]> mj.ucw.cz Git - libucw.git/log
libucw.git
18 years agoDumpers don't return anything, especially not a void pointer :-)
Martin Mares [Wed, 26 Apr 2006 09:47:44 +0000 (11:47 +0200)]
Dumpers don't return anything, especially not a void pointer :-)

18 years agoIf CONFIG_DEBUG is enabled, all programs support a --dumpconfig switch,
Martin Mares [Wed, 26 Apr 2006 09:38:39 +0000 (11:38 +0200)]
If CONFIG_DEBUG is enabled, all programs support a --dumpconfig switch,
which dumps the current configuration.

18 years agoconf2: lookup tables are byte** instead of char**
Robert Spalek [Wed, 26 Apr 2006 09:09:17 +0000 (11:09 +0200)]
conf2: lookup tables are byte** instead of char**

18 years agoconf2: include <getopt.h> again, it's tedious to write it everywhere
Robert Spalek [Tue, 25 Apr 2006 22:23:49 +0000 (00:23 +0200)]
conf2: include <getopt.h> again, it's tedious to write it everywhere

Polish a comment too.

18 years agoCF_LOOKUP lists must end with NULL.
Martin Mares [Tue, 25 Apr 2006 22:06:50 +0000 (00:06 +0200)]
CF_LOOKUP lists must end with NULL.

18 years agoipaccess filters now use CF_USER for (addr,mask) pairs, operations
Martin Mares [Tue, 25 Apr 2006 22:03:13 +0000 (00:03 +0200)]
ipaccess filters now use CF_USER for (addr,mask) pairs, operations
on lists now work fine.

18 years agoUse CF_LOOKUP.
Martin Mares [Tue, 25 Apr 2006 21:53:15 +0000 (23:53 +0200)]
Use CF_LOOKUP.

18 years agoMerge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config
Martin Mares [Tue, 25 Apr 2006 21:47:15 +0000 (23:47 +0200)]
Merge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config

18 years agoDon't end error messages with a comma.
Martin Mares [Tue, 25 Apr 2006 21:46:56 +0000 (23:46 +0200)]
Don't end error messages with a comma.

18 years agoMerge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config
Robert Spalek [Tue, 25 Apr 2006 21:40:13 +0000 (23:40 +0200)]
Merge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config

18 years agoMerge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config
Martin Mares [Tue, 25 Apr 2006 21:39:48 +0000 (23:39 +0200)]
Merge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config

18 years agoconf2: dynamic arrays allow negative maximal length
Robert Spalek [Tue, 25 Apr 2006 21:39:03 +0000 (23:39 +0200)]
conf2: dynamic arrays allow negative maximal length

18 years agoMore ipaccess work-arounds (will be gone soon).
Martin Mares [Tue, 25 Apr 2006 21:37:28 +0000 (23:37 +0200)]
More ipaccess work-arounds (will be gone soon).

18 years agoconf: parsing IP address fixed
Robert Spalek [Tue, 25 Apr 2006 21:31:46 +0000 (23:31 +0200)]
conf: parsing IP address fixed

18 years agoMerge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config
Robert Spalek [Tue, 25 Apr 2006 21:29:35 +0000 (23:29 +0200)]
Merge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git#dev-config

18 years agoconf2: add 2 ASSERT's on checking the user-defined type size
Robert Spalek [Tue, 25 Apr 2006 21:28:04 +0000 (23:28 +0200)]
conf2: add 2 ASSERT's on checking the user-defined type size

18 years agoconf2: added support for user-defined types (tested on u16)
Robert Spalek [Tue, 25 Apr 2006 21:25:11 +0000 (23:25 +0200)]
conf2: added support for user-defined types (tested on u16)

18 years agoFixed operator priorities in cf_parse_ip().
Martin Mares [Tue, 25 Apr 2006 21:21:48 +0000 (23:21 +0200)]
Fixed operator priorities in cf_parse_ip().

