Bearnip Lars LDMud LDMud 3.3 (development)

LDMud 3.3 (development)

Version: 3.3.715 Release Date: 30. Sep 2007 Maintained by: Lars "Mateese" Düning E-mail:
Summary and detailed list of recent changes. Full detailed history of changes since 3.2, including links to early diffs. Download - the source for ldmud-3.3.715.tar.gz (2.5 MB) - svn://svn.bearnip.com/ldmud/version/3.3.715

Everybody else, have fun!

About 3.3.715 (changelog)

New efun tls_refresh_certs() to reload the certificates and certificate revocation information.

New commandline options to handle certification revocations: --tls-crlfile pathname and --tls-crldirectory pathname.

About 3.3.714

Implemented the array/mapping intersection.

Modified a case of undefined behaviour in the creation of the character classification tables during the build. This undefined behavior caused problems with newer GCC versions.

If the driver runs low on memory, it also prints the current object and program in the low-on-memory message.

New efuns hash() and hmac() offer flexible hash and HMAC calculation if the driver is compiled to use OpenSSL (provided by Philipp).

New macros __OPENSSL__ and __GNUTLS__ are set depending on which TLS package is being used.

Corrected a bug with extremely long object names (Bug #466).

Added support for GEN_EMAIL, GEN_URI and GEN_DNS certificates. tls_verify_callback() now prints diagnostics only if the '-d' option was given (thanks, Philipp).

lfun closures now keep track to which program they belong, to avoid ambiguities in case of complex inheritance relationships. This also affects the savefile format: lfun closures are saved with their inheritance path (Bug #443, #444).

If a non-virtual baseclass was inherited multiple times with a virtual class inside the inheritance tree, its members were merged even though they shouldn't (Bug #422).

About 3.3.713

Master function disconnect() receives the remaining input data is passed as the second argument (suggested by Philipp).

The TLS implementation is more robust.

A missing return is noted at runtime only once per function. This is achieved by runtime code modification, but we do that for switch() as well (Bug #430).

Added support for: International Domain Names, SQLite (Bug #431).

Virtual blueprint objects are guaranteed to have their 'clone' flag unset.

Removed the stipulation that CATCH_RESERVED_COST must be twice as much as MASTER_RESERVED_COST, since the master reserve is used automatically anyway.

sscanf(): The new modifier '+' requires the matching of the characters after a field in order to consider the field as matched.

remove_interactive() now preserves the current object (Bug #431).

Corrected the tracedepth handling in efun closures and non-toplevel callbacks (Bug #441).

Private lfuns no longer hide simul-efuns.

Private lfuns can be hidden by public functions without warning, ie. the former are now really invisible to inheritors. This also corrects a crasher with multi-level virtual inherits.

Multiple virtual inheritance of the same program with changing visibility is now handled with a better calculation of the final visibilities (Bug #362).

Calling an unbound lambda cleaned the control stack twice, causing errors (Bug #374).

Only public simul-efuns can be called. This was always valid for sefuns outside the 256-range, but now also applies to normal sefuns as well (Bug #419).

For completeness sake: removed the compilation errors happening when disabling the new inline closures (Bug #423).

present(): Modified the 'n' search behaviour so that the numbering is applied over both inventory and environment together. Before, the numbering was individual in each space, leading to situations where low-numbered objects in the environment were hidden by those in the inventory (Bug #417).

(object.c) Efun function_exists(): The 'FEXISTS_ALL' result also contains information about the function's type, flags, and number of arguments (Bug #410).

Efuns md5()/sha1(): Accept the number of iterations as optional argument (Bug #418).

About 3.3.712

Efun md5() wasn't 64-bit clean (Bug #413).

foreach(): When looping over a mapping with no values, or with only the index being extracted, the loop actually loops over the mapping and not just the indices copied into an array (Bug #388).

Efun set_environment(): clear the RESET_STATE flag of of the moved object's original environment, if any (Bug #387).

Efun md5()/sha1() can now hash arrays of numbers as well (Bug #411).

The string_copy() function checks if the requested memory is actually allocated, throwing an error if not (Bug #397).

Errors during the processing of a fatal() will be recognized and treated as double-fatal (Bug #408).

free_svalue() now recognizes the free of nested structures and de-serializes them to avoid stack overflows (Bug #416).

When re-using a sentence structure for marker_sent, the shadow_ob member wasn't cleared out, causing crashes down the road (Bug #414).

Efun bind_lambda() doesn't ignore unbindable closures when the second argument is missing (Bug #407).

Fixed object reference leaks (Bug #412).

(pkg-tls.c) When accepting a SSL connection, request a client certificate. The driver itself won't check the correctness, but the result will be available in the connection status.

About 3.3.711

Implemented a low/high watermark logic for the backend call to cleanup all objects, in order to prevent the driver from hogging the CPU.

Updates to the OpenSSL interface.

The test for struct equivalence now recurses through member structs; otherwise changes in an contained struct wouldn't be reflected in the containing struct, leading to LPC compilation errors.

to_int() no longer accepts "x<num>" as valid sedecimal notation (Bug #0000402).

Errors during the backend cleanup will now find a consistent virtual machine (Bugs #0000405 and #0000406).

About 3.3.710

Moved the repository from PRCS to Subversion.

About 3.3.709

Oops, I accidentally broke compilation on systems which allow the replacement of the system malloc() function. It should now work again (Bug #0000393).

About 3.3.706-.708

Added new efun pg_conv_string(), courtesy of Philipp.

Added new flag setting '4' to get_type_info() for closures to return the function name of lfun closures (Bug #0000392).

Possibly corrected a refcounting bug in handling prepositions which caused crashers (Bug #0000391).

