1 R Internal Structures

This chapter is the beginnings of documentation about R internal structures. It is written for the core team and others studying the code in the src/main directory.

It is a work-in-progress and should be checked against the current version of the source code. Versions for R 2.x.y contain historical comments about when features were introduced: this version is for the 3.x.y series.

1.1 SEXPs

What R users think of as variables or objects are symbols which are bound to a value. The value can be thought of as either a SEXP (a pointer), or the structure it points to, a SEXPREC (and there are alternative forms used for vectors, namely VECSXP pointing to VECTOR_SEXPREC structures). So the basic building blocks of R objects are often called nodes, meaning SEXPRECs or VECTOR_SEXPRECs.

Note that the internal structure of the SEXPREC is not made available to R Extensions: rather SEXP is an opaque pointer, and the internals can only be accessed by the functions provided.

Both types of node structure have as their first three fields a 64-bit sxpinfo header and then three pointers (to the attributes and the previous and next node in a doubly-linked list), and then some further fields. On a 32-bit platform a node¹ occupies 32 bytes: on a 64-bit platform typically 56 bytes (depending on alignment constraints).

¹ strictly, a SEXPREC node; VECTOR_SEXPREC nodes are slightly smaller but followed by data in the node.

The first five bits of the sxpinfo header specify one of up to 32 SEXPTYPEs.

1.1.1 SEXPTYPEs

Currently SEXPTYPEs 0:10 and 13:25 are in use. Values 11 and 12 were used for internal factors and ordered factors and have since been withdrawn. Note that the SEXPTYPE numbers are stored in saved objects and that the ordering of the types is used, so the gap cannot easily be reused.

no SEXPTYPE Description

0 NILSXP NULL

1 SYMSXP symbols

2 LISTSXP pairlists

3 CLOSXP closures

4 ENVSXP environments

5 PROMSXP promises

6 LANGSXP language objects

7 SPECIALSXP special functions

8 BUILTINSXP builtin functions

9 CHARSXP internal character strings

10 LGLSXP logical vectors

13 INTSXP integer vectors

14 REALSXP numeric vectors

15 CPLXSXP complex vectors

16 STRSXP character vectors

17 DOTSXP dot-dot-dot object

18 ANYSXP make “any” args work

19 VECSXP list (generic vector)

20 EXPRSXP expression vector

21 BCODESXP byte code

22 EXTPTRSXP external pointer

23 WEAKREFSXP weak reference

24 RAWSXP raw vector

25 OBJSXP objects not of simple type

no	SEXPTYPE	Description
`0`	`NILSXP`	`NULL`
`1`	`SYMSXP`	symbols
`2`	`LISTSXP`	pairlists
`3`	`CLOSXP`	closures
`4`	`ENVSXP`	environments
`5`	`PROMSXP`	promises
`6`	`LANGSXP`	language objects
`7`	`SPECIALSXP`	special functions
`8`	`BUILTINSXP`	builtin functions
`9`	`CHARSXP`	internal character strings
`10`	`LGLSXP`	logical vectors
`13`	`INTSXP`	integer vectors
`14`	`REALSXP`	numeric vectors
`15`	`CPLXSXP`	complex vectors
`16`	`STRSXP`	character vectors
`17`	`DOTSXP`	dot-dot-dot object
`18`	`ANYSXP`	make “any” args work
`19`	`VECSXP`	list (generic vector)
`20`	`EXPRSXP`	expression vector
`21`	`BCODESXP`	byte code
`22`	`EXTPTRSXP`	external pointer
`23`	`WEAKREFSXP`	weak reference
`24`	`RAWSXP`	raw vector
`25`	`OBJSXP`	objects not of simple type

Many of these will be familiar from R level: the atomic vector types are LGLSXP, INTSXP, REALSXP, CPLXSP, STRSXP and RAWSXP. Lists are VECSXP and names (also known as symbols) are SYMSXP. Pairlists (LISTSXP, the name going back to the origins of R as a Scheme-like language) are rarely seen at R level, but are for example used for argument lists. Character vectors are effectively lists all of whose elements are CHARSXP, a type that is rarely visible at R level.

Language objects (LANGSXP) are calls (including formulae and so on). Internally they are pairlists with first element a reference² to the function to be called with remaining elements the actual arguments for the call (and with the tags if present giving the specified argument names). Although this is not enforced, many places in the code assume that the pairlist is of length one or more, often without checking.

² a pointer to a function or a symbol to look up the function by name, or a language object to be evaluated to give a function.

Expressions are of type EXPRSXP: they are a vector of (usually language) objects most often seen as the result of parse().

The functions are of types CLOSXP, SPECIALSXP and BUILTINSXP: where SEXPTYPEs are stored in an integer these are sometimes lumped into a pseudo-type FUNSXP with code 99. Functions defined via function are of type CLOSXP and have formals, body and environment.

The SEXPTYPE OBJSXP is for S4 and other objects which do not consist solely of a simple type such as an atomic vector or function.

1.1.2 Rest of header

Note that the size and structure of the header changed in R 3.5.0: see earlier editions of this manual for the previous layout.

The sxpinfo header is defined as a 64-bit C structure by

#define NAMED_BITS 16
struct sxpinfo_struct {
    SEXPTYPE type      :  5;  /* discussed above */
    unsigned int scalar:  1;  /* is this a numeric vector of length 1?
    unsigned int obj   :  1;  /* is this an object with a class attribute? */
    unsigned int alt   :  1;  /* is this an ALTREP object? */
    unsigned int gp    : 16;  /* general purpose, see below */
    unsigned int mark  :  1;  /* mark object as ‘in use’ in GC */
    unsigned int debug :  1;
    unsigned int trace :  1;
    unsigned int spare :  1;  /* debug once and with reference counting */
    unsigned int gcgen :  1;  /* generation for GC */
    unsigned int gccls :  3;  /* class of node for GC */
    unsigned int named : NAMED_BITS; /* used to control copying */
    unsigned int extra : 32 - NAMED_BITS;
}; /*           Tot: 64 */

The debug bit is used for closures and environments. For closures it is set by debug() and unset by undebug(), and indicates that evaluations of the function should be run under the browser. For environments it indicates whether the browsing is in single-step mode.

The trace bit is used for functions for trace() and for other objects when tracing duplications (see tracemem).

The spare bit is used for closures to mark them for one-time debugging.

The named field is set and accessed by the SET_NAMED and NAMED macros, and takes values 0, 1 and 2, or possibly higher if NAMEDMAX is set to a higher value. R has a ‘call by value’ illusion, so an assignment like

b <- a

[The NAMED mechanism has been replaced by reference counting.]

appears to make a copy of a and refer to it as b. However, if neither a nor b are subsequently altered there is no need to copy. What really happens is that a new symbol b is bound to the same value as a and the named field on the value object is set (in this case to 2). When an object is about to be altered, the named field is consulted. A value of 2 or more means that the object must be duplicated before being changed. (Note that this does not say that it is necessary to duplicate, only that it should be duplicated whether necessary or not.) A value of 0 means that it is known that no other SEXP shares data with this object, and so it may safely be altered. A value of 1 is used for situations like

dim(a) <- c(7, 2)

where in principle two copies of a exist for the duration of the computation as (in principle)

a <- `dim<-`(a, c(7, 2))

but for no longer, and so some primitive functions can be optimized to avoid a copy in this case.

The gp bits are by definition ‘general purpose’. We label these from 0 to 15. Bits 0–5 and bits 14–15 have been used as described below (mainly from detective work on the sources).

The bits can be accessed and set by the LEVELS and SETLEVELS macros, which names appear to date back to the internal factor and ordered types and are now used in only a few places in the code. The gp field is serialized/unserialized for the SEXPTYPEs other than NILSXP, SYMSXP and ENVSXP.

Bits 14 and 15 of gp are used for ‘fancy bindings’. Bit 14 is used to lock a binding or an environment, and bit 15 is used to indicate an active binding. (For the definition of an ‘active binding’ see the header comments in file src/main/envir.c.) Bit 15 is used for an environment to indicate if it participates in the global cache.

