ecl.git
15 years agoWhen using threads and sockets, the stream is a two-way streams and its components... ECL.10.2.1
Juan Jose Garcia Ripoll [Sat, 13 Feb 2010 14:35:42 +0000 (15:35 +0100)]
When using threads and sockets, the stream is a two-way streams and its components must be closed separately.

15 years agoExtend the set of types supported by UFFI (Matthew Mondor).
Juan Jose Garcia Ripoll [Sat, 13 Feb 2010 14:29:41 +0000 (15:29 +0100)]
Extend the set of types supported by UFFI (Matthew Mondor).

15 years agoFinalize CHANGELOG for the 10.2.1 release.
Juan Jose Garcia Ripoll [Sat, 13 Feb 2010 13:56:54 +0000 (14:56 +0100)]
Finalize CHANGELOG for the 10.2.1 release.

15 years agoSocket streams are now two-way streams.
Juan Jose Garcia Ripoll [Fri, 12 Feb 2010 17:33:04 +0000 (18:33 +0100)]
Socket streams are now two-way streams.

15 years agoReorganize the code for OP_LABELS/OP_FLET in ecl_interpret() to occupy less space...
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 22:20:29 +0000 (23:20 +0100)]
Reorganize the code for OP_LABELS/OP_FLET in ecl_interpret() to occupy less space and get faster.

15 years agoFactor out some error conditions from the interpreter loop.
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 21:05:57 +0000 (22:05 +0100)]
Factor out some error conditions from the interpreter loop.

15 years agoMark in stacks.h the error conditions with ecl_unlikely().
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 20:49:31 +0000 (21:49 +0100)]
Mark in stacks.h the error conditions with ecl_unlikely().

15 years agoRemove bogus inline forms for FLOOR and TRUNCATE
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 13:26:52 +0000 (14:26 +0100)]
Remove bogus inline forms for FLOOR and TRUNCATE

15 years agoAvoid using ecl_internal_error() also in ecl_bds_overflow and ecl_frs_oveflow.
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 12:01:45 +0000 (13:01 +0100)]
Avoid using ecl_internal_error() also in ecl_bds_overflow and ecl_frs_oveflow.

15 years agoecl_cs_overflow() and SIGSEGV handler now share the same strategy: jump to the outerm...
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 11:46:13 +0000 (12:46 +0100)]
ecl_cs_overflow() and SIGSEGV handler now share the same strategy: jump to the outermost frame when the error can not be cured.

15 years agoECL accepts FTYPE proclamations for SETF-functions.
Juan Jose Garcia Ripoll [Mon, 8 Feb 2010 10:47:30 +0000 (11:47 +0100)]
ECL accepts FTYPE proclamations for SETF-functions.

15 years agoAvoid using _ecl_big_register_normalize() in alloc_2.d
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 22:22:26 +0000 (23:22 +0100)]
Avoid using _ecl_big_register_normalize() in alloc_2.d

15 years agoNumber of allocated bytes was didn't have to be multiplied by anything. Reset GC...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 22:02:29 +0000 (23:02 +0100)]
Number of allocated bytes was didn't have to be multiplied by anything. Reset GC counters every time we call GC-STATS.

15 years agoTurn warnings about ECL not being able to cope with certain types into compiler notes.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 20:05:15 +0000 (21:05 +0100)]
Turn warnings about ECL not being able to cope with certain types into compiler notes.

15 years agoUntil we merge in the new code for type handling, we turn all warnings about types...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 19:56:36 +0000 (20:56 +0100)]
Until we merge in the new code for type handling, we turn all warnings about types into notes.

