| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I/O library rewrite
-------------------
This commit replaces the old C/Haskell I/O implementation with a new
Haskell-only one using the new FFI & hsc2hs.
main points:
- lots of code deleted: we're about 3000 lines of C lighter,
but the amount of Haskell code is about the same.
- performance is ok: some operations are faster, others are
slower. There's still some tuning to do, though.
- the new library is designed to handle read/write streams
much better: a read/write stream gets a special kind of
handle internally called a "DuplexHandle", which actually
contains two separate handles, one for writing and one for
reading. The upshot is that you can do simultaneous reading
and writing to/from a socket or FIFO without any locking
problems. The effect is similar to calling socketToHandle
twice, except that finalization works properly (creating
two separate Handles could lead to the socket being closed
too early when one of the Handles is GC'd).
- hConnectTo and withHandleFor are gone (no one responded to
my mail on GHC users, but we can always bring 'em back if
necessary).
- I made a half-hearted attempt at keeping the system-specific
code in one place: see PrelPosix.hsc.
- I've rearranged the I/O tests and added lots more.
ghc/tests/lib/IO now contains Haskell 98-only IO tests,
ghc/test/lib/{IOExts, Directory, Time} now contain tests for
the relevant libraries. I haven't quite finished in here yet,
the IO tests work but the others don't yet.
- I haven't done anything about Unicode yet, but now we can
start to discuss what needs doing here. The new library
is using MutableByteArrays for its buffers because that
turned out to be a *lot* easier (and quicker) than malloc'd
buffers - I hope this won't cause trouble for unicode
translations though.
WARNING: Windows users refrain from updating until we've had a chance
to fix any issues that arise.
Testing: the basic H98 stuff has been pretty thoroughly tested, but
the new duplex handle stuff is still a little green.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Re-organisation of ghc/lib/std and hslibs/lang
----------------------------------------------
In brief: move deprecated features out of ghc/lib/std and into
hslibs/lang, move new FFI libraries into ghc/lib/std and start
using them.
- foreign import may now return an unboxed type (this was
advertised to work before, but in fact didn't). Subsequent
cleanups in PrelInt/PrelWord.
- Ptr is now defined in ghc/lib/std/PrelPtr.lhs. Ptr is no
longer a newtype of Addr, it is defined directly in terms of
Addr#.
- PrelAddr has disappeared from ghc/lib/std, all uses of Addr in
ghc/lib/std have been replaced with Ptr. The definitions of
Addr has been moved to hslibs/lang/Addr.lhs, as has
lots of other Addr-related stuff.
- ForeignObj has been removed from ghc/lib/std, and replaced with
ForeignPtr. The definition of ForeignObj has been moved to
hslibs/lang/ForeignObj.lhs.
- Most of the new FFI has been moved into ghc/lib/std in the form
of modules PrelMarshalAlloc, PrelCString, PrelCError,
PrelMarshalError, PrelMarshalArray, PrelMarshalUtils,
PrelCTypes, PrelCTypesISO, and PrelStorable. The corresponding
modules in hslibs/lang simply re-export the contents of these
modules.
- PrelPosixTypes defines a few POSIX types (CMode == mode_t,
etc.)
- PrelCError changed to access errno using foreign label and peek
(the POSIX book I have says that errno is guaranteed to be an
extern int, so this should be OK until I get around to making
errno thread-safe).
- Hacked the macros that generate the code for CTypes and
CTypesISO to generate much less code
(ghc/lib/std/cbits/CTypes.h).
- RtsAPI is now a bit more honest when it comes to building heap
objects (it uses the correct constructors).
- the Bits class and related stuff has been moved to ghc/lib/std
(it was simpler this way).
- Directory and System have been converted to use the new FFI.
|
| |
|
|
| |
Change the representation of IOException: add optional filename.
|
| |
|
|
|
| |
merge before-ghci -> before-ghci-branch-merged into the ghc
(non-compiler) parts of the tree.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix bug reported by Hannah Schroeter: reading a file lazily using
hGetContents and then closing it using hClose can cause the program to
fall over with a deadlock.
The reason is that when closing the file in lazyRead{Block,Line,Char},
we set the foreign object in the handle to nullFile__, which causes
the finalizer to run (at some point in the future). The finalizer
takes the MVar in the handle, frees the contents, but never puts the
MVar back. hClose then tries to take the MVar, and deadlocks.
The solution is not to set the foreign object to nullFile__ in the
first place; I'm not sure why it was done this way, and in fact it
leads to a memory leak. hClose itself has a similar problem, leading
to a leak of the fileObject.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rearrange exception stuff, as per my message on glasgow-haskell-users
recently.
The main change is the IOError type is now a synonym for Exception.
IO.ioError can therefore be used for throwing exceptions. IO.catch
still catches only IO exceptions, for backwards compatibility.