The macros ARGUSED and SET_ARGUSED are used when matching actual and formal function arguments, and take the values 0, 1 and 2.

The macros MISSING and SET_MISSING are used for pairlists of arguments. Four bits are reserved, but only two are used (and exactly what for is not explained). It seems that bit 0 is used by matchArgs_NR to mark missingness on the returned argument list, and bit 1 is used to mark the use of a default value for an argument copied to the evaluation frame of a closure.

Bit 0 is used by macros DDVAL and SET_DDVAL. This indicates that a SYMSXP is one of the symbols ..n which are implicitly created when ... is processed, and so indicates that it may need to be looked up in a DOTSXP.

Bit 0 is used for PRSEEN, a flag to indicate if a promise has already been seen during the evaluation of the promise (and so to avoid recursive loops).

Bit 0 is used for HASHASH, on the PRINTNAME of the TAG of the frame of an environment. (This bit is not serialized for CHARSXP objects.)

Bits 0 and 1 are used for weak references (to indicate ‘ready to finalize’, ‘finalize on exit’).

Bit 0 is used by the condition handling system (on a VECSXP) to indicate a calling handler.

Bit 4 is turned on to mark S4 objects.

Bits 1, 2, 3, 5 and 6 are used for a CHARSXP to denote its encoding. Bit 1 indicates that the CHARSXP should be treated as a set of bytes, not necessarily representing a character in any known encoding. Bits 2, 3 and 6 are used to indicate that it is known to be in Latin-1, UTF-8 or ASCII respectively.

Bit 5 for a CHARSXP indicates that it is hashed by its address, that is NA_STRING or is in the CHARSXP cache (this is not serialized). Only exceptionally is a CHARSXP not hashed, and this should never happen in end-user code.

1.1.3 The ‘data’

A SEXPREC is a C structure containing the 64-bit header as described above, three pointers (to the attributes, previous and next node) and the node data, a union

union {
    struct primsxp_struct primsxp;
    struct symsxp_struct symsxp;
    struct listsxp_struct listsxp;
    struct envsxp_struct envsxp;
    struct closxp_struct closxp;
    struct promsxp_struct promsxp;
} u;

All of these alternatives apart from the first (an int) are three pointers, so the union occupies three words.

The vector types are RAWSXP, CHARSXP, LGLSXP, INTSXP, REALSXP, CPLXSXP, STRSXP, VECSXP, EXPRSXP and WEAKREFSXP. Remember that such types are a VECTOR_SEXPREC, which again consists of the header and the same three pointers, but followed by two integers giving the length and ‘true length’³ of the vector, and then followed by the data (aligned as required: on most 32-bit systems with a 28-byte VECTOR_SEXPREC node the data can follow immediately after the node). The data are a block of memory of the appropriate length to store ‘true length’ elements (rounded up to a multiple of 8 bytes, with the 8-byte blocks being the ‘Vcells’ referred in the documentation for gc()).

³ The only current use is for hash tables of environments (VECSXPs), where length is the size of the table and truelength is the number of primary slots in use, for the reference hash tables in serialization (VECSXPs), and for ‘growable’ vectors (atomic vectors, VECSXPs and EXPRSXPs) which are created by slightly over-committing when enlarging a vector during subassignment, so that some number of the following enlargements during subassignment can be performed in place), where truelength is the number of slots in use.

The ‘data’ for the various types are given in the table below. A lot of this is interpretation, i.e. the types are not checked.

NILSXP: There is only one object of type NILSXP, R_NilValue, with no data.
SYMSXP: Pointers to three nodes, the name, value and internal, accessed by PRINTNAME (a CHARSXP), SYMVALUE and INTERNAL. (If the symbol’s value is a .Internal function, the last is a pointer to the appropriate SEXPREC.) Many symbols have SYMVALUE R_UnboundValue.
LISTSXP: Pointers to the CAR, CDR (usually a LISTSXP or NULL) and TAG (a SYMSXP or NULL).
CLOSXP: Pointers to the formals (a pairlist), the body and the environment.
ENVSXP: Pointers to the frame, enclosing environment and hash table (NULL or a VECSXP). A frame is a tagged pairlist with tag the symbol and CAR the bound value.
PROMSXP: Pointers to the value, expression and environment (in which to evaluate the expression). Once an promise has been evaluated, the environment is set to NULL.
LANGSXP: A special type of LISTSXP used for function calls. (The CAR references the function (perhaps via a symbol or language object), and the CDR the argument list with tags for named arguments.) R-level documentation references to ‘expressions’ / ‘language objects’ are mainly LANGSXPs, but can be symbols (SYMSXPs) or expression vectors (EXPRSXPs).
SPECIALSXP
BUILTINSXP: An integer giving the offset into the table of primitives/.Internals.
CHARSXP: length, truelength followed by a block of bytes (allowing for the nul terminator).
LGLSXP
INTSXP: length, truelength followed by a block of C ints (which are 32 bits on all R platforms).
REALSXP: length, truelength followed by a block of C doubles.
CPLXSXP: length, truelength followed by a block of C99 double complexs.
STRSXP: length, truelength followed by a block of pointers (SEXPs pointing to CHARSXPs).
DOTSXP: A special type of LISTSXP for the value bound to a ... symbol: a pairlist of promises.
ANYSXP: This is used as a place holder for any type: there are no actual objects of this type.
VECSXP
EXPRSXP: length, truelength followed by a block of pointers. These are internally identical (and identical to STRSXP) but differ in the interpretations placed on the elements.
BCODESXP: For the ‘byte-code’ objects generated by the compiler.
EXTPTRSXP: Has three pointers, to the pointer, the protection value (an R object which if alive protects this object) and a tag (a SYMSXP?).
WEAKREFSXP: A WEAKREFSXP is a special VECSXP of length 4, with elements key, value, finalizer and next. The key is NULL, an environment or an external pointer, and the finalizer is a function or NULL.
RAWSXP: length, truelength followed by a block of bytes.
OBJSXP: two unused pointers and a tag.

1.1.4 Allocation classes

As we have seen, the field gccls in the header is three bits to label up to 8 classes of nodes. Non-vector nodes are of class 0, and ‘small’ vector nodes are of classes 1 to 5, with a class for custom allocator vector nodes 6 and ‘large’ vector nodes being of class 7. The ‘small’ vector nodes are able to store vector data of up to 8, 16, 32, 64 and 128 bytes: larger vectors are malloc-ed individually whereas the ‘small’ nodes are allocated from pages of about 2000 bytes. Vector nodes allocated using custom allocators (via allocVector3) are not counted in the GC memory usage statistics since their memory semantics is not under R’s control and may be non-standard (e.g., memory could be partially shared across nodes).

1.2 Environments and variable lookup

What users think of as ‘variables’ are symbols which are bound to objects in ‘environments’. The word ‘environment’ is used ambiguously in R to mean either the frame of an ENVSXP (a pairlist of symbol-value pairs) or an ENVSXP, a frame plus an enclosure.

There are additional places that ‘variables’ can be looked up, called ‘user databases’ in comments in the code. These seem undocumented in the R sources, but apparently refer to the RObjectTable package formerly available at https://www.omegahat.net/RObjectTables/.

The base environment is special. There is an ENVSXP environment with enclosure the empty environment R_EmptyEnv, but the frame of that environment is not used. Rather its bindings are part of the global symbol table, being those symbols in the global symbol table whose values are not R_UnboundValue. When R is started the internal functions are installed (by C code) in the symbol table, with primitive functions having values and .Internal functions having what would be their values in the field accessed by the INTERNAL macro. Then .Platform and .Machine are computed and the base package is loaded into the base environment followed by the system profile.

The frames of environments (and the symbol table) are normally hashed for faster access (including insertion and deletion).

