pike.git/ CHANGES

Branch: Tag:

2004-05-05

2004-05-05 14:07:43 by Leif Stensson <leif@lysator.liu.se>

Proof-reading. Also sorted the longer lists of items into alphabetic order
by the item tag lines.

Rev: CHANGES:1.79

4:   New/improved functionality:   ---------------------------    - o The automatic garbage collection strategy is configurable -  The garbage collector can be configured to take both the amount of -  garbage and the cpu time spent in the gc into account. Automatic -  gc runs can also be disabled in critical regions. See the function -  Pike.gc_parameters. + o fewer magical hacks +  Several Pike modules have been cleaned up from old hacks and +  workarounds for problems which now have proper solutions. This +  should reduce size, and improve performance, dumpability and +  readability of the affected code. In addition lots of internal +  types in Pike have been tightened up and several more files now +  use the pragma strict_types, which has helped locating and fixing +  some previously undiscovered bugs.    - o sort() is stable -  Unlike previous releases, sort() will keep the original order -  between elements that are equal. This was not implemented as a -  separate function since it could be done with virtually no extra -  overhead. + o garbage collection strategy is now configurable +  The garbage collector can now be configured to take both the amount +  of garbage and the cpu time spent in the gc into account. Automatic +  garbage collections can also be disabled in critical regions. (Use +  the function Pike.gc_parameters for this.)    - o sort() is portable -  Sort is now locale and platform independent. In other words, -  strings are now compared only according to the unicode value of -  every character. + o hash() modified +  The string hash functions are now byte-order independent. The old +  hash function is available as hash_7_4 or as hash when running in +  7.4 compatibility mode or lower.    - o Reworked codec -  The codec for serialization of Pike data types has been updated -  and has a significally improved successrate at encoding modules. + o --help +  The Pike --help message has been improved, and it is possible to +  retrieve further information about options and environment +  variables. In addition, most of the tools and build scripts now +  accept --help and describes what they do and what options they +  take.    - o Less magic hacks -  Several Pike modules have been cleaned up from old hacks and -  workarounds for problems which now have proper solutions. This -  should improve size, performance, dumpability and readability -  of the affected code. In addition lots of internal types in Pike -  have been tightened up and several more files now use the pragma -  strict_types, which has resulted in some additional bugfixes. -  +    o INT64    On 64 bit architechtures Pike will now use full 64 bits integer    before resorting to Gmp.mpz objects for large numbers.      o IPv6    Networking code in Pike has been updated to work with IPv6. This -  includes both low level fixes in the socket handling as well as -  high level changes such as IP parsing and protocol additions such +  includes both low-level fixes in the socket handling as well as +  high-level changes such as IP parsing and protocol additions such    as AAAA records in Protocol.DNS.    -  + o module system improved +  Both dynamic and static modules now appears as objects, and they both +  support having an identifier _module_value, which will then be +  what is presented by the resolver as the module. _module_value can +  be a constant, which will prevent the program from being automatically +  cloned. +    o mysqls -  The Mysql glue can operate over SSL. Use the "mysqls://" protocol -  in the Sql.Sql module to use this security feature. +  The Mysql glue can operate over SSL, providing additional data +  security. Use the "mysqls://" protocol in the Sql.Sql module to +  access this feature.    -  + o reworked codec +  The codec for serialization of Pike data types has been updated +  and has a significally improved success rate at encoding modules. +  (See the encode_value and decode_value functions.) +  + o sort() is portable +  Sort is now locale and platform independent. In other words, +  strings are now compared only according to the unicode value of +  every character. +  + o sort() is stable +  Unlike previous releases, sort() will keep the original order +  of elements that are equal. At first, we considered keeping the old +  sort function in parallel with the new one, as a "faster but +  unstable" alternative, but it turned out that there was no noticable +  different in speed between the new and the old function. +    o trace() can trace all threads    The configure option --without-thread-trace has been removed, but    instead the trace() function has been extended to make it possible    to set the trace level in all threads simultaneously.    - o --help -  The Pike --help message has been improved and it is possible to -  retrieve further information about options and environment -  variables. In addition most of the tools and build scripts now -  accepts --help and describes what they do and what options they -  take. -  +    o unbug    The Pike debugger unbug has been significtly optimized.    - o Module system. -  Both dynamic and static modules appears as objects and they both -  support having an identifier _module_value, which will then be -  what is presented by the resolver as the module. _module_value can -  be a constant, thus preventing the program from being -  automatically cloned. -  +    o Unicode    Pike 7.6 uses the Unicode 4.0.1 tables for case info and    normalization.    - o upper_case/lower_case -  The built in functions upper_case and lower_case now also works on -  single characters, represented as integers. + o upper_case() and lower_case() extended +  The built-in functions upper_case and lower_case now also works on +  single characters represented as integers.    - o hash -  The string hash functions is now byte-order independent. The old -  hash function is available as hash_7_4 or as hash when running in -  7.4 compatibility mode or lower. -  +    o #warning    A new The Pike preprocessor directive #warning has been added. It    simply writes the rest of the line to stderr.