Since the operators #'&, #'|, #'^ and #'! are now considered efuns, the closure compiler skipped the special treatment of these. As an added bonus of the correction, #'| and #'^ can now take an arbitrary number of arguments, too. (Bug #0000389).

If the number of destructed pending objects exceeds the number of active objects, a cleanup of all objects is triggered. Also, the incremental cleanup in each cycle cleans at least the number of objects destructed in the cycle.

About 3.3.705

runtime_error(), runtime_warning() and heart_beat_error() now receive a flag telling if the error is caught or not (Bug #0000385).

About 3.3.704

Restoring a saved array with one-character operator closures like #'- works again (Bug #0000384).

About 3.3.703

sprintf("%c", 0) returned an empty string "" instead of a string with the 0 character "\0" because internally the system's sprintf() was used for the formatting (Bug #0000378).

About 3.3.702

When printing structs with sprintf("%O"), the output now also contains the names of the structure members (Bug #0000380).

About 3.3.701

The iconv delivered as part of glibc 2.2.5 can get stuck in an endless loop. The configuration script now disables iconv support for glibc 2.0-2.2 (Bug #0000377).

About 3.3.700

There was a problem in the LPC compiler which lead to memory corruption whenever two strings were added in a preprocessor condition. Often enough the corruption was harmless, but sometimes it bit.

About 3.3.699

I was hoping to include fixes for two crashers recently discovered, but no such luck: one is caused by an incompatibility between FreeBSD and ptmalloc which will take some time to nail down, the other one occured only once so far and didn't leave much clues.

When using TLS, the new configuration parameter '--with-tls-path' can be used to specify the installation location for the TLS package.

Corrected a crasher when the TLS certfile locations are printed.

The config macro LOG_NEW_CLEANUP in config.h can be used to enable the logging of the new cleanup calls. Use this if the cleanup seems to hog the CPU to get an idea of how often it is called.

About 3.3.696-.698

Ralph provided a couple of adjustments for FreeBSD 5.4 .

About 3.3.695

Just a little tweaking to the version number determination.

About 3.3.694

Aaaand I had to reintroduce the numeric limit on the number of objects to data-clean per backend cycle, as even with the random factor in the timing sometimes hundreds of objects could become eligible at once.

About 3.3.693

I found the probably reason for the lags experienced by some muds in flaws in the data-cleanup handling (whereas the tunables in the allocators made no difference in my testing):

The data cleanup didn't check whether it still had time to spend or not - as a result the driver could cleanup thousands of objects in a backend cycle even if it had time for only one.

I underestimated the amount of data some muds hold in the default wizlist entry and let the driver clean it up with every object. A mistake, and now the driver applies the cleanup-time timing to the fixed structures as well.

About 3.3.692

input_to() and call_out() no longer accept local references as extra function arguments.

About 3.3.691

Since some muds (but not all) report occasional high CPU loads and lag, I have added a few statistics in order to locate the problem:

• The data-cleanup logs which object is cleant up and how long it took.
• The slaballoc extstats provide an overall and a weighted average of how many search steps per free operation are executed.
• status tables Now provides a weighted average of the VM ticks and time per execution (command, callout, heartbeat).

If you suspect the ordering of free blocks to be the culprit, try undefining the macros MALLOC_ORDER_LARGE_FREELISTS and MALLOC_ORDER_SLAB_FREELISTS in config.h. If the lags get milder by this, but don't go away completely, the root cause is somewhere else.

In either case, if you encounter such situations, please gather information from status tables, status malloc, and status malloc extstats, and send it to me.

About 3.3.690

Made the last-data-cleaned statistics more useful.

About 3.3.689

The synch heartbeat patch wasn't quite correct: one condition was still reversed, and I had left some debug code in. ...it had been a long day.

About 3.3.688

The synchronous heartbeats didn't initialise properly, and thus never executed.

About 3.3.687

Gnomi provided a patch improving the configurability of the call_out() and heartbeat timing at driver compile time; I added some touches to provide commandline options as well:

The basic timing granularity and the heartbeat interval time can be configured at compile time and with the commandline options --alarm-time and --heart-interval. The values are available in LPC with the macros __ALARM_TIME__ and __HEART_BEAT_INTERVAL__.

Heartbeats can be executed synchronously (all at the same time as far as granularity allows, regardless of when inside one hb interval they have executed last). The commandline options are --sync-heart and --async-heart; if enabled, the macro __SYNCHRONOUS_HEART_BEAT__ is defined.

About 3.3.686

Efun tell_object() can now take a non-string as second argument, and will call lfun catch_msg() in the receiver. This functionality is not exactly essential (as it boils down to a call_other()), but brings tell_object() more in line with tell_room() and say() (Bug #0000354).

Efuns tell_room() and say() now accept mappings, structs, and objects as well as arrays for the catch_msg() mechanism.

About 3.3.685

The statistics for hybrid and hash mappings were off.

The data-cleanup is now controlled solely by the configure time interval. Previous drivers imposed an additional limitation on the number of objects data-cleaned in one backend cycle, which as FinalFrontier demonstrates is too restrictive.

About 3.3.684

Finetuned the slab allocator for better use of the slabs.

Corrected a crasher in the error handling for uncallable closures (Bug #0000374).

Corrected the snoop() manpage (Bug #0000372).

About 3.3.683

regreplace() put the stack into an inconsistent state, causing crashes further down the road.

About 3.3.682

The mapping indexing with strings was not correct: under certain circumstances existing entries would not be found.

Corrected various memory leaks and missing error checks when using unique_array() and regreplace() with faulty closures (Bug #0000370).

About 3.3.681

I reduced the overhead used by the string allocator; in particular strings no longer need two memory blocks and their hidden overhead.