By default R maintains a (hashed) global cache of ‘variables’ (that is symbols and their bindings) which have been found, and this refers only to environments which have been marked to participate, which consists of the global environment (aka the user workspace), the base environment plus environments⁴ which have been attached. When an environment is either attached or detached, the names of its symbols are flushed from the cache. The cache is used whenever searching for variables from the global environment (possibly as part of a recursive search).

⁴ Remember that attaching a list or a saved image actually creates and populates an environment and attaches that.

1.2.1 Search paths

S has the notion of a ‘search path’: the lookup for a ‘variable’ leads (possibly through a series of frames) to the ‘session frame’ the ‘working directory’ and then along the search path. The search path is a series of databases (as returned by search()) which contain the system functions (but not necessarily at the end of the path, as by default the equivalent of packages are added at the end).

R has a variant on the S model. There is a search path (also returned by search()) which consists of the global environment (aka user workspace) followed by environments which have been attached and finally the base environment. Note that unlike S it is not possible to attach environments before the workspace nor after the base environment.

However, the notion of variable lookup is more general in R, hence the plural in the title of this subsection. Since environments have enclosures, from any environment there is a search path found by looking in the frame, then the frame of its enclosure and so on. Since loops are not allowed, this process will eventually terminate: it can terminate at either the base environment or the empty environment. (It can be conceptually simpler to think of the search always terminating at the empty environment, but with an optimization to stop at the base environment.) So the ‘search path’ describes the chain of environments which is traversed once the search reaches the global environment.

1.2.2 Namespaces

Namespaces are environments associated with packages (and once again the base package is special and will be considered separately). A package pkg defines two environments namespace:pkg and package:pkg: it is package:pkg that can be attached and form part of the search path.

The objects defined by the R code in the package are symbols with bindings in the namespace:pkg environment. The package:pkg environment is populated by selected symbols from the namespace:pkg environment (the exports). The enclosure of this environment is an environment populated with the explicit imports from other namespaces, and the enclosure of that environment is the base namespace. (So the illusion of the imports being in the namespace environment is created via the environment tree.) The enclosure of the base namespace is the global environment, so the search from a package namespace goes via the (explicit and implicit) imports to the standard ‘search path’.

The base namespace environment R_BaseNamespace is another ENVSXP that is special-cased. It is effectively the same thing as the base environment R_BaseEnv except that its enclosure is the global environment rather than the empty environment: the internal code diverts lookups in its frame to the global symbol table.

1.2.3 Hash table

Environments in R usually have a hash table, and nowadays that is the default in new.env(). It is stored as a VECSXP where length is used for the allocated size of the table and truelength is the number of primary slots in use—the pointer to the VECSXP is part of the header of a SEXP of type ENVSXP, and this points to R_NilValue if the environment is not hashed.

For the pros and cons of hashing, see a basic text on Computer Science.

The code to implement hashed environments is in src/main/envir.c. Unless set otherwise (e.g. by the size argument of new.env()) the initial table size is 29. The table will be resized by a factor of 1.2 once the load factor (the proportion of primary slots in use) reaches 85%.

The hash chains are stored as pairlist elements of the VECSXP: items are inserted at the front of the pairlist. Hashing is principally designed for fast searching of environments, which are from time to time added to but rarely deleted from, so items are not actually deleted but have their value set to R_UnboundValue.

1.3 Attributes

As we have seen, every SEXPREC has a pointer to the attributes of the node (default R_NilValue). The attributes can be accessed/set by the macros/functions ATTRIB and SET_ATTRIB, but such direct access is normally only used to check if the attributes are NULL or to reset them. Otherwise access goes through the functions getAttrib and setAttrib which impose restrictions on the attributes. One thing to watch is that if you copy attributes from one object to another you may (un)set the "class" attribute and so need to copy the object and S4 bits as well. There is a macro/function DUPLICATE_ATTRIB to automate this.

Note that the ‘attributes’ of a CHARSXP are used as part of the management of the CHARSXP cache: of course CHARSXPs are not user-visible but C-level code might look at their attributes.

The code assumes that the attributes of a node are either R_NilValue or a pairlist of non-zero length (and this is checked by SET_ATTRIB). The attributes are named (via tags on the pairlist). The replacement function attributes<- ensures that "dim" precedes "dimnames" in the pairlist. Attribute "dim" is one of several that is treated specially: the values are checked, and any "names" and "dimnames" attributes are removed. Similarly, you cannot set "dimnames" without having set "dim", and the value assigned must be a list of the correct length and with elements of the correct lengths (and all zero-length elements are replaced by NULL).

The other attributes which are given special treatment are "names", "class", "tsp", "comment" and "row.names". For pairlist-like objects the names are not stored as an attribute but (as symbols) as the tags: however the R interface makes them look like conventional attributes, and for one-dimensional arrays they are stored as the first element of the "dimnames" attribute. The C code ensures that the "tsp" attribute is an REALSXP, the frequency is positive and the implied length agrees with the number of rows of the object being assigned to. Classes and comments are restricted to character vectors, and assigning a zero-length comment or class removes the attribute. Setting or removing a "class" attribute sets the object bit appropriately. Integer row names are converted to and from the internal compact representation.

Care needs to be taken when adding attributes to objects of the types with non-standard copying semantics. There is only one object of type NILSXP, R_NilValue, and that should never have attributes (and this is enforced in installAttrib). For environments, external pointers and weak references, the attributes should be relevant to all uses of the object: it is for example reasonable to have a name for an environment, and also a "path" attribute for those environments populated from R code in a package.

When should attributes be preserved under operations on an object? Becker, Chambers & Wilks (1988, pp. 144–6) give some guidance. Scalar functions (those which operate element-by-element on a vector and whose output is similar to the input) should preserve attributes (except perhaps class, and if they do preserve class they need to preserve the OBJECT and S4 bits). Binary operations normally call copyMostAttrib to copy most attributes from the longer argument (and if they are of the same length from both, preferring the values on the first). Here ‘most’ means all except the names, dim and dimnames which are set appropriately by the code for the operator.

Subsetting (other than by an empty index) generally drops all attributes except names, dim and dimnames which are reset as appropriate. On the other hand, subassignment generally preserves such attributes even if the length is changed. Coercion drops all attributes. For example:

> x <- structure(1:8, names=letters[1:8], comm="a comment")
> x[]
a b c d e f g h
1 2 3 4 5 6 7 8
attr(,"comm")
[1] "a comment"
> x[1:3]
a b c
1 2 3
> x[3] <- 3
> x
a b c d e f g h
1 2 3 4 5 6 7 8
attr(,"comm")
[1] "a comment"
> x[9] <- 9
> x
a b c d e f g h
1 2 3 4 5 6 7 8 9
attr(,"comm")
[1] "a comment"

1.4 Contexts

Contexts are the internal mechanism used to keep track of where a computation has got to (and from where), so that control-flow constructs can work and reasonable information can be produced on error conditions (such as via traceback), and otherwise (the sys.xxx functions).

Execution contexts are a stack of C structs:

typedef struct RCNTXT {
    struct RCNTXT *nextcontext; /* The next context up the chain */
    int callflag;               /* The context ‘type’ */
    JMP_BUF cjmpbuf;            /* C stack and register information */
    int cstacktop;              /* Top of the pointer protection stack */
    int evaldepth;              /* Evaluation depth at inception */
    SEXP promargs;              /* Promises supplied to closure */
    SEXP callfun;               /* The closure called */
    SEXP sysparent;             /* Environment the closure was called from */
    SEXP call;                  /* The call that effected this context */
    SEXP cloenv;                /* The environment */
    SEXP conexit;               /* Interpreted on.exit code */
    void (*cend)(void *);       /* C on.exit thunk */
    void *cenddata;             /* Data for C on.exit thunk */
    char *vmax;                 /* Top of the R_alloc stack */
    int intsusp;                /* Interrupts are suspended */
    SEXP handlerstack;          /* Condition handler stack */
    SEXP restartstack;          /* Stack of available restarts */
    struct RPRSTACK *prstack;   /* Stack of pending promises */
} RCNTXT, *context;

plus additional fields for the byte-code compiler. The ‘types’ are from

enum {
    CTXT_TOPLEVEL = 0,  /* toplevel context */
    CTXT_NEXT     = 1,  /* target for next */
    CTXT_BREAK    = 2,  /* target for break */
    CTXT_LOOP     = 3,  /* break or next target */
    CTXT_FUNCTION = 4,  /* function closure */
    CTXT_CCODE    = 8,  /* other functions that need error cleanup */
    CTXT_RETURN   = 12, /* return() from a closure */
    CTXT_BROWSER  = 16, /* return target on exit from browser */
    CTXT_GENERIC  = 20, /* rather, running an S3 method */
    CTXT_RESTART  = 32, /* a call to restart was made from a closure */
    CTXT_BUILTIN  = 64  /* builtin internal function */
};

where the CTXT_FUNCTION bit is on wherever function closures are involved.