15 years agoImplemented SBCL's extension EXT:*INVOKE-DEBUGGER-HOOK*
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 19:42:29 +0000 (20:42 +0100)]
Implemented SBCL's extension EXT:*INVOKE-DEBUGGER-HOOK*

15 years agoIn compiler-debugger, do not muffle errors because they might be caused by user handl...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 15:31:06 +0000 (16:31 +0100)]
In compiler-debugger, do not muffle errors because they might be caused by user handlers. Ideally internal errors should have been already muffled by handle-compiler-internal-error

15 years agoSimplified the SETF expansion of GETF: it no longer produces an ignored variable...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 15:05:32 +0000 (16:05 +0100)]
Simplified the SETF expansion of GETF: it no longer produces an ignored variable when the default value is not supplied

15 years agoSolved a problem with packages to be created in a compiled file not being recognized...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 14:48:19 +0000 (15:48 +0100)]
Solved a problem with packages to be created in a compiled file not being recognized as having been created -- the value of cl_core.packages_to_be_created was not properly updated.

15 years agoFixed typo.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 13:54:16 +0000 (14:54 +0100)]
Fixed typo.

15 years agoExport *SUPPRESS-COMPILER-{NOTES,WARNINGS,MESSAGES}*
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 11:47:01 +0000 (12:47 +0100)]
Export *SUPPRESS-COMPILER-{NOTES,WARNINGS,MESSAGES}*

15 years agoVariables in the PPRINT-LOGICAL-BLOCK macro must be marked as ignorable.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 11:26:39 +0000 (12:26 +0100)]
Variables in the PPRINT-LOGICAL-BLOCK macro must be marked as ignorable.

15 years agoDecrease verbosity in compiler note printing.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 11:17:53 +0000 (12:17 +0100)]
Decrease verbosity in compiler note printing.

15 years agoFixed a typo in EXT:GC-STATS that left statistics printing activated.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 11:17:35 +0000 (12:17 +0100)]
Fixed a typo in EXT:GC-STATS that left statistics printing activated.

15 years agoAdded an ignorable dclaration to the expansion of HANDLER-CASE.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 11:04:38 +0000 (12:04 +0100)]
Added an ignorable dclaration to the expansion of HANDLER-CASE.

15 years agoIn compiler messages, print the actual toplevel form, not the macroexpanded version
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 10:57:34 +0000 (11:57 +0100)]
In compiler messages, print the actual toplevel form, not the macroexpanded version

15 years agoRemove useless compiler note about keywords being reused
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 10:20:41 +0000 (11:20 +0100)]
Remove useless compiler note about keywords being reused

15 years agoFixed typo in the implementation of THE.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 10:13:37 +0000 (11:13 +0100)]
Fixed typo in the implementation of THE.

15 years agoRemove debug statement
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 10:08:35 +0000 (11:08 +0100)]
Remove debug statement

15 years agoThe compiler now uses THE forms with a VALUES type.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 10:08:09 +0000 (11:08 +0100)]
The compiler now uses THE forms with a VALUES type.

15 years agoWhen parsing a function type that contained &optional, the compiler did not eliminate...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 09:50:38 +0000 (10:50 +0100)]
When parsing a function type that contained &optional, the compiler did not eliminate this word before parsing the next argument.

15 years agoSymbols SI:+{IO,ECL}-SYNTAX-PROGV-LIST+ were not declared constants.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 09:36:16 +0000 (10:36 +0100)]
Symbols SI:+{IO,ECL}-SYNTAX-PROGV-LIST+ were not declared constants.

15 years agoInternal variable SI:*ALLOW-WITH-INTERRUPTS* not declared special.
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 00:39:11 +0000 (01:39 +0100)]
Internal variable SI:*ALLOW-WITH-INTERRUPTS* not declared special.

15 years agoFixed typo in CHANGELOG
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 00:29:20 +0000 (01:29 +0100)]
Fixed typo in CHANGELOG

15 years agoVariables declared ignorable no longer produce a warning when removed from LET/LET...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 00:21:26 +0000 (01:21 +0100)]
Variables declared ignorable no longer produce a warning when removed from LET/LET* forms.