About 3.3.678

Corrected a fatal bug in smalloc/slaballoc which appeared on machines where sbrk() could be used (Bug #0000368).

About 3.3.677

It is now possible to set the maximum number of mapping keys (entries) separately from the number of values held. The runtime limits parameter is LIMIT_MAPPING_KEYS (the former LIMIT_MAPPING is an alias hereto), and LIMIT_MAPPING_SIZE respectively. The configuration and commandline parameters changed accordingly (Bug #0000360).

About 3.3.676

Corrected a crasher in smalloc.

Newly allocated core memory is coagulated with existing free memory where possible (Bug #0000364).

About 3.3.675

Corrected a compilation error when GnuTLS was used.

Removing an interim pending input_to should no longer mess up the telnet state (I hope!) (Bug #0000363).

About 3.3.674

During a (re)compilation of the master object, simul-efuns are disabled to avoid accidental use.

About 3.3.673

The permanent block statistics for slaballoc/smalloc was wrong.

Detailed changes since 3.2

30-Sep-2007 (Lars Duening) 3.3.715

• Created snapshot.

04-Nov-2006 (Lars Duening, Philip Hancke)

• (pkg-tls, main.c) Added new options --tls-crlfile and --tls-crldirectory to hold the Certificate Revocation information.
• (pkg-tls, func_spec) New efun tls_refresh_certs() to reload the certificates and certificate revocations.

09-Jul-2006 (Lars Duening) 3.3.714  (diff to 3.3.713: 33 KB: http)

• Created snapshot.

14-Jun-2006 (Lars Duening)

• (array.c) Removed a memory leak when array intersection or subtraction failed.
• (mapping, array, interpret.c, prolang.y) Implemented the array/mapping intersection.

12-Jun-2006 (Lars Duening)

• Oh dear, has it been that long?
• (make_func.y) Modified a case of undefined behaviour in the creation of the character classification tables: only 'unsigned char' types have a defined overflow behaviour (thanks, Erik!)
• (configure) -fno-force-mem is deprecated.

09-May-2006 (Lars Duening)

• (xalloc.c) If the driver runs low on memory, it also prints the current object and program in the low-on-memory message.

06-May-2006 (Lars Duening)

• (pkg-tls) New efuns hash() and hmac() offer flexible hash and HMAC calculation if the driver is compiled to use OpenSSL (provided by Philipp).
• (pkg-tls.c) Removed a compilation error occuring with GnuTLS.
• (lex.c) New macros __OPENSSL__ and __GNUTLS__ are set depending on which TLS package is being used.

06-May-2006 (Gnomi)

• (prolang.y) Corrected the linenumber counting for inline closures.

23-Apr-2006 (Lars Duening)

• (mstring) The max hash length used is now available as macro constant.
• (otable.c) Both mstrings and C-strings are hashed with the same max length. This corrects a bug with extremely long object names (Bug #466).
• (structs.c) The calculation of the first part of a struct's hash uses what the mstring code has already calculated.

30-Mar-2006 (Lars Duening)

• (prolang.y) Removed a dead branch.
• (pkg-tls.c) Removed a possible compiler warning (Bug #464).
• (prolang.y) Removed a bison warning when structs aren't used.

27-Mar-2006 (Lars Duening)

• (pkg-tls.c) Added support for GEN_EMAIL, GEN_URI and GEN_DNS certificates. tls_verify_callback() now prints diagnostics only if the '-d' option was given (thanks, Philipp).

18-Mar-2006 (Gnomi)

• (interpret.c) In apply cache entries for non-existing functions, the garbage collector counted the reference to the string name twice, causing the use of deallocated memory (Bug #463).

16-Mar-2006 (Lars Duening)

• (comm.c) Moved the declaration of a variable which was declared in the middle of a block - gcc doesn't complain, but it's not ISO-C (Bug #462).

14-Mar-2006 (Gnomi)

• (closure, object.c, interpret, exec.h, gcollect.c) lfun closures now keep track to which program they belong, to avoid ambiguities in case of complex inheritance relationships. This also affects the savefile format: lfun closures are saved with their inheritance path (Bug #443, #444).
• (prolang.y) If a non-virtual baseclass was inherited multiple times with a virtual class inside the inheritance tree, its members were merged even though they shouldn't (Bug #422).
• (prolang.y) The LPC compiler used a global variable for typetracking, which caused errors when types changed inside a variable initialisation (Bug #394).

12-Mar-2006 (Lars Duening) 3.3.713  (diff to 3.3.712: 77 KB: http)

• Created snapshot.
• (comm.c) Master function disconnect() receives the remaining input data is passed as the second argument (suggested by Philipp).

08-Mar-2006 (Malcom Tester)

• (lp-245/obj/master.c) Made the move implementation more robust.

07-Mar-2006 (Lars Duening)

• (pkg-tls.c) Two modifications to increase reconnect stability (thanks, Philipp!).
• (prolang.y) Extended a LPC compiler warning when function indices are adjusted.

06-Mar-2006 (Lars Duening)

• (backend, comm.c, object) The logon(obj) function moved from backend.c into object.c, and was renamed logon_object() - it's more sensical this way.
• (comm.c) If the master::connect() initiates a secure connection which gets stuck in the handshake, neither logon() is called nor is the prompt printed. Also, if the tls_init_connection() didn't set a callback, the driver will provide a callback to logon(). This way logon() won't be called unless the connection is up and running.
• (mudlib/psyc-tls.c) Example for using the tls_check_certificate() efun (thanks, Philipp).

03-Mar-2006 (Lars Duening)

• (pkg-tls.c) tls_check_certificate() does more interpretation of the result and returns better data (provided by Philipp).