Contexts are created by a call to begincontext and ended by a call to endcontext: code can search up the stack for a particular type of context via findcontext (and jump there) or jump to a specific context via R_JumpToContext. R_ToplevelContext is the ‘idle’ state (normally the command prompt), and R_GlobalContext is the top of the stack.

Note that whilst calls to closures set a context, internal functions never do and primitive builtins only set it when profiling or when they are interfaces to foreign functions.

The byte-code compiler generates a map of instructions to source references and expressions at compile time, which allows to produce information on error conditions. As an optimization, the byte-code interpreter then does not set a context in some cases, such as in simple loops or when inlining simple builtins or wrappers for internal functions.

Dispatching from a S3 generic (via UseMethod or its internal equivalent) or calling NextMethod sets the context type to CTXT_GENERIC. This is used to set the sysparent of the method call to that of the generic, so the method appears to have been called in place of the generic rather than from the generic.

The R sys.frame and sys.call functions work by counting calls to closures (type CTXT_FUNCTION) from either end of the context stack.

Note that the sysparent element of the structure is not the same thing as sys.parent(). Element sysparent is primarily used in managing changes of the function being evaluated, i.e. by Recall and method dispatch.

CTXT_CCODE contexts are currently used in cat(), load(), scan() and write.table() (to close the connection on error), by PROTECT, serialization (to recover from errors, e.g. free buffers) and within the error handling code (to raise the C stack limit and reset some variables).

1.5 Argument evaluation

As we have seen, functions in R come in three types, closures (SEXPTYPE CLOSXP), specials (SPECIALSXP) and builtins (BUILTINSXP). In this section we consider when (and if) the actual arguments of function calls are evaluated. The rules are different for the internal (special/builtin) and R-level functions (closures).

For a call to a closure, the actual and formal arguments are matched and a matched call (another LANGSXP) is constructed. This process first replaces the actual argument list by a list of promises to the values supplied. It then constructs a new environment which contains the names of the formal parameters matched to actual or default values: all the matched values are promises, the defaults as promises to be evaluated in the environment just created. That environment is then used for the evaluation of the body of the function, and promises will be forced (and hence actual or default arguments evaluated) when they are encountered. (Evaluating a promise sets NAMED = NAMEDMAX on its value, so if the argument was a symbol its binding is regarded as having multiple references during the evaluation of the closure call.) [The NAMED mechanism has been replaced by reference counting.]

If the closure is an S3 generic (that is, contains a call to UseMethod) the evaluation process is the same until the UseMethod call is encountered. At that point the argument on which to do dispatch (normally the first) will be evaluated if it has not been already. If a method has been found which is a closure, a new evaluation environment is created for it containing the matched arguments. (Note that this means changes to the values of the formal arguments in the body of the generic are discarded when calling the method, but actual argument promises which have been forced retain the values found when they were forced. On the other hand, missing arguments have values which are promises to use the default supplied by the method and not by the generic.) If the method found is a primitive it is called with the matched argument list of promises (possibly already forced) used for the generic.

The essential difference⁵ between special and builtin functions is that the arguments of specials are not evaluated before the C code is called, and those of builtins are. Note that being a special/builtin is separate from being primitive or .Internal: quote is a special primitive, + is a builtin primitive, cbind is a special .Internal and grep is a builtin .Internal.

⁵ There is currently one other difference: when profiling builtin functions are counted as function calls but specials are not.

Many of the internal functions are internal generics, which for specials means that they do not evaluate their arguments on call, but the C code starts with a call to DispatchOrEval. The latter evaluates the first argument, and looks for a method based on its class. (If S4 dispatch is on, S4 methods are looked for first, even for S3 classes.) If it finds a method, it dispatches to that method with a call based on promises to evaluate the remaining arguments. If no method is found, the remaining arguments are evaluated before return to the internal generic.

The other way that internal functions can be generic is to be group generic. Most such functions are builtins (so immediately evaluate all their arguments), and all contain a call to the C function DispatchGeneric. There are some peculiarities over the number of arguments for the "Math" group generic, with some members allowing only one argument, some having two (with a default for the second) and trunc allows one or more but the default method only accepts one.

1.5.1 Missingness

Actual arguments to (non-internal) R functions can be fewer than are required to match the formal arguments of the function. Having unmatched formal arguments will not matter if the argument is never used (by lazy evaluation), but when the argument is evaluated, either its default value is evaluated (within the evaluation environment of the function) or an error is thrown with a message along the lines of

argument "foobar" is missing, with no default

Internally missingness is handled by two mechanisms. The object R_MissingArg is used to indicate that a formal argument has no (default) value. When matching the actual arguments to the formal arguments, a new argument list is constructed from the formals all of whose values are R_MissingArg with the first MISSING bit set. Then whenever a formal argument is matched to an actual argument, the corresponding member of the new argument list has its value set to that of the matched actual argument, and if that is not R_MissingArg the missing bit is unset.

This new argument list is used to form the evaluation frame for the function, and if named arguments are subsequently given a new value (before they are evaluated) the missing bit is cleared.

Missingness of arguments can be interrogated via the missing() function. An argument is clearly missing if its missing bit is set or if the value is R_MissingArg. However, missingness can be passed on from function to function, for using a formal argument as an actual argument in a function call does not count as evaluation. So missing() has to examine the value (a promise) of a non-yet-evaluated formal argument to see if it might be missing, which might involve investigating a promise and so on ….

Special primitives also need to handle missing arguments, and in some case (e.g. log) that is why they are special and not builtin. This is usually done by testing if an argument’s value is R_MissingArg.

1.5.2 Dot-dot-dot arguments

Dot-dot-dot arguments are convenient when writing functions, but complicate the internal code for argument evaluation.

The formals of a function with a ... argument represent that as a single argument like any other argument, with tag the symbol R_DotsSymbol. When the actual arguments are matched to the formals, the value of the ... argument is of SEXPTYPE DOTSXP, a pairlist of promises (as used for matched arguments) but distinguished by the SEXPTYPE.

Recall that the evaluation frame for a function initially contains the name=value pairs from the matched call, and hence this will be true for ... as well. The value of ... is a (special) pairlist whose elements are referred to by the special symbols ..1, ..2, … which have the DDVAL bit set: when one of these is encountered it is looked up (via ddfindVar) in the value of the ... symbol in the evaluation frame.

Values of arguments matched to a ... argument can be missing.

Special primitives may need to handle ... arguments: see for example the internal code of switch in file src/main/builtin.c.

1.6 Autoprinting

Whether the returned value of a top-level R expression is printed is controlled by the global boolean variable R_Visible. This is set (to true or false) on entry to all primitive and internal functions based on the eval column of the table in file src/main/names.c: the appropriate setting can be extracted by the macro PRIMPRINT.

The R primitive function invisible makes use of this mechanism: it just sets R_Visible = FALSE before entry and returns its argument.

For most functions the intention will be that the setting of R_Visible when they are entered is the setting used when they return, but there need to be exceptions. The R functions identify, options, system and writeBin determine whether the result should be visible from the arguments or user action. Other functions themselves dispatch functions which may change the visibility flag: examples⁶ are .Internal, do.call, eval, withVisible, if, NextMethod, Recall, recordGraphics, standardGeneric, switch and UseMethod.

