would run in order to self-certify support for a given version of Cypher.
3. Reference implementation:
Distributed under the Apache 2.0 license, the reference implementation is a fully functional implementation of key parts of the stack needed to support Cypher inside a data platform or tool. The first planned deliverable is a parser that will take a Cypher statement and parse it into an AST (abstract syntax tree) representation. The reference implementation complements the documentation and tests by providing working implementations of Cypher – which are permissively licensed – and can be used as examples or as a foundation for one’s own implementation.
4. Cypher language specification:
Licensed under a Creative Commons license, the Cypher language specification is a technical expression of the language syntax to enable parsers to auto-generate the query syntax. A full semantic specification is also planned as a part of the openCypher project.
The same announcement also says that the process is open and that it is possible to submit, review and comment on language proposals.
Update!
Neo4j has changed a lot since this answer was written. In 2017 the simple answer is yes, you can download the grammar files from https://www.opencypher.org/
Below is the old answer, which was accurate in 2014
As far as I can tell, the only formal definition is in the code. That's the bad news.
The good news is that the code uses a scala library to do the parsing which makes the code rules look kinda/sorta like BNF. And there's some documentation on how to read it.
Here's a link into a scala object that defines what a query is.
This general package on github looks to me like it contains all of the cypher command implementations, and should have everything you're asking for.
Code in this package is written in scala, and looks like this:
object Query {
def start(startItems: StartItem*) = new QueryBuilder().startItems(startItems:_*)
def matches(patterns:Pattern*) = new QueryBuilder().matches(patterns:_*)
def optionalMatches(patterns:Pattern*) = new QueryBuilder().matches(patterns:_*).makeOptional()
def updates(cmds:UpdateAction*) = new QueryBuilder().updates(cmds:_*)
def unique(cmds:UniqueLink*) = new QueryBuilder().startItems(Seq(CreateUniqueStartItem(CreateUniqueAction(cmds:_*))):_*)
(...)
This matches roughly with the upper right hand quadrant of the Cypher refcard. You can sorta see that there can be a start clause, a match clause, and so on. This includes links to other implementation classes (like UpdateAction which further define clauses considered update actions.
Make sure to also read How Neo4J Uses Scala's Parser Combinator: Cypher's Internals Part 1 for more information on what's going on here, and the mapping between the scala classes and what we'd normally consider EBNF. This blog post is old (2011) and the specific code examples it gives shouldn't be trusted, but I think it has good general information on how the implementation works, and what to look for if you want to understand the EBNF behind cypher.
Disclaimer: I'm not a scala hardcore, YMMV, IANAL, devs please overrule me if I'm wrong.
(Michael Hunger answered in a comment, so I can't accept his answer. Here's his answer:)
Cypher uses parboiled as parser, the parboiled rule DSL are pretty easy to read and understand. https://github.com/neo4j/neo4j/blob/d18583d260a957ab1a14bd27d34eb5625df42bc5/community/cypher/cypher-compiler-2.2/src/main/scala/org/neo4j/cypher/internal/compiler/v2_2/parser/Clauses.scala
None of these seem to work any more.
I don't see anything on the opencypher.org site that looks like a grammar to download.
None of the github links from Michael Hunger work.
I'd really like access to SOME resource where I can learn how to construct queries for functions like avg that allegedly take a list expression as an argument, yet barf at every variant I can figure out.
I'm currently working with implementing some MISRA C rules in Sonar. My current rule is to avoid recursion. I started with
//statement[#tokenValue=ancestor::functionDefinition/functionDeclarator/functionName/#tokenValue]
To avoid using the same function name within a function definition, but of course it is possible to use other functions with the same names, but different signatures.
Therefore I've two questions:
Is it possible to find out the method signatures (via built in xpath function, etc.)? Here, i could compare the signature with the call statement.
Is it possible to extend the plugin, as there are MISRA rules where it might be more efficient to go through the abstract syntax tree with the sourcecode?
Thank you really much for your replies:)
(ps :- are there any documentions about the SSLR C toolkit / built in xpath rules?)
The C plugin is deprecated. It has been replaced by the C/C++ plugin. See http://www.sonarsource.com/products/plugins/languages/cpp/.
Before implementing a new coding rule, you should consider whether it is specific to your own context or might benefit others. If it might benefit others, you can propose them on the developer mailing-list. If the SonarQube team find your proposed rules interesting, they may be implemented directly in the related language plugin. It means less maintenance for you, and benefit to others. See http://docs.codehaus.org/display/SONAR/Extending+Coding+Rules
Coming from the discussions about the use of vendor specific attributes in another question I asked myself, "what rules should we tell people for using attributes that are not listed in the standard"?