18 years agoconf2: add basic type CT_LOOKUP that maps strings to integers
Robert Spalek [Tue, 25 Apr 2006 20:25:27 +0000 (22:25 +0200)]
conf2: add basic type CT_LOOKUP that maps strings to integers

Since I need to store both type==CT_LOOKUP and a pointer to the lookup
table, I moved the type out the union.  However, I have not blown up the
size of struct cf_item, because I have compressed class and type to 16
bits.

18 years agoA preliminary version of the new ipaccess module.
Martin Mares [Tue, 25 Apr 2006 12:08:34 +0000 (14:08 +0200)]
A preliminary version of the new ipaccess module.

18 years agoAdded CF_ARY_NUM and CF_JOURNAL_VAR.
Martin Mares [Tue, 25 Apr 2006 10:53:35 +0000 (12:53 +0200)]
Added CF_ARY_NUM and CF_JOURNAL_VAR.

18 years ago- fixed accepted double 0x prefix ("0x0x12345678") and
Pavel Charvat [Tue, 25 Apr 2006 07:54:00 +0000 (09:54 +0200)]
- fixed accepted double 0x prefix ("0x0x12345678") and
  negative numbers ("-0.-0.-0.-0") in IP parser

18 years agoThe "don't cf_load on error exit" optimization was wrong.
Martin Mares [Mon, 24 Apr 2006 21:53:48 +0000 (23:53 +0200)]
The "don't cf_load on error exit" optimization was wrong.

18 years agobopen_safe() in conf2 was leaking file descriptors. I've replaced it
Martin Mares [Mon, 24 Apr 2006 21:49:55 +0000 (23:49 +0200)]
bopen_safe() in conf2 was leaking file descriptors. I've replaced it
by a new bopen_try() in the fastbuf library and also cleaned up error
messages on unsuccessful file opens.

18 years agoStarted moving configuration sections to cf/sherlock2.
Martin Mares [Mon, 24 Apr 2006 21:38:47 +0000 (23:38 +0200)]
Started moving configuration sections to cf/sherlock2.

18 years agoDon't load the default config file when an invalid option error is going
Martin Mares [Mon, 24 Apr 2006 21:32:06 +0000 (23:32 +0200)]
Don't load the default config file when an invalid option error is going
to be reported.

18 years agoconf2: damn, I of course forgot to get rid of a debug message
Robert Spalek [Mon, 24 Apr 2006 18:02:34 +0000 (20:02 +0200)]
conf2: damn, I of course forgot to get rid of a debug message

18 years agoconf2: implemented committing only dirty sections
Robert Spalek [Mon, 24 Apr 2006 17:59:47 +0000 (19:59 +0200)]
conf2: implemented committing only dirty sections

This is now needed very much, because several commit-hooks allocate tens
of kilobytes of memory in the journal at each commit (e.g. alphabet.c)

18 years agoconf2: correct unescaping of C-strings
Robert Spalek [Mon, 24 Apr 2006 13:58:28 +0000 (15:58 +0200)]
conf2: correct unescaping of C-strings

18 years agofixed escape sequences parser
Pavel Charvat [Mon, 24 Apr 2006 13:02:54 +0000 (15:02 +0200)]
fixed escape sequences parser

18 years agoSimple parser for escape sequences.
Pavel Charvat [Mon, 24 Apr 2006 11:56:13 +0000 (13:56 +0200)]
Simple parser for escape sequences.

18 years agoconf2: report a bug when mixing line- and block- syntax improperly
Robert Spalek [Mon, 24 Apr 2006 09:12:44 +0000 (11:12 +0200)]
conf2: report a bug when mixing line- and block- syntax improperly

18 years agoconf2: insert a node into a list as soon as possible instead of at the end
Robert Spalek [Mon, 24 Apr 2006 08:54:15 +0000 (10:54 +0200)]
conf2: insert a node into a list as soon as possible instead of at the end

This allows user-defined parsers to see nodes during processing and they
can easier check for duplicates (see lang/lang.c).