⁶ the other current example is left brace, which is implemented as a primitive.

‘Special’ primitive and internal functions evaluate their arguments internally after R_Visible has been set, and evaluation of the arguments (e.g. an assignment as in PR#9263) can change the value of the flag.

The R_Visible flag can also get altered during the evaluation of a function, with comments in the code about warning, writeChar and graphics functions calling GText (PR#7397). (Since the C-level function eval sets R_Visible, this could apply to any function calling it. Since it is called when evaluating promises, even object lookup can change R_Visible.) Internal and primitive functions force the documented setting of R_Visible on return, unless the C code is allowed to change it (the exceptions above are indicated by PRIMPRINT having value 2).

The actual autoprinting is done by PrintValueEnv in file print.c. If the object to be printed has the S4 bit set and S4 methods dispatch is on, show is called to print the object. Otherwise, if the object bit is set (so the object has a "class" attribute), print is called to dispatch methods: for objects without a class the internal code of print.default is called.

1.7 The write barrier and the garbage collector

R has long had a generational garbage collector, and bit gcgen in the sxpinfo header is used in the implementation of this. This is used in conjunction with the mark bit to identify two previous generations.

There are three levels of collections. Level 0 collects only the youngest generation, level 1 collects the two youngest generations and level 2 collects all generations. After 20 level-0 collections the next collection is at level 1, and after 5 level-1 collections at level 2. Further, if a level-n collection fails to provide 20% free space (for each of nodes and the vector heap), the next collection will be at level n+1. (The R-level function gc() performs a level-2 collection.)

A generational collector needs to efficiently ‘age’ the objects, especially list-like objects (including STRSXPs). This is done by ensuring that the elements of a list are regarded as at least as old as the list when they are assigned. This is handled by the functions SET_VECTOR_ELT and SET_STRING_ELT, which is why they are functions and not macros. Ensuring the integrity of such operations is termed the write barrier and is done by making the SEXP opaque and only providing access via functions (which cannot be used as lvalues in assignments in C).

All code in R extensions is behind the write barrier. R extensions cannot directly access the internals of the SEXPRECs. Base code can access internals if USE_RINTERNALS is defined. This is normally defined in Defn.h when R is compiled. To enable a check on the way that the access is used, R can be compiled with flag --enable-strict-barrier which ensures that header Defn.h does not define USE_RINTERNALS and hence that SEXP is opaque in most of R itself. (There are some necessary exceptions: foremost in file memory.c where the accessor functions are defined and also in file size.c which needs access to the sizes of the internal structures.)

For background papers see https://homepage.stat.uiowa.edu/~luke/R/barrier.html and https://homepage.stat.uiowa.edu/~luke/R/gengcnotes.html.

1.8 Serialization Formats

Serialized versions of R objects are used by load/save and also at a slightly lower level by saveRDS/readRDS (and their earlier ‘internal’ dot-name versions) and serialize/unserialize. These differ in what they serialize to (a file, a connection, a raw vector) and whether they are intended to serialize a single object or a collection of objects (typically the workspace). save writes a header at the beginning of the file (a single LF-terminated line) which the lower-level versions do not.

save and saveRDS allow various forms of compression, and gzip compression is the default (except for ASCII saves). Compression is applied to the whole file stream, including the headers, so serialized files can be uncompressed or re-compressed by external programs. Both load and readRDS can read gzip, bzip2 and xz forms of compression when reading from a file, and gzip compression when reading from a connection.

R has used the same serialization format called ‘version 2’ from R 1.4.0 in December 2001 until R 3.5.3 in March 2019. It has been expanded in back-compatible ways since its inception, for example to support additional SEXPTYPEs. Earlier formats are still supported via load and save but such formats are not described here. The current default serialization format is called ‘version 3’, and has been introduced in R 3.5.0.

save works by writing a single-line header (typically RDX2\n for a binary save: the only other current value is RDA2\n for save(files=TRUE)), then creating a tagged pairlist of the objects to be saved and serializing that single object. load reads the header line, unserializes a single object (a pairlist or a vector list) and assigns the elements of the object in the specified environment. The header line serves two purposes in R: it identifies the serialization format so load can switch to the appropriate reader code, and the newline \n allows the detection of files which have been subjected to a non-binary transfer which re-mapped line endings. It can also be thought of as a ‘magic number’ in the sense used by the file program (although R save files are not yet by default known to that program).

Serialization in R needs to take into account that objects may contain references to environments, which then have enclosing environments and so on. (Environments recognized as package or name space environments are saved by name.) There are ‘reference objects’ which are not duplicated on copy and should remain shared on unserialization. These are weak references, external pointers and environments other than those associated with packages, namespaces and the global environment. These are handled via a hash table, and references after the first are written out as a reference marker indexed by the table entry.

Version-2 serialization first writes a header indicating the format (normally X\n for an XDR format binary save, but A\n, ASCII, and B\n, native word-order binary, can also occur) and then three integers giving the version of the format and two R versions (packed by the R_Version macro from Rversion.h). (Unserialization interprets the two versions as the version of R which wrote the file followed by the minimal version of R needed to read the format.) Serialization then writes out the object recursively using function WriteItem in file src/main/serialize.c.

Some objects are written as if they were SEXPTYPEs: such pseudo-SEXPTYPEs cover R_NilValue, R_EmptyEnv, R_BaseEnv, R_GlobalEnv, R_UnboundValue, R_MissingArg and R_BaseNamespace.

For all SEXPTYPEs except NILSXP, SYMSXP and ENVSXP serialization starts with an integer with the SEXPTYPE in bits 0:7⁷ followed by the object bit, two bits indicating if there are any attributes and if there is a tag (for the pairlist types), an unused bit and then the gp field⁸ in bits 12:27. Pairlist-like objects write their attributes (if any), tag (if any), CAR and then CDR (using tail recursion): other objects write their attributes after themselves. Atomic vector objects write their length followed by the data: generic vector-list objects write their length followed by a call to WriteItem for each element. The code for CHARSXPs special-cases NA_STRING and writes it as length -1 with no data. Lengths no more than 2^31 - 1 are written in that way and larger lengths (which only occur on 64-bit systems) as -1 followed by the upper and lower 32-bits as integers (regarded as unsigned).

⁷ only bits 0:4 are currently used for SEXPTYPEs but values 241:255 are used for pseudo-SEXPTYPEs.

⁸ Currently the only relevant bits are 0:1, 4, 14:15.

Environments are treated in several ways: as we have seen, some are written as specific pseudo-SEXPTYPEs. Package and namespace environments are written with pseudo-SEXPTYPEs followed by the name. ‘Normal’ environments are written out as ENVSXPs with an integer indicating if the environment is locked followed by the enclosure, frame, ‘tag’ (the hash table) and attributes.

In the ‘XDR’ format integers and doubles are written in big-endian order: however the format is not fully XDR (as defined in RFC 1832) as byte quantities (such as the contents of CHARSXP and RAWSXP types) are written as-is and not padded to a multiple of four bytes.

The ‘ASCII’ format writes 7-bit characters. Integers are formatted with %d (except that NA_integer_ is written as NA), doubles formatted with %.16g (plus NA, Inf and -Inf) and bytes with %02x. Strings are written using standard escapes (e.g. \t and \013) for non-printing and non-ASCII bytes.

Version-3 serialization extends version-2 by support for custom serialization of ALTREP framework objects. It also stores the current native encoding at serialization time, so that unflagged strings can be converted if unserialized in R running under different native encoding.

1.9 Encodings for `CHARSXP`s

Character data in R are stored in the SEXPTYPE CHARSXP.