01-Mar-2006 (Lars Duening)

• (interpret.c) A missing return is noted at runtime only once per function. This is achieved by runtime code modification, but we do that for switch() as well (Bug #430).

28-Feb-2006 (Lars Duening)

• (pkg-idna) Added support for International Domain Names (provided by Philipp - thanks!).
• (configure, Makefile) Added a target 'install-headers' to install all driver LPC header files.
• (simulate.c) Virtual blueprint objects are guaranteed to have their 'clone' flag unset.
• (string_spec, simulate.c) Renamed COMP_OBJ to the more intuitive COMPILE_OBJECT.
• (interpret.c) Removed the stipulation that CATCH_RESERVED_COST must be twice as much as MASTER_RESERVED_COST, since the master reserve is used automatically anyway (suggested by Gnomi).
• (efuns.c) sscanf(): The new modifier '+' requires the matching of the characters after a field in order to consider the field as matched.

20-Feb-2006 (Lars Duening)

• (Makefile.in) Added a dependency to create patchlevel.h, even if the program is made with 'make ldmud'.

20-Feb-2006 (Gnomi)

• (comm.c) remove_interactive() clobbered the current_object, setting it to NULL. It is now preserved (Bug #431).
• (interpret.c) When executing an efun closure, the tracedepth wasn't incremented accordingly (Bug #441).
• (simulate.c) When executing a non-toplevel callback, the tracedepth should not be zeroed out (Bug #441).

08-Jan-2006 (Lars Duening)

• (make_func.y) When parsing the func_spec file, the preprocessor conditionals were handled in a buffer in a no-longer existing stack frame. So far, we were lucky, but on a 64-bit machine that space was needed by the Bison parser.
• (driver.h, smalloc.c, slaballoc.c) If the machine uses a 8-Byte malloc alignment (e.g. all PowerPC machines), the allocator does not replace the system malloc().
• (mud/lp-245/) Since the recent change to how simul_efuns are called (see Bug #419), the LP-245 lib was broken. This has been corrected.

05-Dec-2005 (Lars Duening)

• (configure, machine.h.in, config.h.in) Added a configuration check for SQLite3.
• (pkg-sqlite, many files) Added SQLite support (Bug #11).

04-Dec-2005 (Gnomi)

• (prolang.y) Private lfuns no longer hide simul-efuns.
• (prolang.y) Private lfuns can be hidden by public functions without warning, ie. the former are now really invisible to inheritors. This also corrects a crasher with multi-level virtual inherits.
• (prolang.y) Multiple virtual inheritance of the same program with changing visibility is now handled with a better calculation of the final visibilities. (Bug #362)

04-Dec-2005 (Lars Duening)

• (interpret.c) Calling an unbound lambda cleaned the control stack twice, causing errors (Bug #374).
• (simul_efun.c) Only public simul-efuns can be called. This was always valid for sefuns outside the 256-range, but now also applies to normal sefuns as well (Bug #419)
• (simul_efun, prolang.y, closure.c, interpret.c) The number of direct callable simul-efuns is now configurable in the simul_efun.h. I also increased it to 65536 from 256.
• (closure.c) Simul-efuns were still called with call_other instead of call_direct.

02-Dec-2005 (Lars Duening)

• (gcollect.c) Added a log output when all objects are data-cleaned, as it might be the cause for some observed slowdowns.
• (closure.c, interpret.c) For completeness sake: removed the compilation errors happening when disabling the new inline closures (Bug #423).

27-Nov-2005 (Lars Duening)

• (object.c) Efun present(): Modified the search behaviour so that the numbering is applied over both inventory and environment together. Before, the numbering was individual in each space, leading to situations where low-numbered objects in the environment were hidden by those in the inventory (Bug #417).
• (object.c) Efun function_exists(): The 'FEXISTS_ALL' result also contains information about the function's type, flags, and number of arguments (Bug #410).

26-Nov-2005 (Lars Duening)

• (efuns.c) Efun sscanf(): Added test whether the constant characters after a %-field also match - if not, the %-field is considered unmatched (suggested by Largo).
• (efuns, func_spec) Efuns md5()/sha1(): Accept the number of iterations as optional argument (Bug #418).

24-Nov-2005 (Lars Duening) 3.3.712  (diff to 3.3.711: 9 KB: http)

• Created snapshot.
• (md5.c) The ROTATE_LEFT macro wasn't 64-bit clean (Bug #413).
• (interpret.c) Using foreach(), when looping over a mapping with no values, or with only the index being extracted, the loop actually loops over the mapping and not just the indices copied into an array (Bug #388).
• (object.c) Efun set_environment(): clear the RESET_STATE flag of of the moved object's original environment, if any (Bug #387).
• (efuns.c) Efun md5()/sha1() can now hash arrays of numbers as well (Bug #411).
• (xalloc.c) The string_copy() function checks if the requested memory is actually allocated, throwing an error if not (Bug #397).
• (simulate.c) Errors during the processing of a fatal() will be recognized and treated as double-fatal (Bug #408).
• (interpret.c) free_svalue() now recognizes the free of nested structures and de-serializes them to avoid stack overflows (Bug #416).
• (Makefile.in, version.sh, patchlevel.h.in) Removed PRCS support.

24-Nov-2005 (Gnomi)

• (actions.c) When re-using a sentence structure for marker_sent, the shadow_ob member wasn't cleared out, causing crashes down the road (Bug #0000414).
• (closure.c) bind_lambda() doesn't ignore unbindable closures when the second argument is missing (Bug #407).

09-Oct-2005 (Lars Duening)

• (closure.c) replace_program_lambda_adjust() forgot to clear the object references from the auxiliary lambda it created (Bug #0000412).