The two attributes that are defined are [[ noreturn ]] and [[ carries_dependencies ]]. The standard leaves open how compilers should react on unknown attributes -- thus, by the standard they may stop with an error message. This is not what e.g. GCC does, it emits a warning and continues. This is probably a behavior to be expected by the most-common compilers. For this reason I would have like to read a "should" in the standard, but we don't have it.
The paper N2553 brings up flexible attributes. It lists further attributes used by GCC (
unused, weak) and MSVC (dllimport). for OpenMP, the widely supported parallelizing framework, scoped attributes are suggested, eg. omp::for(clause, clause), omp::parallel(clause,clause). So, it is very likely that we will se some vendor specific attributes very soon after they support the syntax at all, indeed.
Therefore, when we now go "out in the world" and tell people about C++11, what should the advice be about using attributes?
Only use noreturn and carries_dependencies
Use your compilers old syntax instead, eg. __attribute__((noreturn)) and define a macro when you port the code (the current situation)
Use those attributes your favorite compiler supports freely, knowing this code might not be portable to another standard-conforming compiler, because if the standard allows a compiler to stop with an error, you have to consider this will happen. This sounds a bit like advocating writing non-portable code.
Or, my guess, expect the most-used compilers to warn about unknown attributes, so you can use vendor-specific attributes, keeping in mind that in rare cases you may get problems.
Note the slight difference in the last two bullet-items. While both say "use those attributes you need", item3's message is "do not care about other compilers", while item4 implicitly rephrases the standard texts "implementation defined behavior" to "the compiler should emit a diagnostic message".
What could be the suggestion for an upcoming Best Practice here?
The best practice — the only one that is reasonably portable in practical terms, never mind ambiguity in the Standard — is to use macros. It will be many years before we can forget about compilers that don't support attributes.
The number of compilers and the number of custom __keywords__ defined by those compilers will always be increasing, and it makes sense for the language to define a way to contain the damage. It doesn't need to revolutionize the way people write unportable code, or make unportable code portable (although standard attributes do that). There is a benefit simply to giving caffeine-addled compiler backend engineers a sandbox for when they want to extend the grammar.
It is a bit alarming, though, that no attribute tokens are reserved to the implementation, or to the language besides the ones currently standard. So there will be trouble when they decide to standardize more of them.
I am learning my way around Scheme, and I am especially interested in how the language is constructed. I'm trying to find a nice description of the core syntax for a Scheme implementation. I don't know enough about the standards, but I assume that they all contain macro systems. If not, I'd like to read about a standard that also includes macros (they can't possibly be implemented in simpler Scheme constructs, can they?).
Does anyone have a good reference for the minimal syntax needed for a Scheme dialect?
Just an update:
I also stumbled upon this: http://matt.might.net/articles/compiling-to-java/#sec1. If you also add define-syntax and delay then it seems like it might be a good start.
In the R5RS specification, the following page appears to be what I was looking for: formal syntax
Although it may be a bit dry, you should read over the R5RS spec or the R6RS spec.
The docs really do not take that long to read through and you can just skim most of the sections until you need more detail. But either document does cover all of the minimal syntax required, including macros.
Why didn't C++0x deprecate implicit conversions for user defined types a.k.a. objects? Is there any project which actually uses this (mis)feature? Whenever I see a single argument constructor in a code I get to review or modify I treat it as bug and make it explicit. So far it worked well and nobody complained.
Thank you.
EDIT: Let me quote Alex Stepanov, the creator of STL:
Open your C++ book and read about the
explicit keyword! Also petition your
neighborhood C++ standard committee
member to finally abolish implicit
conversions. There is a common
misconception, often propagated by
people who should know better, that
STL depends on implicit conversions.
Not so!
Reference: A. Stepanov. C++ notes
EDIT AGAIN: No, no debate plz. I am just curious whether anyone uses implicit conversions in their work. I never seen any project which would allow implicit conversion for objects. I thought hard and couldn't come with any hypothetical scenario where implicit conversion wouldn't become a minefield. I mean C++ single argument conversions, not float->double or similar conversions inherited from C.
The obvious answer is that code written and working in C++03 is supposed to continue working with C++0x compilers.
For one thing, it would be a hugely breaking change to remove implicit conversion from the language - even if it were made optional and off-by-default with an implicit keyword.
I've done a search of comp.std.c++ and it doesn't seem to have been discussed at all in that group - though there have been some questions on the subject, no-one seems to have suggested going so far as removing it. I would certainly not go so far either: it's a feature I happily use on occasion and I do not subscribe to making all possibly-converting constructors explicit either - unless it causes real bugs.