Merge of branch 'hex-float'.
- Add support in byterun/floats.c for conversions between floats and strings in hex notation. We cannot rely on the C standard library here because Microsoft consistently fails at supporting hex notation as standardized in C99. Instead, the conversions are implemented from scratch.
- Add support in the lexer so that hex float literals are recognized in OCaml sources.
- Add support in formats. The ISO C99 format letters for hex floats are %a and %A, but %a is already taken. I chose %h and %H, which are rejected today as bad formats (hence no backward incompatibility) and don't mean anything in C either (h is a modifier, not a format letter).
- Add support in printf. All the trimmings are there in the implementation of %h and %H, including sign modifier and fixed precision.
- Benoit Vaugon contributed support in scanf.
Resolved conflicts:
boot/ocamlc
boot/ocamldep
boot/ocamllex
parsing/lexer.mll
The fix consists in representing float literals by their bit patterns (int64) in the Mach and Linear intermediate languages.
A regression test was added to the test suite.
As suggested in the discussion of GPR#272:
- Do not go through fpclassify() (speedup: 2 to 3)
- Add 64-bit variant of the code (additional speedup: 10%-20%)
Unix.sleepf provides sleep with sub-second resolution.
Unix.sleep is implemented on top of Unix.sleepf.
If a handled signal causes the sleep to return early with an EINTR
error, catch it and restart the sleep for the remaining time.
Implementation notes:
- Based on c-cube's GPR#227 code, but many Unix functions were missing.
- For Unix.bind and Unix.connect to a PF_UNIX address, tolerate
file names whose first character is '\000': in Linux, these have
a meaning as "abstract socket addresses", and in other operating
systems, the resulting empty path name causes a EINVAL error
(tested under MacOS X).
- Very lightly tested.
- win32unix remains to be fixed.
Add the following syntactic sugars, in both expressions and patterns:
- { f : typ = x } to mean { f = (x : typ) }
- { f : typ } to mean { f = (f : typ) }
In expressions, the type annotation can also be a coercion. For
instance: { f :> typ } to mean { f = (f :> typ) }.
Patch by Valentin Gatien-Baron
(Leo White)
Add support for simple functor types of the form:
S -> T
equivalent to:
functor (_ : S) -> T
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16546 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
(Jérémie Dimino)
- use inline-records to avoid using Obj.magic while keeping
indirections to the minumum
- change the representation from a cyclic list to a simply-linked
list. The drawback is one more word per queue, but it makes the
implementation clearer
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16545 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
to warn when the argument is matches against a constant pattern.
This attributes is applied on predefined exception constructors
which take an purely informational (with no stability guarantee)
message.
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16502 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
compiling the run-time system and C stub code.
Also: select "gnu99" mode (= ISO C99 + GNU extensions).
Also: gcc/clang warnings are errors when in development mode.
(Turned off for release builds.)
Also: clean up and modernize the autoconf tests in config/auto-aux
(Merge of branch 'cc-optim'.)
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16379 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
There is currently a GC bug in the bytecode debug-info handling, due
to the fact that
void read_main_debug_info(struct debug_info *di)
is passed a internal pointer in the middle of a custom block inside
the OCaml heap. I could only observe the bug when such custom blocks
are allocated on the minor heap -- which does not happen with the
current implementation, but becomes possible after GPR#92 for example
(which let custom blocks with finalizer be allocated in the
minor heap).
This commit fixes this issue by moving debug_info chunks from the
OCaml heap to the C land, stored in a dynamic table. They are
allocated when caml_add_debug_info is called, and removed when
caml_remove_debug_info is called.
(Another approach would be to keep the debug_info inside the OCaml
heap, but make sure that there are no dangling internal pointers. See
GPR#228 for an attack of this by Mark Shinwell.)
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16356 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
* PR#6816: reject identifiers following numeric literals without spacing, e.g.
let foo = 1 in (fun a b -> a,b) 23foo
(Hugo Heuzard)
(Thanks to Daniel Bünzli for reminding me.)
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16317 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
(Hongbo Zhang)
when printing recursive module the old version would print no space
before 'and':
A : ... = struct
...
endand B : ... = ...
this tiny PR fixes such problem
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16307 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
(Pierre Chambart)
It now reallocates and reinitialises the whole table only when it is
too small. This avoids quadratic behavior when loading a lot of module
with dynlink.
This was problematic on frama-c when inlining increase the code
size. The frame table initialisation took ~0.5 second. This is quite
noticeable on real examples where the whole frama-c analysis is ~1.5s
long.
Also allows to unregister a frametable.
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16260 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
(Clark Gaebel)
These functions are called every tick of the Async scheduler, and
are the only remaining calls to [caml_c_call] every cycle. It would
be nice to remove them, especially since these functions don't
allocate.
git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@16259 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02
Damien Doligez
Xavier Leroy
Thomas Refis
Jacques Garrigue
Xavier Leroy
Gabriel Scherer
François Bobot
Leo White
Jeremie Dimino
Mark Shinwell
Leo White
alainfrisch
Jérémie Dimino
Jeremy Yallop
Jacques Garrigue
Damien Doligez