08-Oct-2005 (Lars Duening)

• (closure.c) When freeing bound unbound_lambdas, the code forgot to remove the reference to the creator object of the unbound_lambda.
• (closure.c) Efun symbol_function() leaked an object reference (Bug #0000412).
• (closure.c, object.h) Made the object reference logs a bit more expressive.

31-Oct-2005 (Philipp Hancke)

• (pkg-tls.c) When accepting a SSL connection, request a client certificate. The driver itself won't check the correctness, but the result will be available in the connection status.

30-Oct-2005 (Lars Duening) 3.3.711  (diff to 3.3.710: 31 KB: http)

• Created snapshot.

29-Oct-2005 (Lars Duening)

• (backend.c) The backend call to cleanup_all_objects() is now controlled with a low/high watermark logic, to prevent repeated calls in consecutive backend cycles: if one cleanup didn't remove enough destructed objects, it is unlikely that immediately following ones will.
• (pkg-tls.c) The OpenSSL read/writes are now retried when necessary.
• (configure) When searching for the mysqlclient library in the usual places, the configure script forgot to clear out the cache flag.
• (smalloc.c) Removed a compilation error when SBRK_OK wasn't defined.
• (mstrings, many files) The find-in-table statistics were done incompletely.
• (structs.c) The test for struct equivalence now recurses through member structs; otherwise changes in an contained struct wouldn't be reflected in the containing struct.
• (prolang.y) Clarified the 'Unknown type' error messages when derefencing structs.

19-Oct-2005 (Lars Duening)

• (pkg-tls.c) tls_check_certificate() now returns the low-level API values, and throws an error if the connection is not secure.

11-Oct-2005 (Lars Duening)

• (erq/erq.c, xerq/socket.c) Corrected compilation warnings.
• (lex.c) to_int() no longer accepts "x<num>" as valid sedecimal notation (Bug #0000402).

10-Oct-2005 (Lars Duening)

• (backend.c) Added a clear_state() at the beginning of the process_objects() loop, so that errors during the cleanup find a consistent machine state (Bug #0000405).

09-Oct-2005 (Lars Duening)

situation, the catch() is ignored. This also fixes an endless recursion (Bug #0000406).

04-Oct-2005 (Lars Duening)

• (doc/concepts/overloading) New man page (thanks, Coogan!).

28-Sep-2005 (Lars Duening)

• (doc/efun/clone_object) Clarified the variable initialisation.

23-Aug-2005 (Lars Duening)

• (Makefile.in) The PRCS rules first check if PRCS is available.

20-Jun-2005 (Lars Duening) 3.3.710  (diff to 3.3.709: 3 KB: http)

• (patchlevel.h.in, version.sh, mk-patchlevel.sh, Makefiles) Modified the version numbering system to be compatible between Subversion and PRCS.

20-Jun-2005 (Lars Duening) 3.3.709  (diff to 3.3.708: 2 KB: http)

• (xalloc.c) xalloc used REPLACE_MALLOC before it was defined, leading to compilation errors on systems supporting it (Bug #0000393).

19-Jun-2005 (Lars Duening) 3.3.708  (diff to 3.3.705: 16 KB: http)

• (pkg-pgsql, func_spec) Added new efun pg_conv_string(), courtesy of Philipp.
• (efuns.c) Added new flag setting '4' to get_type_info() for closures (Bug #0000392).
• (parse.c) Possibly corrected a refcounting bug in handling prepositions (Bug #0000391).
• (closure.c) Since the operators #'&, #'|, #'^ and #'! are now considered efuns, the closure compiler skipped the special treatment of these. As an added bonus of the correction, #'| and #'^ can now take an arbitrary number of arguments, too. (Bug #0000389)

10-Jun-2005 (Lars Duening) (3.3.707)

• (gcollect, backend.c) If the number of destructed pending objects exceeds the number of active objects, a cleanup of all objects is triggered. Also, the incremental cleanup in each cycle cleans at least the number of objects destructed in the cycle.

06-Jun-2005 (Lars Duening) (3.3.706)

• (object.c, sprintf.c, structs.c) Removed a couple compilation warnings on x86-64.

15-May-2005 (Lars Duening) 3.3.705  (diff to 3.3.687: 104 KB: http)  (diff to 3.3.704: 4 KB: http)

• (simulate.c) runtime_error(), runtime_warning() and heart_beat_error() received a new argument 'caught' (Bug #0000385).

15-May-2005 (Lars Duening) 3.3.704  (diff to 3.3.703: 2 KB: http)

• (object.c) Single-character operator closures like #'- couldn't be restored in arrays due to some outdated hack in the parser (Bug #0000384 and #0000386).

15-May-2005 (Lars Duening) 3.3.703  (diff to 3.3.702: 2 KB: http)

- (sprintf) sprintf("%c", 0) returned an empty string instead of a string with the 0 character, because the code used the system sprintf() for this formatting internally (Bug #0000378).

14-May-2005 (Lars Duening) 3.3.702  (diff to 3.3.701: 3 KB: http)

• (sprintf) Printing structures with "%O" or "%Q" in non-compact mode also prints the names of the structure members (Bug #0000380).

13-May-2005 (Lars Duening) 3.3.701  (diff to 3.3.700: 2 KB: http)

• (configure) Added a check to see if the system is using a broken iconv (part of glibc 2.0-2.2) - this iconv can throw the driver into an endless loop (Bug #0000377).

13-May-2005 (Lars Duening) 3.3.700  (diff to 3.3.699: 2 KB: http)

• (lex.c) When adding strings in preprocessor conditionals, the lhs string was freed twice, causing memory corruption (reported by Tobias).