15 years agoWhen inlining an APPLY form, the variable for processing the argument list may be...
Juan Jose Garcia Ripoll [Sun, 7 Feb 2010 00:21:09 +0000 (01:21 +0100)]
When inlining an APPLY form, the variable for processing the argument list may be declared ignorable, in order to prevent warnings.

15 years ago(EXT:GC-STATS :FULL) now tells the garbage collector to produce full statistics.
Juan Jose Garcia Ripoll [Sat, 6 Feb 2010 23:02:52 +0000 (00:02 +0100)]
(EXT:GC-STATS :FULL) now tells the garbage collector to produce full statistics.

15 years agoImprove previous change.
Juan Jose Garcia Ripoll [Sat, 6 Feb 2010 22:57:49 +0000 (23:57 +0100)]
Improve previous change.

15 years agoMingw defines 'environ' to be a macro instead of a global variable.
Juan Jose Garcia Ripoll [Sat, 6 Feb 2010 22:51:25 +0000 (23:51 +0100)]
Mingw defines 'environ' to be a macro instead of a global variable.

15 years agoUpgraded to ASDF 1.604 to solve a problem with restarts.
Juan Jose Garcia Ripoll [Fri, 5 Feb 2010 18:26:12 +0000 (19:26 +0100)]
Upgraded to ASDF 1.604 to solve a problem with restarts.

15 years agoUpgraded ASDF to 1.603.
Juan Jose Garcia Ripoll [Thu, 4 Feb 2010 22:27:12 +0000 (23:27 +0100)]
Upgraded ASDF to 1.603.

15 years agoUse __builtin_expect() in the dispatch functions.
Juan Jose Garcia Ripoll [Thu, 4 Feb 2010 20:25:17 +0000 (21:25 +0100)]
Use __builtin_expect() in the dispatch functions.

15 years agoReduce the size of the core using GCC's __attribute__((noreturn))
Juan Jose Garcia Ripoll [Thu, 4 Feb 2010 20:18:24 +0000 (21:18 +0100)]
Reduce the size of the core using GCC's __attribute__((noreturn))

15 years agoReplace use of __builtin_expect() with two macros, ecl_{likely,unlikely}
Juan Jose Garcia Ripoll [Thu, 4 Feb 2010 15:54:05 +0000 (16:54 +0100)]
Replace use of __builtin_expect() with two macros, ecl_{likely,unlikely}

15 years agoMark FEprogram_error as noreturn.
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 23:31:03 +0000 (00:31 +0100)]
Mark FEprogram_error as noreturn.

15 years agoFEinvalid_function and FEundefined_function marked as __attribute((noreturn)).
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 23:05:01 +0000 (00:05 +0100)]
FEinvalid_function and FEundefined_function marked as __attribute((noreturn)).

15 years agoUse of __builtin_expect() to mark rare error conditions. Use of __attribute__((noretu...
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 22:30:16 +0000 (23:30 +0100)]
Use of __builtin_expect() to mark rare error conditions. Use of __attribute__((noreturn)) on error functions which are used at the beginning of a function and do not have possibility of triggering a longjmp int the same function.

15 years agoCLOS::*NEXT-METHODS* and CLOS::.COMBINED-METHOD-ARGS. are now globally declared special.
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 22:29:08 +0000 (23:29 +0100)]
CLOS::*NEXT-METHODS* and CLOS::.COMBINED-METHOD-ARGS. are now globally declared special.

15 years agoUndo previous garbage collector change.
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 21:43:04 +0000 (22:43 +0100)]
Undo previous garbage collector change.

15 years agoPushed version number forward.
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 21:21:45 +0000 (22:21 +0100)]
Pushed version number forward.

15 years agoNew variables c:*user-ld-flags* and c:*user-cc-flags*
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 21:14:07 +0000 (22:14 +0100)]
New variables c:*user-ld-flags* and c:*user-cc-flags*

15 years agoWhen marking the environment, only mark pages that are dirty.
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 20:01:05 +0000 (21:01 +0100)]
When marking the environment, only mark pages that are dirty.