18 years agoPointers to lists should be relative to current section, too.
Martin Mares [Sun, 23 Apr 2006 22:09:25 +0000 (00:09 +0200)]
Pointers to lists should be relative to current section, too.

18 years agoMerge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git
Robert Spalek [Sun, 23 Apr 2006 20:33:45 +0000 (22:33 +0200)]
Merge with git+ssh://cvs.ucw.cz/projects/sherlock/GIT/sherlock.git

18 years agolib: revert unwanted commit
Robert Spalek [Sun, 23 Apr 2006 19:41:18 +0000 (21:41 +0200)]
lib: revert unwanted commit

damn, I tried to just delete one unwanted file from the list when
editing the commit-log, but cogito ignored the deletion and committed the
file anyway.  rollback then.

18 years agolib: converted to use conf2.c instead of conf.c
Robert Spalek [Sun, 23 Apr 2006 19:33:56 +0000 (21:33 +0200)]
lib: converted to use conf2.c instead of conf.c

18 years agoconf2: tested a few more test-cases and fixed/polished the code
Robert Spalek [Sun, 23 Apr 2006 19:13:08 +0000 (21:13 +0200)]
conf2: tested a few more test-cases and fixed/polished the code

- journal blocks are "transactions", don't overload the word "sections" too much
- added 2 long-term TODO's and 1 short-term FIXME
- the include command must be the last one on a line

18 years agoconf2: sorry, fix the commit-hook messages
Robert Spalek [Sun, 23 Apr 2006 13:04:20 +0000 (15:04 +0200)]
conf2: sorry, fix the commit-hook messages

18 years agoconf2: fix commit hooks
Robert Spalek [Sun, 23 Apr 2006 13:00:22 +0000 (15:00 +0200)]
conf2: fix commit hooks

18 years agoconf2: MJ's objections taken care of
Robert Spalek [Sun, 23 Apr 2006 12:44:40 +0000 (14:44 +0200)]
conf2: MJ's objections taken care of

18 years agoconf2: bugfixes in journaling
Robert Spalek [Sun, 23 Apr 2006 12:14:21 +0000 (14:14 +0200)]
conf2: bugfixes in journaling

- commented usage of cf_malloc() vs. xmalloc() in init-hooks
- commented buggy DYN_ALLOC(double) macro
- don't bzero() static sections
- use bopen_safe()

18 years agoconf2: verified and fixed all parsers
Robert Spalek [Sun, 23 Apr 2006 10:51:54 +0000 (12:51 +0200)]
conf2: verified and fixed all parsers

18 years agoconf2: fixed journaling of list operations
Robert Spalek [Sun, 23 Apr 2006 10:24:40 +0000 (12:24 +0200)]
conf2: fixed journaling of list operations

18 years agoconf2: lists seem to be debugged
Robert Spalek [Sun, 23 Apr 2006 10:03:23 +0000 (12:03 +0200)]
conf2: lists seem to be debugged

18 years agoconf2: 2 bugfixes concerning dynamic arrays and lists
Robert Spalek [Sun, 23 Apr 2006 09:41:15 +0000 (11:41 +0200)]
conf2: 2 bugfixes concerning dynamic arrays and lists

18 years agoconf2: debug dumping of memory into a fastbuf partially implemented
Robert Spalek [Sun, 23 Apr 2006 00:56:08 +0000 (02:56 +0200)]
conf2: debug dumping of memory into a fastbuf partially implemented

18 years agoconf2: another bug hunted
Robert Spalek [Sun, 23 Apr 2006 00:24:40 +0000 (02:24 +0200)]
conf2: another bug hunted

18 years agoconf2: removed handling of dirty pages
Robert Spalek [Sun, 23 Apr 2006 00:13:05 +0000 (02:13 +0200)]
conf2: removed handling of dirty pages

2 XXX's and 1 FIXME deleted by just omitting the whole functionality and
commiting _all_ pages after each reload.  it should not cost much more than
a bit of memory.

