|
|
@8063
|
13 years |
gb |
FORCE-OUTPUT to *DEBUG-IO* after backtrace in ABNORMAL-PROGRAM-EXIT.
|
|
|
@8056
|
13 years |
wws |
Stop make-instantiate-lambda-for-class-cell from doing (setf (cdr nil) …
|
|
|
@8042
|
13 years |
gb |
Bump FASL version.
|
|
|
@8041
|
13 years |
gb |
Set save_rbp slot of newly-created tsp frames.
Ensure that linked …
|
|
|
@8040
|
13 years |
gb |
Set save_rbp slot of newly-created tsp frames.
|
|
|
@8039
|
13 years |
gb |
tsp_frame: second slot is used to save paired rbp, has never been
used …
|
|
|
@8038
|
13 years |
gb |
Ignore SIGQUIT.
|
|
|
@8037
|
13 years |
gb |
Lots of changes to APPLY-IN-FRAME and supporting functions.
Define …
|
|
|
@8036
|
13 years |
gb |
%PRINT-CALL-HISTORY-INTERNAL: mark exeption frames, not their children.
|
|
|
@8035
|
13 years |
gb |
Break-loop commands for :apply-in-frame, :return-from-frame.
|
|
|
@8034
|
13 years |
gb |
SYMBOL-VALUE-IN-PROCESS (and SETF therof) check for process = …
|
|
|
@8033
|
13 years |
gb |
Ensure that %UPDATE-CPL always returns the CPL.
|
|
|
@8032
|
13 years |
gb |
Handle unquoted type T in TYPEP.
In REQUIRE-TYPE, don't devolve to …
|
|
|
@8031
|
13 years |
gb |
When making frames on the tstack or cstack, store %rbp in the
word …
|
|
|
@8030
|
13 years |
gb |
Add definitions for TSP-FRAME/CSP-FRAME.
|
|
|
@8028
|
13 years |
wws |
make-instantiate-lambda-for-class-cell, the function that generates …
|
|
|
@8027
|
13 years |
gb |
Remove #+are-you-kidding version of %%APPLY-IN-FRAME.
Fix a few bugs …
|
|
|
@8026
|
13 years |
gb |
Apply-in-frame support for x86-64.
|
|
|
@8025
|
13 years |
gb |
BACKTRACE-CALL-ARGUMENTS: try to show real arguments if before #args
trap.
|
|
|
@8024
|
13 years |
gb |
Move #+ppc-target stuff from backtrace-lds.lisp here.
|
|
|
@8023
|
13 years |
gb |
FRAME-SUPPLIED-ARGS: return actual argument values if PC is before/at …
|
|
|
@8022
|
13 years |
gb |
ARG-NAMES-FROM-MAP: give up if PC is before/at …
|
|
|
@8021
|
13 years |
gb |
Define xp-mxcsr; use it to set STATUS slot of ARITHMETIC-ERROR.
|
|
|
@8020
|
13 years |
gb |
ARITHMETIC-ERROR has "status" slot.
|
|
|
@8019
|
13 years |
gb |
Use CHECK-MIN-MAX-NARGS to anchor both UUOs.
|
|
|
@8018
|
13 years |
gb |
New CHECK-NARGS lapmacro uses anchored UUOs.
|
|
|
@8017
|
13 years |
gb |
Use multiple frag-lists ("sections"), so that we can move UUOs out of …
|
|
|
@8016
|
13 years |
gb |
Do unsigned comparisons for arg-checking.