There is support for encodings other than that of the current locale, in particular UTF-8 and the multi-byte encodings used on Windows for CJK languages. A limited means to indicate the encoding of a CHARSXP is via two of the ‘general purpose’ bits which are used to declare the encoding to be either Latin-1 or UTF-8. (Note that it is possible for a character vector to contain elements in different encodings.) Both printing and plotting notice the declaration and convert the string to the current locale (possibly using <xx> to display in hexadecimal bytes that are not valid in the current locale). Many (but not all) of the character manipulation functions will either preserve the declaration or re-encode the character string.

Strings that refer to the OS such as file names need to be passed through a wide-character interface on some OSes (e.g. Windows).

When are character strings declared to be of known encoding? One way is to do so directly via Encoding. The parser declares the encoding if this is known, either via the encoding argument to parse or from the locale within which parsing is being done at the R command line. (Other ways are recorded on the help page for Encoding.)

It is not necessary to declare the encoding of ASCII strings as they will work in any locale. ASCII strings should never have a marked encoding, as any encoding will be ignored when entering such strings into the CHARSXP cache.

The rationale behind considering only UTF-8 and Latin-1 was that most systems are capable of producing UTF-8 strings and this is the nearest we have to a universal format. For those that do not (for example those lacking a powerful enough iconv), it is likely that they work in Latin-1, the old R assumption. Then the parser can return a UTF-8-encoded string if it encounters a \uxxxx escape for a Unicode point that cannot be represented in the current charset. (This needs MBCS support, and in the past was only enabled⁹ on Windows.) Now this is enabled for all platforms, and a \uxxxx or \Uxxxxxxxx escape ensures that the parsed string will be marked as UTF-8.

⁹ See define USE_UTF8_IF_POSSIBLE in old versions of file src/main/gram.c.

Most of the character manipulation functions now preserve UTF-8 encodings: there are some notes as to which at the top of file src/main/character.c and in file src/library/base/man/Encoding.Rd.

Graphics devices are offered the possibility of handing UTF-8-encoded strings without re-encoding to the native character set, by setting hasTextUTF8 to be TRUE and supplying functions textUTF8 and strWidthUTF8 that expect UTF-8-encoded inputs. Normally the symbol font is encoded in Adobe Symbol encoding, but that can be re-encoded to UTF-8 by setting wantSymbolUTF8 to TRUE. The Windows’ port of cairographics has a rather peculiar assumption: it wants the symbol font to be encoded in UTF-8 as if it were encoded in Latin-1 rather than Adobe Symbol: this is selected by wantSymbolUTF8 = NA_LOGICAL.

Windows with MSVCRT as the C runtime has no UTF-8 locales, but rather expects to work with UCS-2¹⁰ strings. R (being written in standard C) would not work internally with UCS-2 without extensive changes. The Rgui console¹¹ uses UCS-2 internally, but communicates with the R engine in the native encoding. To allow UTF-8 strings to be printed in UTF-8 in Rgui.exe, an escape convention is used (see header file rgui_UTF8.h) by cat, print and autoprinting.

¹⁰ or UTF-16 if support for surrogates is enabled in the OS, which it used not to be when encoding support was added to R.

¹¹ but not the GraphApp toolkit.

‘Unicode’ (UCS-2LE) files are common in the Windows world, and readLines and scan will read them into UTF-8 strings on Windows if the encoding is declared explicitly on an unopened connection passed to those functions.

Windows have multiple notions of the current locale encoding, one is in the C runtime (C library) and another is the active code page (system locale). The active code page is used when calling non-UTF-16 variants of Windows API functions (earlier referred to as ANSI calls), either directly or indirectly via POSIX wrappers inside MinGW-w64, from R, R packages and libraries they link to. While R has handled many cases by calling directly the UTF-16 variants of the Windows API, it still may sometimes use the non-UTF-16 ones, and external libraries also primarily developed for POSIX systems typically do that. Therefore, for R to reliably work with (non-ASCII) strings on Windows, both the C locale encoding and the active code page on Windows must be the same, and by default they are.

The Windows UCRT C runtime supports UTF-8 locales. Historically, the active code page was a system-wide setting, changing which required a reboot, and UTF-8 was not supported. Later a "BETA: Use Unicode UTF-8 for worldwide language support" feature has been added to set the active code page to UTF-8, but this still required a reboot and impacted all applications, many of which would not work correctly with that unexpected setting, so it could not be used in practice.

Windows since Windows 10 (version 1903), Windows Server 2022 (LTSC), and Windows Server 1903 (semi-annual channel) allow setting the active code page to UTF-8 in the application manifest. This changes the active code page only for the given application and does so together with changing also the UCRT C locale to UTF-8. R 4.2 for Windows uses this feature to get UTF-8 as the native encoding on Windows. To make that possible, R had to switch to UCRT, which in turn required creation of Rtools42.

Older versions of Windows still rely on the previous encoding support where the native encoding cannot be UTF-8. R 4.2 requires UCRT to work, but UCRT can be installed on Windows since Vista SP2 and Windows Server 2008 SP2. It is shipped with Windows since Windows 10 and Windows Server 2016.

1.10 The `CHARSXP` cache

There is a global cache for CHARSXPs created by mkChar — the cache ensures that most CHARSXPs with the same contents share storage (‘contents’ including any declared encoding). Not all CHARSXPs are part of the cache – notably NA_STRING is not. CHARSXPs reloaded from the save formats of R prior to 0.99.0 are not cached (since the code used is frozen and very few examples still exist).

The cache records the encoding of the string as well as the bytes: all requests to create a CHARSXP should be via a call to mkCharLenCE. Any encoding given in mkCharLenCE call will be ignored if the string’s bytes are all ASCII characters.

1.11 Warnings and errors

Each of warning and stop have two C-level equivalents, warning, warningcall, error and errorcall. The relationship between the pairs is similar: warning tries to fathom out a suitable call, and then calls warningcall with that call as the first argument if it succeeds, and with call = R_NilValue if it does not. When warningcall is called, it includes the deparsed call in its printout unless call = R_NilValue.

warning and error look at the context stack. If the topmost context is not of type CTXT_BUILTIN, it is used to provide the call, otherwise the next context provides the call. This means that when these functions are called from a primitive or .Internal, the imputed call will not be to primitive/.Internal but to the function calling the primitive/.Internal . This is exactly what one wants for a .Internal, as this will give the call to the closure wrapper. (Further, for a .Internal, the call is the argument to .Internal, and so may not correspond to any R function.) However, it is unlikely to be what is needed for a primitive.

The upshot is that warningcall and errorcall should normally be used for code called from a primitive, and warning and error should be used for code called from a .Internal (and necessarily from .Call, .C and so on, where the call is not passed down). However, there are two complications. One is that code might be called from either a primitive or a .Internal, in which case probably warningcall is more appropriate. The other involves replacement functions, where the call was once of the form

> length(x) <- y ~ x
Error in "length<-"(`*tmp*`, value = y ~ x) : invalid value

which is unpalatable to the end user. For replacement functions there will be a suitable context at the top of the stack, so warning should be used. (The results for .Internal replacement functions such as substr<- are not ideal.)

1.12 S4 objects

[This section is currently a preliminary draft and should not be taken as definitive. The description assumes that R_NO_METHODS_TABLES has not been set.]

1.12.1 Representation of S4 objects

S4 objects can be of any SEXPTYPE. They are either an object of a simple type (such as an atomic vector or function) with S4 class information or of type OBJSXP. In all cases, the ‘S4 bit’ (bit 4 of the ‘general purpose’ field) is set, and can be tested by the macro/function IS_S4_OBJECT.

S4 objects are created via new()¹² and thence via the C function R_do_new_object. This duplicates the prototype of the class, adds a class attribute and sets the S4 bit. All S4 class attributes should be character vectors of length one with an attribute giving (as a character string) the name of the package (or .GlobalEnv) containing the class definition. Since S4 objects have a class attribute, the OBJECT bit is set.

¹² This can also create non-S4 objects, as in new("integer").

It is currently unclear what should happen if the class attribute is removed from an S4 object, or if this should be allowed.