18 years agoconf2: debugged a lot
Robert Spalek [Sun, 23 Apr 2006 00:09:32 +0000 (02:09 +0200)]
conf2: debugged a lot

- closing_brace() confused current operation with the one on the stack a lot
- get_word() fix when recognizing = in the parameter name
- split_line() totally rewritten to 2 nice procedures: get_token() and
  split_command()
- #include back to include, otherwise it would be a comment
- local fastbufs in parse_fastbuf()

- the test file updated accordingly

18 years agoconf2: journaling of lists still doesn't work, doubles don't accept "." yet
Robert Spalek [Sat, 22 Apr 2006 15:41:14 +0000 (17:41 +0200)]
conf2: journaling of lists still doesn't work, doubles don't accept "." yet

18 years agoconf2-test: add a simple tester configuration file
Robert Spalek [Sat, 22 Apr 2006 15:32:26 +0000 (17:32 +0200)]
conf2-test: add a simple tester configuration file

18 years agoconf2: several bugfixes, and added interpolation of C-strings
Robert Spalek [Sat, 22 Apr 2006 15:31:15 +0000 (17:31 +0200)]
conf2: several bugfixes, and added interpolation of C-strings

18 years agoconf2: call commit-hooks on dirty sections
Robert Spalek [Sat, 22 Apr 2006 14:38:19 +0000 (16:38 +0200)]
conf2: call commit-hooks on dirty sections

I have, however, added 2 XXX's, because the whole concept of dirtiness has
not been thoroughly thought about.  It's indeed dirty.

18 years agoipaccess: move here temporarily cf_parse_ip()
Robert Spalek [Sat, 22 Apr 2006 13:53:21 +0000 (15:53 +0200)]
ipaccess: move here temporarily cf_parse_ip()

18 years agoconf: comment out parsers before I get rid of the module whatsoever
Robert Spalek [Sat, 22 Apr 2006 13:49:33 +0000 (15:49 +0200)]
conf: comment out parsers before I get rid of the module whatsoever

The parsers collide with conf2.c and I want to be able to have both modules
in the library before I convert all programs.

18 years agoconf2: fill in the skeleton of the tester
Robert Spalek [Sat, 22 Apr 2006 13:48:04 +0000 (15:48 +0200)]
conf2: fill in the skeleton of the tester

18 years agoconf2.c: enable the journal by default
Robert Spalek [Sat, 22 Apr 2006 13:47:06 +0000 (15:47 +0200)]
conf2.c: enable the journal by default

18 years agoconf2: added interface for accessing cf_items from outside
Robert Spalek [Sat, 22 Apr 2006 13:27:39 +0000 (15:27 +0200)]
conf2: added interface for accessing cf_items from outside

- find_item() changed to return a real pointer too
- added cf_find_item() and cf_write_item()
- enum operation moved to the header

18 years agoconf2: implemented a command-line parser and sections supporting unknown items
Robert Spalek [Sat, 22 Apr 2006 13:00:06 +0000 (15:00 +0200)]
conf2: implemented a command-line parser and sections supporting unknown items

18 years agoconf2: full implementation of the parser; only command-line parsing is missing
Robert Spalek [Sat, 22 Apr 2006 12:27:21 +0000 (14:27 +0200)]
conf2: full implementation of the parser; only command-line parsing is missing

plus:
- added macro TRY() polishing the code a bit
- cf_(re)load() and cf_set() return an int, not an error message
- names of operations stringified

18 years agoconf2: added a parser for IP addresses
Robert Spalek [Fri, 21 Apr 2006 18:06:41 +0000 (20:06 +0200)]
conf2: added a parser for IP addresses

plus added two error messages when trying to set an attribute without
giving any value

18 years agoconf2: bugfix in editing and adding before/after a node of a link list
Robert Spalek [Fri, 21 Apr 2006 17:44:13 +0000 (19:44 +0200)]
conf2: bugfix in editing and adding before/after a node of a link list

interpret_add_list() allocates memory himself, hence it is not suitable for
these operations.  since the real code is only 2 lines long, I moved the
functionality to closing_brace().  I changed its interface too, hence it
can be used much simpler and interpret_line() can be returned to its
original form.