91:   ----------------      o The new option --dumpversion prints out the Pike version in a more -  easily parsed way than --version. +  easily pared way than --version.    - o The master now supplies argv to the started application's create() -  callback, so it can be used for execution. Some people find it a -  more elegant solution than having a main function. + o The master now supplies argv to the started application, so it can +  use the create() method for execution. Some people find it a more +  elegant solution than having a main function.    - o The -e Pike environment has been improved with a lot of convenience -  wrappings. When compiling code with pike -e the macro NOT(X) will + o The "-e" flag to Pike has been improved with a number of convenience +  wrappers. When compiling code with "pike -e", the macro NOT(X) will    expand to !(X) and CHAR(X) to 'X', making it easier to solve quote    problems from a shell. If the return from a pike -e statement is a    positive integer it will be used as pikes exit code. If the return -  value is a string it will be output, and with a trailing "\n", if it -  lacked one. Finally there is a predefined variable env that contains -  the environment variables as a string:string mapping. +  value is a string it will be output (with a "\n" added, if the string +  didn't already end in "\n"). Finally there is a predefined variable +  env that contains the environment variables as a string:string mapping.         Building and installing pike:
112:      o To simplify the process of building Pike source distributions it is    now possible to bundle less common libraries such as Nettle with -  Pike. The bundles, found in the bundle/ directory, will be compiled -  and statically linked if the no matching library is found on the -  system. +  Pike. The bundles, found in the "bundles" directory, will be compiled +  and statically linked if no matching library is found on the system.      o The Pike build process no longer builds a bootstrap pike, used to    run various small programs to generate source files. This means that    in order to build Pike from CVS you need the latest major version of -  Pike installed on your system. Pike source distributions will still -  contain the generated files and are not affected (except they compile -  in less time). +  Pike installed on your system. Pike source distributions contain +  these generated files, and are not affected.    - o By default pike want you to build a documentation source file, i.e. + o By default, pike want to build a documentation source file, i.e.    "make && make doc && make install". This makes it possible for third    party modules to integrate its documentation into the system Pike    documentation. If you do not want to build a documentation source
133:   Changes and additions to -x utilities:   --------------------------------------    - o rsif -  rsif can now work recursively in a file structure if given the -  flag -r or --recursive. Example: -  pike -x rsif -r 7.4 7.6 -  +    o benchmark    A set of benchmarks is available in the benchmark tool. They are    based on the Great Computer Language Shootout, but have been -  modified to make more sense in the context of Pike and to test +  modified to make more sense in the context of Pike, and to test    various aspects of Pike we suspect (or at least at one time -  suspected) perform badly. +  suspected) to have bad performance.      o cgrep -  Context sensitive grep (grepping through tokenized code, at -  present handling Pike and C source files) is very helpful when -  you are looking for substrings that are common. The implemented -  search modes are: +  Context sensitive grep is very helpful when you are looking for +  substrings that are common. The implemented search modes are:    o Complete token    o Part of a token    o Complete string literal
176:      o monger    A new utility to download, install and otherwise manage external -  modules from the Fresh Pike project (modules.gotpike.org). +  modules from the Fresh Pike project (modules.gotpike.com).    -  + o rsif +  rsif can now work recursively in a file structure if given the +  flag -r or --recursive. Example: +  pike -x rsif -r 7.4 7.6    -  +  +    New modules / classes / methods added:   --------------------------------------    - o Regexp.PCRE added to give libpcre support. -  Are you a closet-perlet? For those who missed the richer flora of -  regexp syntax that Perl supports, the Perl Compatible Regular -  Expression module is the answer. If you compile libpcre with UTF-8 -  support you will also get transparent wide string support. Several -  convenience functions are also added to the Regexp module to -  enable easier access to regular expressions. + o ADT.BitBuffer +  When operating on data formats that are defined on bit level, the +  ADT.BitBuffer presents a convenient interface. It operates as a +  FIFO buffer on bit level, which allows you to read and write bits +  and bytes.    -  +  > ADT.BitBuffer b=ADT.BitBuffer(); +  > b->put1(2); +  (1) Result: ADT.BitBuffer(11) +  > b->put0(15); +  (2) Result: ADT.BitBuffer("À\0"0) +  > b->drain(); +  (3) Result: "À\0" +  > sizeof(b); +  (4) Result: 1 +  + o ADT.CircularList +  This is an circular list implemented as a C wrapper around an +  array. It has a constant time complexity for pop and push. It has +  a limited max size but it can be increased manually by calling the +  allocate() method. It supports pushing, popping and peeking at both +  ends of the list, as well as removing and insering elements at any +  position in the list. +  + o ADT.Sequence +  The sequence work similar to an array but has the ability to +  insert and remove elements at any position. It also has functions +  to retrieve iterators for forward and backward iteration over the +  elements in the sequence. +  + o ADT.Struct +  This module makes it possible to work with binary formats of the +  "packed struct" type. Simply create a class inheriting ADT.Struct, +  specify its members in order, create an object with a file object +  as argument, and then read the decoded values from the object as +  ordinary variables. An example, reading ID3 metadata headers from +  an MP3 file: +  +  class ID3 { +  inherit ADT.Struct; +  Item head = Chars(3); +  Item title = Chars(30); +  Item artist = Chars(30); +  Item album = Chars(30); +  Item year = Chars(4); +  Item comment = Chars(30); +  Item genre = Byte(); +  } +  +  Stdio.File f = Stdio.File("foo.mp3"); +  f->seek(-128); +  ADT.Struct tag = ID3(f); +  if(tag->head=="TAG") { +  write("Title: %s\n", tag->title); +  tag->title = "A new title" + "\0"*19; +  f->seek(-128); +  f->write( (string)tag ); +  } +  + o Array +  The Array module has been extended with Array.all and Array.any +  which determine if all or any of the elements in an array meet a +  given condition. The function Array.partition has also been added, +  which divides the elements of an array into two arrays, selected +  according to a given condition. +  + o Bz2 +  Support for the compression format used by the bzip2 program. The +  module interface is almost identical to that of the Gz module. +  + o Calendar module +  The %S format was added to Calendar.parse() to parse seconds since +  the Epoch. The TimeRange class is extended with a format method +  format_commonlog(), a tzname_utc_offset (+HHMM) and +  tzname_location(), which, when available, returns zone names like +  "Europe/Stockholm". The database has been updated with Swedish +  pre-1901 namedays and Finnish namedays. +    o Crypto/Nettle -  The crypto module has been almost completely rewritten and now -  uses libnettle as backend instead of the built in crypto +  The crypto module has been almost completely rewritten, and now +  uses libnettle as backend instead of the built-in crypto    algorithms that Pike had before. This way Pike will benefit from    the more active development and optimization work in the Nettle    project. The most common hash and cipher methods are already
209:    future.       The interfaces for the hash and cipher algorithms have been -  improved and they no longer only conform to predefined interfaces, +  improved, and no longer only conform to predefined interfaces,    but actually inherits Hash/HashState and Cipher/CipherState. Every    algorithm is in itself a module in the Crypto module, and inherits    the Hash/Cipher class. This module contains instance non-specific
223:       Crypto.MD5()->update("Hash this ")->update("string")->digest();    -  The confusing and hard to use Crypto.randomness has been obsoleted +  The confusing and hard-to-use Crypto.randomness has been obsoleted    by the Nettle implementation of the Yarrow random generator. The    Nettle implementation of Yarrow uses AES and SHA256. The    Crypto.Random module acts as a frontend before system crypto -  sources and ensure that no matter how bad your OS is, you will -  still get cryptographically strong random data. +  sources, and ensure that almost no matter how bad your OS is, you +  will still get cryptographically strong random data.    -  +    o Crypto.PGP    The start of a PGP/GPG toolkit. It still cannot handle all types    of signatures nor generate any, but it is useful to verify    selfsigned code and data.    -  - o ADT.Struct -  This module makes it possible to work with binary formats of the -  packed-struct-type. Simply create a class inheriting ADT.Struct, -  specify the members with type and size, in order, instantiate an -  object with a file object as argument and then read the decoded -  values from the object as ordinary variables. -  -  class ID3 { -  inherit ADT.Struct; -  Item head = Chars(3); -  Item title = Chars(30); -  Item artist = Chars(30); -  Item album = Chars(30); -  Item year = Chars(4); -  Item comment = Chars(30); -  Item genre = Byte(); -  } -  -  Stdio.File f = Stdio.File("foo.mp3"); -  f->seek(-128); -  ADT.Struct tag = ID3(f); -  if(tag->head=="TAG") { -  write("Title: %s\n", tag->title); -  tag->title = "A new title" + "\0"*19; -  f->seek(-128); -  f->write( (string)tag ); -  } -  -  - o ADT.BitBuffer -  When operating on data formats that are defined on bit level, the -  ADT.BitBuffer presents a convenient interface. It operates as a -  FIFO buffer on bit level, which allows you to read and write bits -  and bytes. -  -  > ADT.BitBuffer b=ADT.BitBuffer(); -  > b->put1(2); -  (1) Result: ADT.BitBuffer(11) -  > b->put0(15); -  (2) Result: ADT.BitBuffer("À\0"0) -  > b->drain(); -  (3) Result: "À\0" -  > sizeof(b); -  (4) Result: 1 -  -  +    o Debug.Wrapper    A simple litte wrapper that can be placed around another object to    get printouts about what is happening to it. Only a few LFUNs are
298:    (2) Result: 0       - o Image.NEO -  Support for one of the major image formats on Atari ST/STE, -  including decoding of color cycling. The C bitmap and palette -  handler can be used to implement other atari image formats. + o Error +  The root of the Pike exception hiearchy. Even though it is work +  in progress, it is already usable.    -  +  > mixed err = catch( 1/0 ); +  > err; +  (1) Result: Error.Math("Division by zero.\n") +  > object_program(err)==Error.Math; +  (2) Result: 1 +  + o EXIF +  The EXIF module is now much more robust than before, and can handle +  some of the erroneous EXIF tags some cameras procduce. Extended +  support for Nikon cameras. +  + o Filesystem.Traversion +  The Filesystem.Traversion iterator is a handy way to descend into +  a directory structure (depth first) and perform actions on all or +  some of the files. +  +  Example: +  +  foreach(Filesystem.Traversion("."); string dir; string file) +  if(file[-1]=='~' || file=="core") rm(dir+file); +  + o Float +  Here constants and functions that deal specifically with the +  float data type is stored. Currently, it only contains various +  constants that describe the precision and limitations of the float +  type with which Pike was compiled. +  + o Function +  The Function module has been extended with the fixpoint combinator +  "Y", useful when writing recursive lambdas, a curry function, to +  partially evaluate a function call, and a callback caller function +  that protects the caller from exceptions in the callback function. +  + o Geography +  The Geography.Position object can now also convert to and from the +  Swedish RT38 (Rikets Triangulering 1938) position format. +    o GLUE    The GL Universal Environment is a GL toolkit that makes    implementing GL applications in Pike really simple. GLUE has -  several orthogonal components like object orientation -  encapsulation of GL, drawing primitives, texture handling and text -  generation. The most fundamental part is however an abstract -  interface to a GL area driver, like GTK or SDL. +  several orthogonal components like object oriented encapsulation +  of GL, drawing primitives, texture handling and text generation. +  The most fundamental part is however an abstract interface to a +  GL area driver, like GTK or SDL.       int main() {    GLUE.init();
323:    GLUE.swap_buffers();    }    - o ADT.CircularList -  This is a circular list implemented as a C wrapper around an -  array. It has a constant time complexity for pop and push. It has -  a limited max size but it can be increased manually by calling -  allocate. It supports pushing, popping and peeking at both ends of -  the list, as well as removing and insering elements at any -  position in the list. + o Image.NEO +  Support for one of the major image formats on Atari ST/STE, +  including decoding of color cycling. The C bitmap and palette +  handler can be used to implement other atari image formats.    - o ADT.Sequence -  The sequence works similar to an array but has the possibilities -  to insert and remove elements at any position. It also has methods -  to retrieve iterators that work from either side of the sequence. -  Both of the above ADT:s share a common set of interface primitives -  described in a thesis work available for download and review at -  -  http://pike.ida.liu.se/projects/docs/cmods/ADTs%20in%20Pike.pdf -  -  - o Geography -  The Geography.Position object can now convert to and from the -  Swedish RT38 (Rikets Triangulering 1938) position format. -  +    o Int    The Int module contains a few functions and constants that deal    specifically with the integer data type. Currently it contains -  parity and byte swapping functions and constants for the system +  parity and byte swapping functions, and constants for the system    maxint and minint.      o Pike.Security