15 years agoECL now ships with ASDF version 1.596
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 20:00:39 +0000 (21:00 +0100)]
ECL now ships with ASDF version 1.596

15 years agoECL now ships with ASDF version 1.596
Juan Jose Garcia Ripoll [Wed, 3 Feb 2010 20:00:23 +0000 (21:00 +0100)]
ECL now ships with ASDF version 1.596

15 years agoFixed typo.
Juan Jose Garcia Ripoll [Tue, 2 Feb 2010 18:39:35 +0000 (19:39 +0100)]
Fixed typo.

15 years agoOwn marking routine for garbage collection.
Juan Jose Garcia Ripoll [Tue, 2 Feb 2010 15:53:53 +0000 (16:53 +0100)]
Own marking routine for garbage collection.

15 years agoDispatch allocation to functions using pointers.
Juan Jose Garcia Ripoll [Tue, 2 Feb 2010 09:30:07 +0000 (10:30 +0100)]
Dispatch allocation to functions using pointers.

15 years agoDo not define the free list and new kind unless needed.
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 23:05:50 +0000 (00:05 +0100)]
Do not define the free list and new kind unless needed.

15 years agoWe are using __builtin_expect() which is not portable. Define a macro for non-GCC...
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 23:04:38 +0000 (00:04 +0100)]
We are using __builtin_expect() which is not portable. Define a macro for non-GCC compilers.

15 years agoActivated code for inline allocation.
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 23:00:14 +0000 (00:00 +0100)]
Activated code for inline allocation.

15 years agoChanged header so that we can use GC_arrays from ECL.
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 22:56:21 +0000 (23:56 +0100)]
Changed header so that we can use GC_arrays from ECL.

15 years agoGroup the precise marking code so that it builds also without --enable-precisegc.
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 15:47:27 +0000 (16:47 +0100)]
Group the precise marking code so that it builds also without --enable-precisegc.

15 years agoImplemented precise marking of objects using Boehm's typed objects.
Juan Jose Garcia Ripoll [Mon, 1 Feb 2010 13:54:59 +0000 (14:54 +0100)]
Implemented precise marking of objects using Boehm's typed objects.

15 years agoDo not shrink bignum registers.
Juan Jose Garcia Ripoll [Fri, 29 Jan 2010 13:10:49 +0000 (14:10 +0100)]
Do not shrink bignum registers.

15 years agoMoved more uses of _ecl_big_register*() out of sources into big.d. Introduced _ecl_bi...
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 22:36:01 +0000 (23:36 +0100)]
Moved more uses of _ecl_big_register*() out of sources into big.d. Introduced _ecl_big_gcd().

15 years agoIntoduced _ecl_fix_times_fix, _ecl_big_ceiling and _ecl_big_floor.
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 22:08:16 +0000 (23:08 +0100)]
Intoduced _ecl_fix_times_fix, _ecl_big_ceiling and _ecl_big_floor.

15 years agoCreated _ecl_fix_times_fix and used it in num_arith.d
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 21:13:40 +0000 (22:13 +0100)]
Created _ecl_fix_times_fix and used it in num_arith.d

15 years agoReactivate allocating compact bignums.
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 20:42:34 +0000 (21:42 +0100)]
Reactivate allocating compact bignums.

15 years agoIntroduced _ecl_big_negate.
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 17:14:04 +0000 (18:14 +0100)]
Introduced _ecl_big_negate.

15 years agoFixed and used _ecl_fix_minus_big and _ecl_big_minus_big.
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 17:03:59 +0000 (18:03 +0100)]
Fixed and used _ecl_fix_minus_big and _ecl_big_minus_big.