CHECK-MIN-MAX-NARGS: need …
|
|
|
@8012
|
13 years |
gb |
%APPLY-IN-FRAME-PROTO.
|
|
|
@8011
|
13 years |
gb |
Build a stack frame in %FIXNUM-TRUNCATE, to avoid endless rounds of
mail.
|
|
|
@8010
|
13 years |
gb |
ARG-CHECK-TRAP-PC-LIMIT.
|
|
|
@8009
|
13 years |
gb |
check-min-max-args: use the same anchor in both anchored UUOs.
|
|
|
@8008
|
13 years |
gb |
In finish_function_entry, ensure that the xp's notion of where
rsp is …
|
|
|
@8007
|
13 years |
gb |
Add a couple of fields to XCF. Note that this will cause a benign …
|
|
|
@8005
|
13 years |
gb |
Slot-unbound-trap terminates a block (why did it ever not do so ?)
by …
|
|
|
@8004
|
13 years |
gb |
Set foreign_sp (at time of exception) and previous-xframe slots
of xcf …
|
|
|
@8003
|
13 years |
gb |
Add a couple of fields to exception_callback_frame, to support …
|
|
|
@7992
|
13 years |
gb |
Default *TYPECHECK-SLOTS-IN-OPTIMIZED-MAKE-INSTANCE* to T.
In the …
|
|
|
@7967
|
13 years |
gb |
Need %STRING-HASH for PPC.
|
|
|
@7966
|
13 years |
gb |
PPC32 vinsns: :jump attribute use for pc-relative jump only.
|
|
|
@7965
|
13 years |
gb |
ppc64 vinsns: :jump vinsn attribute now means "pc-relative jump" only.
|
|
|
@7961
|
13 years |
gb |
Spell UVSIZE with an "S".
|
|
|
@7960
|
13 years |
gb |
Try to avoid branching around UUOs, since some implemenations don't
do …
|
|
|
@7959
|
13 years |
gb |
new hash-table iteration scheme.
|
|
|
@7958
|
13 years |
gb |
CLASS-WRAPER caches CPL.
|
|
|
@7957
|
13 years |
gb |
New hash-table iteration scheme.
|
|
|
@7956
|
13 years |
gb |
New hash-table iteration scheme.
New WITH-PROCESS-WHOSTATE just binds …
|
|
|
@7955
|
13 years |
gb |
GCTIMES reported in INTERNAL-TIME-UNITS-PER-SECOND.
|
|
|
@7954
|
13 years |
gb |
Time-reporting changes.
|
|
|
@7953
|
13 years |
gb |
Fix comment indentation.
|
|
|
@7952
|
13 years |
gb |
CHAR/SET-CHAR: define in terms of SCHAR.
|
|
|
@7951
|
13 years |
gb |
Add timeval->microseconds.
|
|
|
@7950
|
13 years |
gb |
Continue Bill's changes to remove brain-death from %%TYPEP on arrays …
|
|
|
@7949
|
13 years |
gb |
Keep whostate in a per-thread variable, not in the PROCESS-OBJECT.
|
|
|
@7948
|
13 years |
gb |
INTERNAL-TIME-UNITS-PER-SECOND, whostate tweaks.
|
|
|
@7947
|
13 years |
gb |
INTERNAL-TIME-UNITS-PER-SECOND is 1000000 on 64-bit machines.
(Can …
|
|
|
@7946
|
13 years |
gb |
Maintain CPL in wrapper ...
Slot type-predicates can be NULL (instead …
|
|
|
@7945
|
13 years |
gb |
New CLASS-CELL-TYPEP, tries to get CPL from (valid) wrapper.
Maintain …
|
|
|
@7944
|
13 years |
gb |
Process-whostate is now the per-thread binding of *WHOSTATE*.
|
|
|
@7943
|
13 years |
gb |
A few small changes; there may be bad bugs in EQUAL/EQUALP hashing. …
|
|
|
@7942
|
13 years |
gb |
%STRING-HASH: like %PNAME-HASH, but with a (possibly non-zero)
offset.
|
|
|
@7941
|
13 years |
gb |
Try to avoid jumping/branching to jumps. (NB: this requires some …
|
|
|
@7940
|
13 years |
gb |
Try harder (maybe much harder) to inline TYPEP calls involving …
|
|
|
@7939
|
13 years |
gb |
Recognize CHAR-CODE's result type.
|
|
|
@7938
|
13 years |
gb |
Peephole on the fly: turn
(push r1)
(movq top-of-stack r2)
into
…
|
|
|
@7937
|
13 years |
gb |
PRINT-OBJECT for DISASSEMBLED-INSTRUCTION.
|
|
|
@7936
|
13 years |
gb |
More tweaks.
|
|
|
@7935
|
13 years |
gb |
Don't use a bignum in HASHED-BY-IDENTITY, since there's only one …
|
|
|
@7925
|
13 years |
gb |
Add :monitor-futex-wait to *known-optional-features*.