24-Apr-2005 (Lars Duening) 3.3.699  (diff to 3.3.698: 6 KB: http)

• (config.h.in, gcollect.c) Added a specific define LOG_NEW_CLEANUP to enable the logging of the automatic cleanup activity. This is useful if the cleanup is suspicious of hogging the CPU.
• (main.c, pkg-tls.c) Improved the documentation for the TLS commandline options, removed a crasher when only one of the two cert locations is specified (thanks, Coogan!).
• (xalloc.c) Added warning if ptmalloc is used under FreeBSD.
• (configure) New configuration parameter '--with-tls-path' allows to specifiy a non-standard location for the TLS installation.

01-Apr-2005 (Lars Duening) 3.3.698  (diff to 3.3.695: 41 KB: http)

• (pkg-tls.c) Added FreeBSD 5.4 compatibility (thanks, Ralph!).
• (simulate, many files) Renamed error() to errorf() due to a collision with gnutls.h include files under newer compilers.

01-Apr-2005 (Lars Duening) (3.3.696, .697)

• (port.c) Corrected a C99 incompatibility (thanks, Ralph!).
• (simulate.c, backend) The commands/s and compiled lines/s statistics are part of the 'status tables' output.
• (gcollect.c) Added a missing variable initialisation (thanks, Ralph!).

27-Mar-2005 (Lars Duening) 3.3.695  (diff to 3.3.694: 3 KB: http)

• (patchlevel.h, main.c, lex.c) Simplified the version number handling in anticipation of the 3.4 branch.

27-Mar-2005 (Lars Duening) 3.3.694  (diff to 3.3.693: 4 KB: http)

• (backend.c) I reintroduced the numeric limit on the number of objects to data-clean, as even with the random factor it could happen that the driver tries to cleanup hundreds of objects in one cycle. The numeric limit is calculated such that all listed objects can be cleaned in half of the configured cleanup cycle.

27-Mar-2005 (Lars Duening) 3.3.693  (diff to 3.3.692: 9 KB: http)

• (slaballoc.c) Corrected the free-steps statistics.
• (backend.c) The data-cleanup didn't pay attention to the backend timing, occasionally causing the driver to clean up thousands of objects in one cycle.
• (gcollect.c) The cleanup of the fixed datastructures (wizlist, driver hooks) did not observe the overall cleanup timing - on muds with lots of data in the default wizlist entry this lead to annoying lags.

25-Mar-2005 (Lars Duening) 3.3.692  (diff to 3.3.691: 5 KB: http)

• (simulate.c) Modified the callback setup so that efun callbacks can take normal lvalues; and timed function/closure callbacks take protected lvalues - but not vice versa.

24-Mar-2005 (Lars Duening) 3.3.691  (diff to 3.3.690: 8 KB: http)

• (gcollect.c) Added a log output for the time it takes to data-clean objects.
• (config.h.in, slaballoc.c) The sorting of the partial slab freelists is now configurable by hand.
• (slaballoc.c) Added weighted-average statistics for number of steps per freed block.
• (interpret, backend.c, call_out.c, efuns.c heart_beat.c, simulate.c) Added a statistic (last and time-weighted average) about the total execution ticks and time per LPC thread.

15-Feb-2003 (Lars Duening) 3.3.391  (diff to 3.3.390: 36 Bytes: http)

20-Mar-2005 (Lars Duening) 3.3.690  (diff to 3.3.689: 36 KB: http)

• (backend, main, simulate.c) Updated the last-data-cleaned statistics to be weighted over time; and generalized the statistics implementation in general.

18-Mar-2005 (Lars Duening) 3.3.689  (diff to 3.3.688: 2 KB: http)

• (backend.c) Improved the synch heartbeat condition expression.
• (heartbeat.c) I had forgotten to remove all remnants of an exploratory change.

17-Mar-2005 (Lars Duening) 3.3.688  (diff to 3.3.687: 3 KB: http)

• (backend.c) The synchronous heartbeat timing wasn't initialised properly.

13-Mar-2005 (Lars Duening) 3.3.687  (diff to 3.3.672: 70 KB: http)  (diff to 3.3.686: 11 KB: http)

• (main, backend.c, heartbeat.c) The ALARM_TIME and HEART_BEAT_INTERVAL time are now configurable at compile and start time (based on a patch by Gnomi).

12-Mar-2005 (Lars Duening) 3.3.686  (diff to 3.3.685: 7 KB: http)

• (func_spec, object.c) The efun tell_object() can now take an array, mapping, struct or object as second argument, calling catch_msg() in the target in that case (Bug #0000354).
• (func_spec, object.c) The efuns tell_room() and say() can now take a mapping, object or struct instead of an array as second argument for the catch_msg() mechanism.

09-Mar-2005 (Lars Duening) 3.3.685  (diff to 3.3.684: 4 KB: http)

• (mapping.c) The hash/hybrid statistics were somewhat off.
• (gcollect, backend.c, simulate.c) The default time interval for a data cleanup is now available as macro in gcollect.h . And since we have that time interval, the number-to-process limit in process_objects() didn't make much sense anymore.

06-Mar-2005 (Lars Duening) 3.3.684  (diff to 3.3.683: 14 KB: http)

• (slaballoc.c) Tuned the allocator: the partially-used slab lists are now lazily order for better usage.
• (interpret.c) For an uncallable closure, the code forgot to clean out the CSP before throwing the error (Bug #0000374).
• (doc/efun/snoop) Corrected the description of the return value (Bug #0000372).

26-Feb-2005 (Lars Duening) 3.3.683  (diff to 3.3.682: 2 KB: http)

• (efuns.c) regreplace() didn't update the stack pointers properly, causing crashes down the road (reported by Bardioc).