15 years agoFixed and used _ecl_big_plus_{fix,big}
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 16:57:43 +0000 (17:57 +0100)]
Fixed and used _ecl_big_plus_{fix,big}

15 years agoNew functions for adding, subtracting and multiplying bignums.
Juan Jose Garcia Ripoll [Thu, 28 Jan 2010 16:25:59 +0000 (17:25 +0100)]
New functions for adding, subtracting and multiplying bignums.

15 years agoFixed typo in big.d
Juan Jose Garcia Ripoll [Wed, 27 Jan 2010 12:52:41 +0000 (13:52 +0100)]
Fixed typo in big.d

15 years agoIn 'dpp.c', the value of the environment is reused in the @(return) statements.
Juan Jose Garcia Ripoll [Wed, 27 Jan 2010 12:18:17 +0000 (13:18 +0100)]
In 'dpp.c', the value of the environment is reused in the @(return) statements.

15 years agoReplaced ecl_symbol_slot and ecl_set_symbol with ecl_bds_read and ecl_bds_set.
Juan Jose Garcia Ripoll [Tue, 26 Jan 2010 22:14:33 +0000 (23:14 +0100)]
Replaced ecl_symbol_slot and ecl_set_symbol with ecl_bds_read and ecl_bds_set.

15 years agoNo bds check is needed in compiled functions, since this is included in the ecl_bds_b...
Juan Jose Garcia Ripoll [Tue, 26 Jan 2010 21:29:53 +0000 (22:29 +0100)]
No bds check is needed in compiled functions, since this is included in the ecl_bds_bind routine.

15 years agoIn critical circumstances, a newly bound special variable that causes the thread...
Juan Jose Garcia Ripoll [Tue, 26 Jan 2010 13:23:15 +0000 (14:23 +0100)]
In critical circumstances, a newly bound special variable that causes the thread-local stack to grow could result in the variable not being bound.

15 years agoChanges to make the MSVC port work again, including fixes for EXT:RUN-PROGRAM and...
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 21:58:06 +0000 (22:58 +0100)]
Changes to make the MSVC port work again, including fixes for EXT:RUN-PROGRAM and EXT:EXTERNAL-PROCESS-WAIT.

15 years agoFixed typo in ext:external-process-status
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 20:26:19 +0000 (21:26 +0100)]
Fixed typo in ext:external-process-status

15 years agoSometimes, in external-process-wait, the value of code was undetermined.
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 20:25:56 +0000 (21:25 +0100)]
Sometimes, in external-process-wait, the value of code was undetermined.

15 years agoIntroduced EXT:EXTERNAL-PROCESS structure. RUN-PROCESS creates that structure. New...
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 19:51:36 +0000 (20:51 +0100)]
Introduced EXT:EXTERNAL-PROCESS structure. RUN-PROCESS creates that structure. New function EXT:EXTERNAL-PROCESS-WAIT.

15 years agoMissing argument for a warning
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 19:50:31 +0000 (20:50 +0100)]
Missing argument for a warning

15 years agoMissing arguments to several warnings (S. Bourakev).
Juan Jose Garcia Ripoll [Mon, 25 Jan 2010 08:09:10 +0000 (09:09 +0100)]
Missing arguments to several warnings (S. Bourakev).

15 years agoMP:PROCESS-ENABLE did not return the process object.
Juan Jose Garcia Ripoll [Sun, 24 Jan 2010 22:39:20 +0000 (23:39 +0100)]
MP:PROCESS-ENABLE did not return the process object.

15 years agoEXT:RUN-PROGRAM now accepts an :ENVIRON argument.
Juan Jose Garcia Ripoll [Sun, 24 Jan 2010 22:09:28 +0000 (23:09 +0100)]
EXT:RUN-PROGRAM now accepts an :ENVIRON argument.