1.12.2 S4 classes

S4 classes are stored as R objects in the environment in which they are created, with names .__C__classname: as such they are not listed by default by ls.

The objects are S4 objects of class "classRepresentation" which is defined in the methods package.

Since these are just objects, they are subject to the normal scoping rules and can be imported and exported from namespaces like other objects. The directives importClassesFrom and exportClasses are merely convenient ways to refer to class objects without needing to know their internal ‘metaname’ (although exportClasses does a little sanity checking via isClass).

1.12.3 S4 methods

Details of the methods are stored in environments (typically hidden in the respective namespace) with a non-syntactic name of the form .__T__generic:package containing objects of class MethodDefinition for all methods defined in the current environment for the named generic derived from a specific package (which might be .GlobalEnv). This is sometimes referred to as a ‘methods table’.

For example,

 length(nM <- asNamespace("Matrix") )                    # 941 for Matrix 1.2-6
 length(meth <- grep("^[.]__T__", names(nM), value=TRUE))# 107 generics with methods
 length(meth.Ops <- nM$`.__T__Ops:base`) # 71 methods for the 'Ops' (group)generic
 head(sort(names(meth.Ops))) ## "abIndex#abIndex" ... "ANY#ddiMatrix" "ANY#ldiMatrix" "ANY#Matrix"

During an R session there is an environment associated with each non-primitive generic containing objects .AllMTable, .Generic, .Methods, .MTable, .SigArgs and .SigLength. .MTable and AllMTable are merged methods tables containing all the methods defined directly and via inheritance respectively. .Methods is a merged methods list.

Exporting methods from a namespace is more complicated than exporting a class. Note first that you do not export a method, but rather the directive exportMethods will export all the methods defined in the namespace for a specified generic: the code also adds to the list of generics any that are exported directly. For generics which are listed via exportMethods or exported themselves, the corresponding environment is exported and so will appear (as hidden object) in the package environment.

Methods for primitives which are internally S4 generic (see below) are always exported, whether mentioned in the NAMESPACE file or not.

Methods can be imported either via the directive importMethodsFrom or via importing a namespace by import. Also, if a generic is imported via importFrom, its methods are also imported. In all cases the generic will be imported if it is in the namespace, so importMethodsFrom is most appropriate for methods defined on generics in other packages. Since methods for a generic could be imported from several different packages, the methods tables are merged.

When a package is attached methods:::cacheMetaData is called to update the internal tables: only the visible methods will be cached.

1.12.4 Mechanics of S4 dispatch

This subsection does not discuss how S4 methods are chosen: see https://developer.r-project.org/howMethodsWork.pdf.

For all but primitive functions, setting a method on an existing function that is not itself S4 generic creates a new object in the current environment which is a call to standardGeneric with the old definition as the default method. Such S4 generics can also be created via a call to setGeneric¹³ and are standard closures in the R language, with environment the environment within which they are created. With the advent of namespaces this is somewhat problematic: if myfn was previously in a package with a name space there will be two functions called myfn on the search paths, and which will be called depends on which search path is in use. This is starkest for functions in the base namespace, where the original will be found ahead of the newly created function from any other package.

¹³ although this is not recommended as it is less future-proof.

Primitive functions are treated quite differently, for efficiency reasons: this results in different semantics. setGeneric is disallowed for primitive functions. The methods namespace contains a list .BasicFunsList named by primitive functions: the entries are either FALSE or a standard S4 generic showing the effective definition. When setMethod (or setReplaceMethod) is called, it either fails (if the list entry is FALSE) or a method is set on the effective generic given in the list.

Actual dispatch of S4 methods for almost all primitives piggy-backs on the S3 dispatch mechanism, so S4 methods can only be dispatched for primitives which are internally S3 generic. When a primitive that is internally S3 generic is called with a first argument which is an S4 object and S4 dispatch is on (that is, the methods namespace is loaded), DispatchOrEval calls R_possible_dispatch (defined in file src/main/objects.c). (Members of the S3 group generics, which includes all the generic operators, are treated slightly differently: the first two arguments are checked and DispatchGroup is called.) R_possible_dispatch first checks an internal table to see if any S4 methods are set for that generic (and S4 dispatch is currently enabled for that generic), and if so proceeds to S4 dispatch using methods stored in another internal table. All primitives are in the base namespace, and this mechanism means that S4 methods can be set for (some) primitives and will always be used, in contrast to setting methods on non-primitives.

The exception is %*%, which is S4 generic but not S3 generic as its C code contains a direct call to R_possible_dispatch.

The primitive as.double is special, as as.numeric and as.real are copies of it. The methods package code partly refers to generics by name and partly by function, and maps as.double and as.real to as.numeric (since that is the name used by packages exporting methods for it).

Some elements of the language are implemented as primitives, for example }. This includes the subset and subassignment ‘functions’ and they are S4 generic, again piggybacking on S3 dispatch.

.BasicFunsList is generated when methods is installed, by computing all primitives, initially disallowing methods on all and then setting generics for members of .GenericArgsEnv, the S4 group generics and a short exceptions list in file BasicFunsList.R: this currently contains the subsetting and subassignment operators and an override for c.

1.13 Memory allocators

R’s memory allocation is almost all done via routines in file src/main/memory.c.

The rest of R should where possible make use of the allocators made available by file src/main/memory.c, which are also the methods recommended in Memory allocation in Writing R Extensions for use in R packages, namely the use of R_alloc, R_Calloc, R_Realloc and R_Free. Memory allocated by R_alloc is freed by the garbage collector once the ‘watermark’ has been reset by calling vmaxset. This is done automatically by the wrapper code calling primitives and .Internal functions (and also by the wrapper code to .Call and .External), but vmaxget and vmaxset can be used to reset the watermark from within internal code if the memory is only required for a short time.

All of the methods of memory allocation mentioned so far are relatively expensive. All R platforms support alloca, and in almost all cases¹⁴ this is managed by the compiler, allocates memory on the C stack and is very efficient.

¹⁴ but apparently not on Windows.

There are two disadvantages in using alloca. First, it is fragile and care is needed to avoid writing (or even reading) outside the bounds of the allocation block returned. Second, it increases the danger of overflowing the C stack. It is suggested that it is only used for smallish allocations (up to tens of thousands of bytes), and that

    R_CheckStack();

is called immediately after the allocation (as R’s stack checking mechanism will warn far enough from the stack limit to allow for modest use of alloca). (do_makeunique in file src/main/unique.c provides an example of both points.)

There is an alternative check,

    R_CheckStack2(size_t extra);

to be called immediately before trying an allocation of extra bytes.

An alternative strategy has been used for various functions which require intermediate blocks of storage of varying but usually small size, and this has been consolidated into the routines in the header file src/main/RBufferUtils.h. This uses a structure which contains a buffer, the current size and the default size. A call to

    R_AllocStringBuffer(size_t blen, R_StringBuffer *buf);

sets buf->data to a memory area of at least blen+1 bytes. At least the default size is used, which means that for small allocations the same buffer can be reused. A call to R_FreeStringBufferL releases memory if more than the default has been allocated whereas a call to R_FreeStringBuffer frees any memory allocated.

The R_StringBuffer structure needs to be initialized, for example by

static R_StringBuffer ex_buff = {NULL, 0, MAXELTSIZE};

which uses a default size of MAXELTSIZE = 8192 bytes. Most current uses have a static R_StringBuffer structure, which allows the (default-sized) buffer to be shared between calls to e.g. grep and even between functions: this will need to be changed if R ever allows concurrent evaluation threads. So the idiom is

static R_StringBuffer ex_buff = {NULL, 0, MAXELTSIZE};
...
    char *buf;
    for(i = 0; i < n; i++) {
        compute len
        buf = R_AllocStringBuffer(len, &ex_buff);
        use buf
    }
    /*  free allocation if larger than the default, but leave
        default allocated for future use */
   R_FreeStringBufferL(&ex_buff);

1.13.1 Internals of `R_alloc`

The memory used by R_alloc is allocated as R vectors, of type RAWSXP. Thus the allocation is in units of 8 bytes, and is rounded up. A request for zero bytes currently returns NULL (but this should not be relied on). For historical reasons, in all other cases 1 byte is added before rounding up so the allocation is always 1–8 bytes more than was asked for: again this should not be relied on.