18 years agoconf2: bugfix in tracing the recursion
Robert Spalek [Fri, 21 Apr 2006 16:28:16 +0000 (18:28 +0200)]
conf2: bugfix in tracing the recursion

18 years agoconf2: implemented all list operations in the interpreter
Robert Spalek [Fri, 21 Apr 2006 16:20:38 +0000 (18:20 +0200)]
conf2: implemented all list operations in the interpreter

- added operations OP_EDIT, OP_AFTER, OP_BEFORE, and flags OP_1ST, OP_2ND
  denoting the phase (forming a search query vs. filling in the data)
- add_to_list() extended to support all list operations
- definition of stack moved closer to the interpreter, because many
  functions don't need it
- implemented find_list_node(), opening_brace (former increase_stack), and
  closing_brace() that fully support all operations
- the top of interpret_line() rewritten too

18 years agoconf2: record used selectors when searching in lists
Robert Spalek [Fri, 21 Apr 2006 12:45:26 +0000 (14:45 +0200)]
conf2: record used selectors when searching in lists

- store number of entries to section flags
- find_item() fixed so that it correctly handles accesses to unknown
  top-level sections.  also setting an absolute attribute (i.e. not
  relative to the current section) starts with a caret now.
- added an u32 mask to the stack and record_selector() that fills it

18 years agoconf2: implement clearing a dynamic array
Robert Spalek [Fri, 21 Apr 2006 12:04:16 +0000 (14:04 +0200)]
conf2: implement clearing a dynamic array

18 years agoconf2: fix skipping unspecified parameters at the end of the line
Robert Spalek [Fri, 21 Apr 2006 11:49:37 +0000 (13:49 +0200)]
conf2: fix skipping unspecified parameters at the end of the line

Along the way, I precompute flags for each section at the initialization.
Some headers unified.

18 years agoconf2: resolve an issue concerning variable numbers of parameters
Robert Spalek [Fri, 21 Apr 2006 10:40:57 +0000 (12:40 +0200)]
conf2: resolve an issue concerning variable numbers of parameters

When a node of a link list is parsed on one line, then the inner section
is checked for dynamic attributes at the end.  If they are present, the
outer node is just parsed once and the flag allowing dynamic parsing is
passed inside.  If we did not allow this, then an error would have occured
anyway, and this behaviour is actually useful.

Along the way, I unified the headers and semantics of most interpreters and
cleaned up the code.

18 years agoconf2: enhanced the interpreter so that it can parse almost anything on 1 line
Robert Spalek [Fri, 21 Apr 2006 09:26:24 +0000 (11:26 +0200)]
conf2: enhanced the interpreter so that it can parse almost anything on 1 line

I cleaned up the subroutines so that they form a nice recursive structure
that allows to parse almost anything (static and dynamic arrays, lists,
parsed value, and sections in all combinations) on 1 line, while detecting
ambiguities.

18 years agosmall fixes
Robert Spalek [Thu, 20 Apr 2006 23:15:19 +0000 (01:15 +0200)]
small fixes

18 years agoadded CF_UNS, because (int*) and (uns*) are uncompatible pointers
Robert Spalek [Thu, 20 Apr 2006 23:14:25 +0000 (01:14 +0200)]
added CF_UNS, because (int*) and (uns*) are uncompatible pointers

18 years agoimplemented parser of a single line and top-level sections unified
Robert Spalek [Thu, 20 Apr 2006 22:56:54 +0000 (00:56 +0200)]
implemented parser of a single line and top-level sections unified