…
|
|
|
@7924
|
13 years |
wws |
TYPEP was consing on a user-defined array type, due to the call to …
|
|
|
@7923
|
13 years |
wws |
optimize-generic-function-dispatching now calls …
|
|
|
@7918
|
13 years |
gb |
Add TYPE-PREDICATEs for ARGS-CTYPE and its subclasses.
|
|
|
@7917
|
13 years |
gb |
Provide type-predicate for NAMED-CTYPE; the type system otherwise …
|
|
|
@7908
|
13 years |
gb |
Cache a symbol's hash code in the plist slot along with the plist, so …
|
|
|
@7901
|
13 years |
gb |
Implement HASH-TABLE-OWNER and CLAIM-HASH-TABLE; the latter probably …
|
|
|
@7900
|
13 years |
gb |
Don't generate (often bad ...) typechecks in in the optimized …
|
|
|
@7897
|
13 years |
gb |
Conditionally hash symbols by address.
|
|
|
@7896
|
13 years |
gb |
Revert eep.adress to old (anchored UUO out-of-line) behavior, since …
|
|
|
@7895
|
13 years |
gb |
After a couple of failed attempts, get the improvement to the
&key …
|
|
|
@7894
|
13 years |
gb |
Copy the build-time value of *BUILD-TIME-OPTIONAL-FEATURES* into the …
|
|
|
@7893
|
13 years |
gb |
Add :HASH-SYMBOLS-BY-ADDRESS to *KNOWN-OPTIONAL-FEATURES*.
|
|
|
@7892
|
13 years |
gb |
Fix type in call to %CHECK-ANCHORED-UUO in the decode of #xb* UUOs.
|
|
|
@7891
|
13 years |
gb |
Try to inorporate *optional-features* into *openmcl-version* format string.
|
|
|
@7890
|
13 years |
gb |
Get 3d array right in %TYPE-ERROR-TYPESPECS%.
|
|
|
@7889
|
13 years |
gb |
Try to avoide a GF call in %SHARED-INITIALIZE.
Look at %CLASS.CPL …
|
|
|
@7888
|
13 years |
gb |
Define STANDARD-INSTANCE-P (which is inlined) before its first use.
|
|
|
@7887
|
13 years |
gb |
*OPTIONAL-FEATURES*.
|
|
|
@7886
|
13 years |
gb |
Spin a little faster (when we use spinlocks at all)
|
|
|
@7870
|
13 years |
gb |
back out of last change
|
|
|
@7869
|
13 years |
gb |
Try to handle the &key &allow-other-keys (no explicit keywords)
case …
|
|
|
@7868
|
13 years |
gb |
Fix absolute pc if relative PC was adjusted in lisp handler in …
|
|
|
@7867
|
13 years |
gb |
Control transfer after anchored UUO not working right; revert …
|
|
|
@7866
|
13 years |
gb |
If we have to wait in %PROMOTE-RWLOCK, release the spinlock/futex …
|
|
|
@7865
|
13 years |
gb |
GETHASH locks hash-table for reading, promotes to write-access only if …
|
|
|
@7864
|
13 years |
gb |
Don't just do MEMQ on a long list of classes in optimized reader …
|
|
|
@7863
|
13 years |
gb |
RELOCATABLE_FULLTAG_MASK has to include fulltag_symbol, …
|
|
|
@7861
|
13 years |
gb |
If the callback to handle a UUO claims that we should skip -1 bytes, …
|
|
|
@7860
|
13 years |
gb |
Use the new scheme so that most UUOs (at least those used to signal …
|
|
|
@7859
|
13 years |
gb |
Support the use of 2 (x86-asm) frag-lists for vinsn expansion (one for …
|
|
|
@7858
|
13 years |
gb |
Always use NOPs to pad for alignment. (We might want to align loop …
|
|
|