26-Feb-2005 (Lars Duening) 3.3.682  (diff to 3.3.681: 8 KB: http)

• (mapping.c) Indexing a mapping with strings didn't quite work yet: the hash computation didn't take into effect that strings are no longer tabled just to use them as rvalue indices.
• (simulate.c) If a callback didn't have a valid object associated with it (e.g. unbound_lambda()s), the code forgot to remove the arguments from the stack (Bug #0000370).
• (simulate.c) Uncalleable closures given as callbacks are now detected and cause a runtime error (Bug #0000370).
• (simulate.c) A runtime in a regreplace() substitution closure leaked the regexp structure. I changed the whole thing to automatic cleanup (Bug #0000370).

24-Feb-2005 (Lars Duening) 3.3.681  (diff to 3.3.678: 23 KB: http)

• (mstrings) Re-integrated the stringtable structures into the strings: on the whole, the majority of strings will be tabled so reducing the overhead for this case is beneficial.

24-Feb-2004 (Lars Duening) (3.3.680)

• (mstrings, closure.c, interpret.c) Debugging 3.3.679: reallocating a string 'under the hood' is no longer possible as the string data is now an integral part of the string structure. This disables the mstr_append() optimizations.

21-Feb-2004 (Lars Duening) (3.3.679)

• (mapping.c) get_map_lvalue() didn't correctly check the mapping size: it checked against the current and not the future size (Bug #0000369).
• (mapping, simulate.c) Added a statistic for hash and hybrid mappings (for now just the count).
• (mstrings, gcollect.c, array.c, closure.c, i-svalue_cmp.h, interpret.c, simulate.c) Simplified the string implementation to reduce the overhead.

21-Feb-2005 (Lars Duening) 3.3.678  (diff to 3.3.677: 5 KB: http)

• (slaballoc.c, smalloc.c) I hate crashers which don't happen on my machine: the merging of the esbrk() result with adjacent free blocks crashed the driver (Bug #0000368).

20-Feb-2005 (Lars Duening) 3.3.677  (diff to 3.3.676: 11 KB: http)

• (configure, mapping.c, simulate) It is now possible to determine the max number of mapping entries (keys) in addition to the total mapping size (Bug #0000360).
• (main.c) New commandline option '--max-mapping-keys' to determine separately the maximum number of entries in a mapping.

20-Feb-2005 (Lars Duening) 3.3.676  (diff to 3.3.675: 2 KB: http)

• (smalloc.c) Corrected a crasher in realloc().
• (smalloc.c, slaballoc.c) New memory allocated with esbrk() is now coagulated with existing free memory if possible (Bug #0000364).

20-Feb-2005 (Lars Duening) 3.3.675  (diff to 3.3.674: 2 KB: http)

• (pkg-tls.c) Removed compilation errors when using GnuTLS.
• (comm.c) The charmode/noecho mode is changed in remove_input_to() only if the first pending input_to() is removed. Hopefully this fixes Bug #0000363.

20-Feb-2005 (Lars Duening) 3.3.674  (diff to 3.3.673: 4 KB: http)

• (prolang, simulate.c) During the compilation of the master object, simul efuns are disabled.

20-Feb-2005 (Lars Duening) 3.3.673  (diff to 3.3.672: 4 KB: http)

• (smalloc.c, slaballoc.c) The memory block size computation failed on large blocks, giving bad statistics.

The older changelog entries can be read here.