- we store registered sections in the same format as subsections so that we
  can search/process them in a uniform way
- implemented find_item() that can parse both absolute and relative paths
- all parsers of basic types simplified to take care of only 1 value;
  arrays are handled by the caller
- implemented a part of the interpretter: it gets parsed line (attribute
  name, desired operations, and parameters) and loads the values into
  memory.  since declarations can be nested and they can span over multiple
  lines, the interpretter operates on a stack

18 years agothe init-hook of the main section inserts a few nodes into the link-list
Robert Spalek [Thu, 20 Apr 2006 17:31:55 +0000 (19:31 +0200)]
the init-hook of the main section inserts a few nodes into the link-list

18 years agoadded declaration and initialization of sections
Robert Spalek [Thu, 20 Apr 2006 17:31:05 +0000 (19:31 +0200)]
added declaration and initialization of sections

18 years agofixes
Pavel Charvat [Thu, 20 Apr 2006 16:56:41 +0000 (18:56 +0200)]
fixes

18 years agocf_pool exported, journal not mandatory, and simplified swapping data
Robert Spalek [Thu, 20 Apr 2006 16:24:08 +0000 (18:24 +0200)]
cf_pool exported, journal not mandatory, and simplified swapping data

18 years agoseparated size of bucket ID (4 vs 5) and size of pointer (4 vs 8)
Robert Spalek [Thu, 20 Apr 2006 16:11:03 +0000 (18:11 +0200)]
separated size of bucket ID (4 vs 5) and size of pointer (4 vs 8)

the first one belongs to sherlock.h and depends on the configuration and
the second one belongs to lib/config.h and depends on the architecture

18 years agoSome description and corrections...
Pavel Charvat [Thu, 20 Apr 2006 15:56:09 +0000 (17:56 +0200)]
Some description and corrections...

18 years agoimplemented journal and safe (re)loading of cf/files and setting variables
Robert Spalek [Thu, 20 Apr 2006 13:32:53 +0000 (15:32 +0200)]
implemented journal and safe (re)loading of cf/files and setting variables

18 years agominor changes to KMP:
Pavel Charvat [Thu, 20 Apr 2006 12:42:22 +0000 (14:42 +0200)]
minor changes to KMP:
- added state initializer (KMP_STATE_INIT)
- fixed some formatting rules

18 years agoparsers of basic types exported and got rid of the journaling switch
Robert Spalek [Thu, 20 Apr 2006 12:01:42 +0000 (14:01 +0200)]
parsers of basic types exported and got rid of the journaling switch

18 years agomacros don't end by a comma, and CF_ITEMS/CF_END reintroduced
Robert Spalek [Thu, 20 Apr 2006 11:52:46 +0000 (13:52 +0200)]
macros don't end by a comma, and CF_ITEMS/CF_END reintroduced

the reason is auto-indentation.  in this setting, editors don't get
confused and indent everything well

18 years agoRemoved KMP_ADD_EXTRA_VAR and some useless parenteses.
Pavel Charvat [Thu, 20 Apr 2006 11:36:42 +0000 (13:36 +0200)]
Removed KMP_ADD_EXTRA_VAR and some useless parenteses.

18 years agosubstantially changed the interface again
Robert Spalek [Thu, 20 Apr 2006 10:50:44 +0000 (12:50 +0200)]
substantially changed the interface again

I realised that in the last interface I couldn't distinguish between static
and dynamic arrays of length exactly 1.  thinking about how to resolve the
problem led me to a different solution that is as easy to use as before,
but easier to implement, easier to extend by new basic types, and as a
bonus I get a natural interface for static arrays.

I then immediately scratched a duplicate code for parsing dynamic arrays,
because it can be easily unified now.  also, CF_ITEMS does not expect a
parameter now, because they the compiler messes up line numbers when an
error occurs.

