Commits on Oct 22, 2011
  1. Temporal fix for long directory name support

    * src/code/win32-pathname.lisp (unparse-native-win32-namestring):
    add "fudge factor" of 12 characters for long directory names.
    committed Oct 22, 2011
Commits on Aug 18, 2011
  1. * src/runtime/win32-os.c (scratch): added new function references for

    calls that the LFN-handling code _definitely_ needs. Probably forgot some others...
    committed Aug 18, 2011
Commits on Aug 17, 2011
  1. * src/code/unix.lisp: don't define unix-open here on win32, see win32…

    * src/code/win32.lisp (#!+sb-unicode): define binding for _wopen/_open,
    use it in sb-unix:unix-open implementation.
    committed Aug 17, 2011
  2. * src/code/filesys.lisp (query-file-system): fixed

    typo (:file-write-date => :write-date)
    committed Aug 17, 2011
  3. * src/code/win32-pathname.lisp (unparse-native-win32-namestring): reo…

    …rdered merge-pathname args, clarified the intent.
    committed Aug 17, 2011
  4. * src/code/win32-pathname.lisp (unparse-native-win32-namestring): rem…

    …oved #P"", as it can't be dumped by the cross-compiler.
    committed Aug 17, 2011
  5. * src/code/win32-pathname.lisp (unparse-native-win32-namestring):

    fixed typo (sb-win32 => sb!win32, as usually)
    committed Aug 17, 2011
  6. * src/code/win32.lisp: moved some constant definitions, so they come …

    …before use
    committed Aug 17, 2011
  7. * src/code/filesys.lisp (query-file-system): on win32, require non-ni…

    …l KIND form native-probe-file-name
    committed Aug 17, 2011
  8. * src/code/win32.lisp (native-delete-file): don't signal win32-error,

    just return false on failure
    (native-delete-directory): ditto
    * src/code/filesys.lisp (query-file-system): split out win32
    implementation (based on native-probe-file-name); remove code that become dead
    (delete-file): merge *default-pathname-defaults* before translation,
    like other filesystem functions do (FIXME should I?);
    Use native-delete-file on win32.
    Use native-delete-directory on win32.
    merge *default-pathname-defaults* before translation,
    like other filesystem functions do (FIXME should I?)
    (map-directory): account for multiple-valued directory iterator: don't
    query file kind if it's returned already.
    committed Aug 17, 2011
  9. * src/code/win32-pathname.lisp (unparse-native-win32-namestring): fix…

    … -- missing write-string stream argument.
    committed Aug 17, 2011
  10. Win32: support lfn-prefix (including unc variant) in native namestring

    parse/unparse methods.
    * src/code/win32-pathname.lisp:
    (+long-file-name-prefix+): would be string constants if not eql-annoyance,
    (+unc-file-name-prefix+): symbol-macros by now (to allow constant folding).
    (parse-native-win32-namestring): accept UNC or LFN prefix, if it is there
    (unparse-native-win32-namestring): generate UNC/LFN prefix for
    absolute pathnames; strip them (adding #\\ for UNC) if resulting
    namestring is short enough.
    Interpret :up as :back in absolute pathnames.
    committed Aug 17, 2011
  11. Added native windows directory iterator using FindFirstFile* (not cal…

    …led yet).
    * src/code/win32.lisp (with-handle): new macro for dynamic-extent
    handle, taking care of interrupts as well as unwind-protect.
    ("FindClose"): new alien routine.
    (native-file-write-date): test syscall result.
    (native-probe-file-name): ditto
    (native-call-with-directory-iterator): reimplementation of
    call-with-native-directory-iterator, with (kind as secondary value)
    committed Aug 17, 2011
  12. * src/code/win32.lisp: added definitions for native FS access.

    (+common-lisp-epoch-filetime-seconds+): new constants
    (file-attributes):  new alien types
    (native-delete-directory): low-level functions to be used in code/filesys
    committed Aug 17, 2011
  13. * src/code/filesys.lisp (map-directory): a step toward the (future) F…

    …indFirstFileExW-based iterator, that will be able to return file kind for each directory entry
    committed Aug 17, 2011
  14. * src/code/filesys.lisp (map-directory): query-file-system :existence…

    … works for networks shares now (and will cease to be an awful kludge after LFN-patch); use it instead of realpath.
    committed Aug 17, 2011
  15. * src/code/win32.lisp (FILETIME-MEMBER): defined a version of FILETIM…

    …E that is less convenient, but is aligned correctly when used in structures.
    committed Aug 17, 2011
  16. * src/code/filesys.lisp (query-file-system): use top-level VALUES to …

    …get rid of extra return values, instead of repeated nth-value 0
    committed Aug 17, 2011
  17. * src/code/filesys.lisp (query-file-system): use top-level VALUES to …

    …get rid of extra return values, instead of repeated nth-value 0
    committed Aug 17, 2011
  18. * src/code/pathname.lisp: added :UNC pathname token

    * src/compiler/generic/vm-type.lisp (pathname-device): added :UNC as a
    possible value
    * src/code/win32-pathname.lisp (extract-device): for UNC names,
    extract :UNC and eat one slash.
    (unparse-win32-device): for :UNC device, return a [back-?]slash
    committed Aug 16, 2011
Commits on Aug 15, 2011
  1. fix bogus deadlocks from interrupts and GCs