The interface exported by Exception has changed somewhat:
try :: IO a -> IO (Either Exception a)
tryJust :: (Exception -> Maybe b) -> a -> IO (Either b a)
catch :: IO a -> (Exception -> IO a) -> IO a
catchJust :: (Exception -> Maybe b) -> IO a -> (b -> IO a) -> IO a
ioErrors :: Exception -> Maybe IOError
arithExceptions :: Exception -> Maybe ArithException
errorCalls :: Exception -> Maybe String
dynExceptions :: Exception -> Maybe Dynamic
assertions :: Exception -> Maybe String
asyncExceptions :: Exception -> Maybe AsyncException
raiseInThread is now called throwTo.
Where possible, the old functions have been left around, but marked
deprecated.
|
| |
|
|
| |
Fix mondo bogosity in hGetLine, probably caused by cut-n-pasting with eyes shut.
|
| |
|
|
|
|
|
| |
Fix bug in new version of hGetLine: it didn't handle unbuffered
handles properly.
The fix involves re-importing the old code for hGetLine :-(
|
| |
|
|
|
|
|
| |
New version of hGetLine that is roughly 4 times faster than the
original, and is tail-recursive to boot.
I'm not entirely happy with the code, but it needs to get some testing.
|
| |
|
|
| |
tidy up conditionals a bit.
|
| |
|
|
|
|
|
|
| |
Fixes to commitBuffer and commitAndReleaseBuffer to maintain the
invariant that we never leave the handle buffer in a completely full
state. This fixes the crashes seen in recent Sparc builds.
Found with help from: Electric Fence &:-)
|
| |
|
|
|
| |
catch exceptions around commitBuffer and free the buffer. This closes
one memory leak in the new I/O stuff, there may be another small one left.
|
| |
|
|
| |
Removed -fcompiling-prelude flag (now removed from compiler)
|
| |
|
|
|
|
|
| |
Fix a bug in commitBuffer, and tweak the semantics of
commitBuffer/commitAndReleaseBuffer.
Add some comments on the algorithms used here.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit fixes the trace/stderr problem, and also fixes some other
problems with the I/O library.
- handles now contain a list of free buffers, which are
guaranteed to be the same size as the primary handle buffer.
- hPutStr now doesn't evaluate any part of the input string with
the handle locked. Instead, it acquires a buffer from the handle
copies characters into it, then commits the buffer. This is
better for concurrency too, because the handle is only locked
while we're actually reading/writing, not while evaluating.
- there were an even number of off-by-one errors in the I/O system
which compensated for each other. This has been fixed.
- made the I/O subsystem a little more exception-safe. It still
isn't totally exception-safe, but I can't face doing that
without a complete rewrite of this thing in Haskell.
- add hPutBufFull and hGetBufFull. The compiler probably needs to
be updated to use these too.
|
| |
|
|
|
|
|
|
|
|
| |
Make it so that -fcompiling-prelude applies only
for Prelude modules (i.e. ones called Prelxxx).
I've done this with an {-# OPTIONS #-} line in each
such module (they all has -fno-implicit-prelude anyway)
but a less repetitive approach in the Makefile would
be welcome.
|
| |
|
|
|
|
|
|
|
| |
Complete the previous commit. I think things should
be OK now; but the Hugs versioning may well need
adjusting.
NOTE: I've added PrelIO.lhs; it contains all the GHC parts
of what was IO.lhs.
|
| |
|
|
| |
Surplus to requirements (cf. last IO commit)
|
|
|
Library re-organisation:
All libraries now live under ghc/lib, which has the following structure:
ghc/lib/std -- all prelude files (libHS.a)
ghc/lib/std/cbits
ghc/lib/exts -- standard Hugs/GHC extensions (libHSexts.a)
-- available with '-fglasgow-exts'
ghc/lib/posix -- POSIX library (libHSposix.a)
ghc/lib/posix/cbits -- available with '-syslib posix'
ghc/lib/misc -- used to be hslibs/ghc (libHSmisc.a)
ghc/lib/misc/cbits -- available with '-syslib misc'
ghc/lib/concurrent -- Concurrent libraries (libHSconc.a)
-- available with '-concurrent'
Also, several non-standard prelude modules had their names changed to begin
with 'Prel' to reduce namespace pollution.
Addr ==> PrelAddr (Addr interface available in 'exts')
ArrBase ==> PrelArr
CCall ==> PrelCCall (CCall interface available in 'exts')
ConcBase ==> PrelConc
GHCerr ==> PrelErr
Foreign ==> PrelForeign (Foreign interface available in 'exts')
GHC ==> PrelGHC
IOHandle ==> PrelHandle
IOBase ==> PrelIOBase
GHCmain ==> PrelMain
STBase ==> PrelST
Unsafe ==> PrelUnsafe
UnsafeST ==> PrelUnsafeST
|