18 years agocustom arguments and return type of kmp_search replaced
Pavel Charvat [Thu, 20 Apr 2006 10:01:09 +0000 (12:01 +0200)]
custom arguments and return type of kmp_search replaced
with user's access to struct search

18 years agoMore changes in KMP:
Pavel Charvat [Thu, 20 Apr 2006 08:05:43 +0000 (10:05 +0200)]
More changes in KMP:
- renamed P(context) -> P(struct)
- removed KMP_NO_DUPS
- changed syntax of user-defined variables in P(struct) and P(state)

18 years agoRemoved old KMP completely
Pavel Charvat [Thu, 20 Apr 2006 07:19:58 +0000 (09:19 +0200)]
Removed old KMP completely

18 years agoRemoved previously added parameter {KMP,HASH}_PARAM_POOL from
Pavel Charvat [Thu, 20 Apr 2006 06:58:00 +0000 (08:58 +0200)]
Removed previously added parameter {KMP,HASH}_PARAM_POOL from
hashtable and KMP... it is not really necessary.
Added full support for non-integer character types and custom
allocation routines to KMP.

18 years agoFixed implicit declaration warning in hash-test.c
Pavel Charvat [Thu, 20 Apr 2006 06:28:55 +0000 (08:28 +0200)]
Fixed implicit declaration warning in hash-test.c

18 years agoEnables the user to give non-global memory pool to hash table or KMP.
Pavel Charvat [Wed, 19 Apr 2006 12:56:42 +0000 (14:56 +0200)]
Enables the user to give non-global memory pool to hash table or KMP.

18 years agoimplemented a few functions of the new configuration mechanism:
Robert Spalek [Wed, 19 Apr 2006 10:02:37 +0000 (12:02 +0200)]
implemented a few functions of the new configuration mechanism:
- memory allocation
- skeleton of future journal functions
- parsers of all basic types and arrays of these types

18 years ago- added prototypes of memory allocation and journal functions
Robert Spalek [Wed, 19 Apr 2006 10:00:02 +0000 (12:00 +0200)]
- added prototypes of memory allocation and journal functions
- rearranged some declarations
- added proper copyright

18 years agoFIxed indentation.
Martin Mares [Wed, 19 Apr 2006 09:50:21 +0000 (11:50 +0200)]
FIxed indentation.

18 years agoSlight improvements to comments.
Martin Mares [Wed, 19 Apr 2006 09:38:03 +0000 (11:38 +0200)]
Slight improvements to comments.

#error when asked to unaccent ASCII characters.

18 years agomissing commit
Pavel Charvat [Wed, 19 Apr 2006 07:58:02 +0000 (09:58 +0200)]
missing commit

18 years ago'make tests' executes also KMP tests
Pavel Charvat [Wed, 19 Apr 2006 07:53:32 +0000 (09:53 +0200)]
'make tests' executes also KMP tests

18 years agoLanguage recognizer uses the new KMP... at last :-)
Pavel Charvat [Wed, 19 Apr 2006 07:29:29 +0000 (09:29 +0200)]
Language recognizer uses the new KMP... at last :-)

18 years agosignificant simplifications of the interface
Robert Spalek [Tue, 18 Apr 2006 12:37:31 +0000 (14:37 +0200)]
significant simplifications of the interface

- instead of letting the user to set cf_section.{size,init,commit}
  directly, macros CF_{TYPE,INIT,COMMIT} should be used
- another macro CF_ITEMS creates a CT_END-delimited list of cf_items
- all these macros automatically add commas so that declaring a list of
  cf_items behaves similarly to other preprocessor-made lists
- removed parameter name from parser hooks, because either the parser
  doesn't care and only uses *ptr, or it does and then it should be set up
  to pass the name in ptr
- no need to define hooks using _void_ *ptr; one can use any pointers
  according to his taste and the hook-function is automatically retyped
- renamed (to make it shorter) some enumerations and macros
- CHECK_VAR_TYPE and CF_FIELD renamed to CHECK_PTR_TYPE and PTR_TO, and
  moved to lib/lib.h

18 years agoadded macros CHECK_PTR_TYPE and PTR_TO
Robert Spalek [Tue, 18 Apr 2006 12:26:59 +0000 (14:26 +0200)]
added macros CHECK_PTR_TYPE and PTR_TO