     Going in despite the freeze since this is a regression that can
     semi-randomly break correct code. *ouch*
     Thanks to Bart Bortta and #sbcl for the analysis.
     Problem 1:
       T1 holds L1
       T2 is waiting for L1
       T2 is interrupted, interrupt handler grabs L2
       T1 starts waiting on L2
       Prior to this patch, when GET-MUTEX in T2's interrupt handler grabbed
       L2 is marked T2 as still waiting for L1 -- which is not true until
       the interrupt handler returns.
     Problem 2:
       T1 holds L1
       T2 is waiting for L1
       GC is triggered in T2 inside GET-MUTEX
       T2 grabs *ALREADY-IN-GC* lock
       GC is triggered in T1, T1 tries to get *ALREADY-IN-GC* lock.
       Prior to this patch, when T1 detects a bogus deadlock as T2 has
       been marked as waiting for L1 -- which is not true until the GC is
       finished and normal execution resumes.
     Problem 3:
       T1 holds L1
       T2 is waiting for L1
       GC is triggered in T2 inside GET-MUTEX
       T2 grabs lock L2 due to a finalizer or an after-gc-hook
       GC is triggered in T1
       T1 tries to grab L2 due to a finalizer, etc.
       Same as problem 2, but with a user-lock and POST-GC instead of
       *ALREADY-IN-GC* and SUB-GC.
     This patch fixes the issue by saving, clearing, and restoring
     the waiting-for mark in
      1) interrupt handlers
      2) SUB-GC
      3) POST-GC
    sb-studio committed Aug 15, 2011
Commits on Aug 14, 2011
  1. Fix automatic &rest to &more conversion in unsafe code

     Applying &rest lists to known/typed functions can lead to a
     :fixed values call to %more-arg-values.  In that case, unroll
     it into several %more-arg of constant indices.
     Reported, with test case, by Lutz Euler (lp#826459).
    pkhuong committed Aug 14, 2011
  2. New function: SB-EXT:SPIN-LOOP-HINT

     Some architectures have developed ways to help the processor execute
     spin loops efficiently; expose them, where applicable, via
     Assembles to PAUSE on x86oids, and to nothing on other platforms.
    pkhuong committed Aug 14, 2011
  3. Missing NEWS entry for 4e431db

     Also, missing credit for Lutz Euler's bug report on a7b24b5.
    pkhuong committed Aug 14, 2011
  4. Handle SIMPLE-TYPE-ERROR when propagating bounds

     Type conversions (e.g. bignum to float) may signal a
     SIMPLE-TYPE-ERROR rather than returning an infinity.
     Treat that case like an unknown value.
     Add two tests.
     Fixes lp#819269.
    pkhuong committed Aug 14, 2011
  5. New function: SB!KERNEL:%MULTIPLY-HIGH

     * Does the same thing as only returning the first value of %MULTIPLY,
       only better on some platforms.
     * Implemented vas VOPs on x86, x86-64 and PPC.  The PPC code sequence
       is fully untested, and merely looks correct.
     * VOPs for fixnum first argument are included, but will only be used
       when the result is forcibly marked as fixnum, e.g., with TRULY-THE.
       Questionnable, but I'd rather err on the side of straightforwardness
       rather than put even more pressure on representation selection.
     * Use it in the division-by-multiplication transform for unsigned
       TRUNCATE by constant.
    pkhuong committed Aug 14, 2011
  6. More efficient integer division by multiplication

     * Exploit restricted range for inputs (e.g. fixnums).
     * When the divisor is even, simplify with a mask instead of a shift.
     * Clean up the code a bit: we don't want modular arithmetic, it's
       actually all guaranteed not to wrap around.
     * Change the test so that larger divisors are a bit more likely to
       get tested too.
     * Lots more things can be done, including generalizing the transform
       to pretty much arbitrary rational divisor, or avoiding the costly
       general code sequence in nearly all cases.  Unfortunately, it's a
       lot of (somewhat original, at that) code, and can be fairly slow;
       if it matters to someone, I can try and find a compromise (contrib?).
    pkhuong committed Aug 14, 2011
  7. Various (mostly) SSE x86-64 instruction definition fixes and additions

     * Fix typos in mnemonics and non-temporal store definitions;
     * Add some instructions;
     * CLFLUSH was totally broken.
    Taken from Alexander Gavrilov's SSE intrinsics branch.
    pkhuong committed Aug 14, 2011
Commits on Aug 13, 2011
  1. fix DESCRIBE on compiled closures from EVAL

    nikodemus committed Aug 13, 2011
Commits on Aug 12, 2011
  1. allow user-defined STRING synonyms in MAKE-SEQUENCE

    ...and MAP, MERGE, COERCE and CONCATENATE too.  This also meant
    working a bit on TYPEXPAND, to make (STRING <N>) be unexpanded in the
    same way as STRING, and consequently on VALUES-SPECIFIER-TYPE to
    ensure that the system would still recognize those types.
    Include test cases for the sequence functions but not for TYPEXPAND.
    csrhodes committed Aug 12, 2011
Commits on Aug 11, 2011
  1. Declaim sb-di::valid-lisp-pointer-p inline.

    Making it inline prevents from creating a SAP, and thus reduces
    consing and improves performance. It's called a lot by
    map-allocated-objects when GCing is disabled, so consing less is a
    desirable property.
    Reported by `lisper' on #lisp
    stassats committed Aug 11, 2011
  2. Fix QUERY-FILE-SYSTEM for Windows UNC and device file names

    Thanks to Anton Kovalenko.
    lichtblau committed Aug 11, 2011
Commits on Aug 10, 2011
  1. sb-bsd-sockets: Fix error code handling on Windows

      Factored out into a new function for the win32-specific
      WSAGetLastError call.
      Run-time results now match the existing constant definitions.
    Thanks to Anton Kovalenko.
    committed with lichtblau Aug 18, 2010
  2. sb-simple-streams: In test lp491087, merge the pathname

      Needed for Windows.
    Thanks to Anton Kovalenko.
    lichtblau committed Aug 10, 2011
  3. sb-simple-streams: Close files before deleting them on revert

    Thanks to Anton Kovalenko.
    lichtblau committed Aug 10, 2011