The vectors allocated are protected via the setting of R_VStack, as the garbage collector marks everything that can be reached from that location. When a vector is R_allocated, its ATTRIB pointer is set to the current R_VStack, and R_VStack is set to the latest allocation. Thus R_VStack is a single-linked chain of the vectors currently allocated via R_alloc. Function vmaxset resets the location R_VStack, and should be to a value that has previously be obtained via vmaxget: allocations after the value was obtained will no longer be protected and hence available for garbage collection.

1.14 Internal use of global and base environments

This section notes known use by the system of these environments: the intention is to minimize or eliminate such uses.

1.14.1 Base environment

The graphics devices system maintains two variables .Device and .Devices in the base environment: both are always set. The variable .Devices gives a list of character vectors of the names of open devices, and .Device is the element corresponding to the currently active device. The null device will always be open.

There appears to be a variable .Options, a pairlist giving the current options settings. But in fact this is just a symbol with a value assigned, and so shows up as a base variable.

Similarly, the evaluator creates a symbol .Last.value which appears as a variable in the base environment.

Errors can give rise to objects .Traceback and last.warning in the base environment.

1.14.2 Global environment

The seed for the random number generator is stored in object .Random.seed in the global environment.

Some error handlers may give rise to objects in the global environment: for example dump.frames by default produces last.dump.

The windows() device makes use of a variable .SavedPlots to store display lists of saved plots for later display. This is regarded as a variable created by the user.

1.15 Modules

R makes use of a number of shared objects/DLLs stored in the modules directory. These are parts of the code which have been chosen to be loaded ‘on demand’ rather than linked as dynamic libraries or incorporated into the main executable/dynamic library.

For the remaining modules the motivation has been the amount of (often optional) code they will bring in via libraries to which they are linked.

internet: The internal HTTP and FTP clients and socket support, which link to system-specific support libraries. This may load libcurl and on Windows will load wininet.dll and ws2_32.dll.
lapack: The code which makes use of the LAPACK library, and is linked to libRlapack or an external LAPACK library.
X11: (Unix-alikes only.) The X11(), jpeg(), png() and tiff() devices. These are optional, and links to some or all of the X11, pango, cairo, jpeg, libpng and libtiff libraries.

1.16 Visibility

1.16.1 Hiding C entry points

We make use of the visibility mechanisms discussed in Controlling visibility in Writing R Extensions. C entry points not needed outside the main R executable/dynamic library (and in particular in no package nor module) should be prefixed by attribute_hidden. Minimizing the visibility of symbols in the R dynamic library will speed up linking to it (which packages will do) and reduce the possibility of linking to the wrong entry points of the same name. In addition, on some platforms reducing the number of entry points allows more efficient versions of PIC to be used: somewhat over half the entry points are hidden. A convenient way to hide variables (as distinct from functions) is to declare them extern0 in header file Defn.h.

The visibility mechanism used is only available with some compilers and platforms, and in particular not on Windows, where an alternative mechanism is used. Entry points will not be made available in R.dll if they are listed in the file src/gnuwin32/Rdll.hide. Entries in that file start with a space and must be strictly in alphabetic order in the C locale (use sort on the file to ensure this if you change it). It is possible to hide Fortran as well as C entry points via this file: the former are lower-cased and have an underline as suffix, and the suffixed name should be included in the file. Some entry points exist only on Windows or need to be visible only on Windows, and some notes on these are provided in file src/gnuwin32/Maintainers.notes.

Because of the advantages of reducing the number of visible entry points, they should be declared attribute_hidden where possible. Note that this only has an effect on a shared-R-library build, and so care is needed not to hide entry points that are legitimately used by packages. So it is best if the decision on visibility is made when a new entry point is created, including the decision if it should be included in header file Rinternals.h. A list of the visible entry points on shared-R-library build on a reasonably standard Unix-alike can be made by something like

nm -g libR.so | grep ' [BCDT] ' | cut -b20-

1.16.2 Variables in Windows DLLs

Windows is unique in that it conventionally treats importing variables differently from functions: variables that are imported from a DLL need to be specified by a prefix (often _imp_) when being linked to (‘imported’) but not when being linked from (‘exported’). The details depend on the compiler system, and have changed for MinGW during the lifetime of that port. They are in the main hidden behind some macros defined in header file R_ext/libextern.h.

A (non-function) variable in the main R sources that needs to be referred to outside R.dll (in a package, module or another DLL such as Rgraphapp.dll) should be declared with prefix LibExtern. The main use is in Rinternals.h, but it needs to be considered for any public header and also Defn.h.

It would nowadays be possible to make use of the ‘auto-import’ feature of the MinGW port of ld to fix up imports from DLLs (and if R is built for the Cygwin platform this is what happens). However, this was not possible when the MinGW build of R was first constructed in ca 1998, allows less control of visibility and would not work for other Windows compiler suites.

It is only possible to check if this has been handled correctly by compiling the R sources on Windows.

1.17 Lazy loading

Lazy loading is always used for code in packages but is optional (selected by the package maintainer) for datasets in packages. When a package/namespace which uses it is loaded, the package/namespace environment is populated with promises for all the named objects: when these promises are evaluated they load the actual code from a database.

There are separate databases for code and data, stored in the R and data subdirectories. The database consists of two files, name.rdb and name.rdx. The .rdb file is a concatenation of serialized objects, and the .rdx file contains an index. The objects are stored in (usually) a gzip-compressed format with a 4-byte header giving the uncompressed serialized length (in XDR, that is big-endian, byte order) and read by a call to the primitive lazyLoadDBfetch. (Note that this makes lazy-loading unsuitable for really large objects: the unserialized length of an R object can exceed 4GB.)

The index or ‘map’ file name.rdx is a compressed serialized R object to be read by readRDS. It is a list with three elements variables, references and compressed. The first two are named lists of integer vectors of length 2 giving the offset and length of the serialized object in the name.rdb file. Element variables has an entry for each named object: references serializes a temporary environment used when named environments are added to the database. compressed is a logical indicating if the serialized objects were compressed: compression is always used nowadays. We later added the values compressed = 2 and 3 for bzip2 and xz compression (with the possibility of future expansion to other methods): these formats add a fifth byte to the header for the type of compression, and store serialized objects uncompressed if compression expands them.

Source references are treated specially for performance reasons: bindings lines and parseData from srcfile environments are loaded lazily. This uses a mechanism that allows loading selected bindings from an environment lazily. The key for such environment is a list with two elements: eagerKey gives the length-two integer key for the bindings loaded eagerly and lazyKeys gives a vector of length-two integer keys, one for each lazily loaded binding.

The loader for a lazy-load database of code or data is function lazyLoad in the base package, but note that there is a separate copy to load base itself in file R_HOME/base/R/base.

Lazy-load databases are created by the code in src/library/tools/R/makeLazyLoad.R: the main tool is the unexported function makeLazyLoadDB and the insertion of database entries is done by calls to .Call("R_lazyLoadDBinsertValue", ...).

Lazy-load databases of less than 10MB are cached in memory at first use: this was found necessary when using file systems with high latency (removable devices and network-mounted file systems on Windows).

Lazy-load databases are loaded into the exports for a package, but not into the namespace environment itself. Thus they are visible when the package is attached, and also via the :: operator. This was a deliberate design decision, as packages mostly make datasets available for use by the end user (or other packages), and they should not be found preferentially from functions in the package, surprising users who expected the normal search path to be used. (There is an alternative mechanism, sysdata.rda, for ‘system datasets’ that are intended primarily to be used within the package.)

The same database mechanism is used to store parsed Rd files. One or all of the parsed objects is fetched by a call to tools:::fetchRdDB.

Footnotes