Summary of changes since 3.2

• Changed efuns:
  • Removed obsolete and deprecated efuns: add_verb(), add_xverb(), allocate_mapping(), copy_mapping(), efun308(), extract(), file_name(), filter_array(), filter_mapping(), m_sizeof(), map_array(), map_mapping(), mapping_contains(), member_array(), set_auto_include_string(). transfer().
  • copy(), deep_copy(), get_type_info(), to_array(), to_string(): Added struct support.
  • add_action(): Historical usage with just one argument no longer supported.
  • assoc(), insert_alist(), order_alist(), intersect_alist(): Optional as package 'alists'.
  • call_out(): The minimum delay can now be 0, meaning: as soon as possible.
  • catch(): If there is not enough time left (less than __CATCH_EVAL_COST__), this is thrown as an error from inside the catch instead of causing the catch to return 0.
  • catch(): Modifier reserve expr can be used to specify a computing reserve different from __CATCH_EVAL_COST__.
  • debug_info(): Option DINFO_DUMP:"memory" to dump a list of all allocated memory blocks (currently supported just by smalloc).
  • debug_info(): Added more data to the option DINFO_DATA:DID_MEMORY result.
  • debug_info(): DINFO_DATA:DID_ST_PACKETS_IN and DID_ST_PACKET_SIZE_IN are the statistics for the incoming network traffic.
  • deep_inventory(): The result can be restricted to a certain depth.
  • filter(): The efun can now filter strings as well.
  • function_exists(): now also returns the filename and linenumber of a function.
  • function_exists(): The 'FEXISTS_ALL' result also contains information about the function's type, flags, and number of arguments.
  • garbage_collection(): An optional flag argument allows to change the name of the default GC log file.
  • get_type_info(): The name of the defining program of a lfun or context closure, or the function name itself, can be returned.
  • map(): The efun can now map strings as well.
  • make_shared_string(): Deprecated.
  • member(), rmember(): The start position for the search can be specified.
  • mkmapping(): Convert structs to mappings, too.
  • pg_*(): Postgres efuns trigger a privilege violation ("pgsql", "efun name").
  • process_string(): Spaces are now allowed in function arguments and no longer terminate the function specification.
  • regexp(), regexplode(): Optionally take RE interpretation options.
  • regexplode(): New flag RE_OMIT_DELIM allows to omit the matched delimiters from the result.
  • send_erq(): ERQ callback closures can now accept the data as a string instead of an array.
  • sprintf(), printf(): Formats %O and %Q print floats always with a decimal point.
  • say(), tell_room(), tell_object(): The efuns can now take a mapping, an object or a struct instead of an array to use with the catch_msg() mechanism.
  • transfer(): Available (in compat mode) only when USE_DEPRECATED is in effect.
  • unique_array() The separator function can now be a closure, and can also be given extra arguments.
  • wizlist_info(): Added the number of mappings to the result.
  • write_file() Allowed to remove the file before writing.
• Corrected efuns:
  • restore_object(): An exception during the restore no longer has the danger of leaking values and memory.
  • unique_array(): An exception during the process no longer has the danger of leaking values and memory.
• New efuns:
  • baseof(): Test the relation between two structs.
  • call_direct(), call_direct_resolved(): To perform calls which are not resolved by a default method.
  • convert_charset(): Convert a string between character sets.
  • match_command(): Find all matching actions for a command.
  • md5_crypt(): Compute an Apache-compatible password encryption using the MD5 message digest algorithm.
  • pg_connect(), pg_query(), pg_pending(), pg_close(): Optional: PostgreSQL support.
  • regmatch(): To match a string against a regexp and optionally retrieve the match results for subexpressions.
  • regexp_package(): Returns the currently selected default regexp package.
  • reverse(): Reverse the content of an integer, a string or an array.
  • sha1(): Compute the SHA1 message digest.
  • struct_info(): Return information about a struct.
  • structp(): Predicate to detect struct values.
  • tls_refresh_certs(): Reload the certificates and certificate revocation information.
  • to_struct(): Convert an array into a struct.
• Master Object:
  • Function runtime_warning() is called for runtime warnings.
  • privilege_violation(): For "set_limits" and "limited" the requested limits are passed as argument.
  • During the compilation of the master object, simul-efuns are disabled (so they won't be used by accident).
• Fixed Crashers:
  • Removed a crasher when a private lfun shadowing a public sefun was virtually inherited twice in parallel.
  • Removed a crasher when a walk_mapping() over a mapping deleted elements from the very same mapping.
• Language:
  • New operator ... to flatten array arguments in a function call at runtime.
  • Strings can now hold all characters, including '\0'.
  • foreach() can loop over references, integers and integer ranges.
  • nosave and catch(; nolog) are now official features.
  • New indexing method [>] to index either from the beginning or the end, depending on the sign of the index value.
  • New pragma range_check allows to check array range indexing at runtime.
  • If switch() is used on a value which is neither a string nor a number, the driver no longer generates a runtime error messge, but instead jumps to the default label.
  • Operators | and |= can compute the union of two arrays.
  • Operators ^ and ^= can compute the symmetric difference of two arrays.
  • Closure operators #'| and #'^ can take more than two arguments.
  • Support for structs; preprocessor symbol __LPC_STRUCTS__ is defined.
  • Preprocessor symbol __PCRE__ is always defined.
  • Real inline closures; preprocessor symbol __LPC_INLINE_CLOSURES__ is defined.
  • New macro __BOOT_TIME__ expands to the time() the driver was started.
  • Blueprint variables are now always initialized using __INIT(). Clone variables are usually initialized the same way, however with the pragmas share_variables and init_variables the programmer can instruct the driver on a pre-program basis to initialize the variables of clone from the current values of the corresponding variables in the blueprint.
  • Bugfix: The type recognition of the variables for a foreach() no longer gets confused by visibility flags.
• Runtime:
  • Revamped the string handling to use refcounted strings everywhere.
  • If input is requested in CHARMODE, the NL characters are passed verbatim to the caller (old drivers passed an empty string instead).
  • The driver hook H_DEFAULT_METHOD can be used to resolve call_other()/call_resolved()s to non-existing functions.
  • The driver hook H_DEFAULT_PROMPT can be used to specify a default prompt different from "> ".
  • The driver hook H_PRINT_PROMPT can be used to do special magic when printing the prompt.
  • The driver regularly checks the variables of all objects for references to destructed objects and removes them; it is therefore no longer necessary to "clean" mappings by hand.
  • The driver can now warn about missing return statements for value returning functions (this applies only for functions with a specified return type). If possible the driver tries to detect this situation at compile time, but also inserts a special 'return' bytecode to catch certain pathological cases at runtime. The check is controlled by the pragma '(no_)warn_missing_return'.
  • New runtime 'limit' LIMIT_COST to control how much a limited execution should cost.
  • The basic timing granularity and the heartbeat interval time can be configured at compile time and with the commandline options --alarm-time and --heart-interval. The values are available in LPC with the macros __ALARM_TIME__ and __HEART_BEAT_INTERVAL__.
  • Heartbeats can be executed synchronously (all at the same time as far as granularity allows, regardless of when inside one hb interval they have executed last). The commandline options are --sync-heart and --async-heart; if enabled, the macro __SYNCHRONOUS_HEART_BEAT__ is defined.
• Other Changes:
  • The input escape character can be configured at compile time to something different from '!'.
  • The MAX_MAPPING configuration parameter now determines the total number of elements (keys + values) a mapping can hold, not just the maximum number of keys.
  • The driver can be configured to use either the traditional regular expression code, or the PCRE package.
  • New commandline options --share-variables and --init-variables to defined the default for clone variable initialization.
  • New configuration option --enable-malloc-sbrk to selectively allow the memory allocator to use sbrk() instead of malloc() where available (default is to allow sbrk()).
  • New commandline options --tls-key and --tls-cert to define the x509 key- and certfile for the TLS package.
  • New commandline options to handle certification revocations: --tls-crlfile and --tls-crldirectory.

Bearnip Lars LDMud LDMud 3.3 (development)	Lars Düning () - PGP Key
Validate this page.