15 years agoNew function EXT:ENVIRON
Juan Jose Garcia Ripoll [Sun, 24 Jan 2010 21:37:43 +0000 (22:37 +0100)]
New function EXT:ENVIRON

15 years agoUse inline functions also for the single threaded version
Juan Jose Garcia Ripoll [Sun, 24 Jan 2010 08:49:36 +0000 (09:49 +0100)]
Use inline functions also for the single threaded version

15 years agoThe code for binding special variables is now in inline functions.
Juan Jose Garcia Ripoll [Sat, 23 Jan 2010 22:45:21 +0000 (23:45 +0100)]
The code for binding special variables is now in inline functions.

15 years agoIn alloc_2.d, when recycling a symbol binding index we cannot cons inside
Juan Jose Garcia Ripoll [Sat, 23 Jan 2010 22:29:57 +0000 (23:29 +0100)]
In alloc_2.d, when recycling a symbol binding index we cannot cons inside
a THREAD_OP_LOCK() region because this may be interrupted by a garbage
collection process - instead perform the allocation outside and then
do an atomic update of the indices list.

15 years agoInline ecl_bds_push() and simplify ecl_symbol_slot(), which will never be called...
Juan Jose Garcia Ripoll [Sat, 23 Jan 2010 00:35:17 +0000 (01:35 +0100)]
Inline ecl_bds_push() and simplify ecl_symbol_slot(), which will never be called on Cnil.

15 years agoRemove unused variables in stacks.d
Juan Jose Garcia Ripoll [Fri, 22 Jan 2010 18:37:57 +0000 (19:37 +0100)]
Remove unused variables in stacks.d

15 years agoSave another few bytes by redefining ecl_bds_overflow() to return a pointer.
Juan Jose Garcia Ripoll [Fri, 22 Jan 2010 18:15:52 +0000 (19:15 +0100)]
Save another few bytes by redefining ecl_bds_overflow() to return a pointer.

15 years agoAnother variant in which the memory address of the bindings and their size is stored...
Juan Jose Garcia Ripoll [Fri, 22 Jan 2010 18:07:18 +0000 (19:07 +0100)]
Another variant in which the memory address of the bindings and their size is stored in the environment itself.

15 years agoRemove one test from the implementation of ecl_bds_bind().
Juan Jose Garcia Ripoll [Fri, 22 Jan 2010 17:47:55 +0000 (18:47 +0100)]
Remove one test from the implementation of ecl_bds_bind().

15 years agoSpecial variable bindings using a thread-local array instead of a hash table.
Juan Jose Garcia Ripoll [Fri, 22 Jan 2010 16:47:10 +0000 (17:47 +0100)]
Special variable bindings using a thread-local array instead of a hash table.

15 years agoSome comments on thread_entry_point().
Juan Jose Garcia Ripoll [Thu, 21 Jan 2010 14:56:44 +0000 (15:56 +0100)]
Some comments on thread_entry_point().

15 years agoThe environment of a thread is now allocated and managed by the parent thread. This...
Juan Jose Garcia Ripoll [Thu, 21 Jan 2010 14:39:41 +0000 (15:39 +0100)]
The environment of a thread is now allocated and managed by the parent thread. This decreases the time that passes since a thread is started till it is interruptible.

15 years agoecl_init_env() split into ecl_init_env and ecl_cs_set_org() so that the former does...
Juan Jose Garcia Ripoll [Thu, 21 Jan 2010 13:31:56 +0000 (14:31 +0100)]
ecl_init_env() split into ecl_init_env and ecl_cs_set_org() so that the former does not need to be executed in the same thread where it is going to be used.

15 years agoNew macro _hash_eq
Juan Jose Garcia Ripoll [Wed, 20 Jan 2010 21:55:53 +0000 (22:55 +0100)]
New macro _hash_eq

15 years agoReorganize the code for variable binding for better caching.
Juan Jose Garcia Ripoll [Wed, 20 Jan 2010 21:55:14 +0000 (22:55 +0100)]
Reorganize the code for variable binding for better caching.