357:    are stored in this module. Beware that the Pike security system is    still very experimental.    - o Protocols.HTTP.Session -  Handles an HTTP connection as a session, with cookies, redirects, -  HTTP keep-alive et cetera. -  - o Bz2 -  Support for the compression format used by the bzip2 program. The -  module interface is almost identical with the Gz module, so -  changing between Bz2 and Gz is just a matter of a few characters' -  difference. -  +    o Process.Process -  The Process.Process class is extended to handle a time out -  callback and a read callback, to better manage processes in an -  asynchronous way. +  The Process.Process class is extended to handle a time-out +  callback and a read callback, to make it easier to manage +  processes in an asynchronous way.    - o Error -  The beginning of the Pike exception hiearchy. Even though it is -  work in progress, it is already usable. -  -  > mixed err = catch( 1/0 ); -  > err; -  (1) Result: Error.Math("Division by zero.\n") -  > object_program(err)==Error.Math; -  (2) Result: 1 -  -  - o Float -  Here, constants and functions that deals specifically with the -  float data type are stored. Currently contains only various -  constants that describe the precision and limitations of the float -  type Pike is compiled with. -  +    o Protocols.Bittorrent    This is a Pike implementation of the client side of the popular -  Bittorrent protocol. Both creating torrent files and seeding, as -  well as downloading and verifying files is supported. There are +  Bittorrent protocol. Both creating torrent files and seeding as +  well as downloading and verifying files are supported. There are    also several functions to retreive auxillary information such as -  identifying the other peer clients. Contains all the primitives -  you would expect, except for the tracker. +  identifying the other peer clients.    -  + o Protocols.HTTP.Session +  Handles an HTTP connection as a session, with cookies, redirects, +  HTTP keep-alive, etc. +    o Protocols.LMTP/Protocols.SMTP    Protocols.LMTP implements a server for the Local Mail Transfer    Protocol, LMTP, as specified in RFC2033. This module has been well
408:    a protocol for sending and receiving binary objects to mobile    devices using transports such as IR and Bluetooth.    + o Regexp.PCRE added to give libpcre support. +  Are you a closet perler? For those who missed the richer flora of +  regexp syntax that Perl supports, the Perl Compatible Regular +  Expression module is the answer. If you compile libpcre with UTF-8 +  support you will also get transparent wide string support. Several +  convenience functions are also added to the Regexp module to +  enable easier access to regular expressions. +  + o SSL +  Large parts of the SSL module have been rewritten, either as part +  of the rewrite of the Crypto module or in an attempt to fix various +  issues regarding blocking/nonblocking files, threads, etc. +  + o Standards.IDNA +  This module implements various algorithms specified by the +  Internationalizing Domain Names in Applications (IDNA) memo by the +  Internet Engineering Task Force (IETF), RFC 3490. +  + o Thread +  Thread.Fifo and Thread.Queue have been extended with try_* +  varieties of all functions that can block, and with functions that +  make them both implement the same interface. Both also have +  fallbacks for running in nonthreaded mode. +    o Web.RDF/RDFS/OWL/RSS    Pike now has support for the W3C semantic web formats OWL, RDF    and RDFS, which are used for expressing structure and metadata
428:       foreach(owlset->find_statements(0,    owlset->rdf_type, -  owlset->owl_Class), +  owlset->owl_Class);    array statement)    write("- %O\n", statement[0]); -  +  +  return 0;    }       - o Array -  The Array module has been extended with Array.all and Array.any -  which determines if all or any of the elements in an array meet a -  given condition. Also the function Array.partition has been added, -  which divides the elements of an array into two arrays, selected -  by a given condition. -  - o Function -  The Function module has been extended with the fixpoint combinator -  "Y", useful when writing recursive lambdas, a curry function, to -  partially evaluate a function call and a callback caller function -  that protects the caller from exceptions in the callback function. -  - o Thread -  Thread.Fifo and Thread.Queue have been extended with try_* -  varieties of all functions that can block and with functions that -  make them both implement the same interface. Both also have -  fallbacks for running in nonthreaded mode. -  - o Filesystem.Traversion -  The Filesystem.Traversion iterator is a handy way to descend into -  a directory structure (depth first) and perform actions on all or -  some of the files. -  -  Example: -  -  foreach(Filesystem.Traversion("."); string dir; string file) -  if(file[-1]=='~' || file=="core") rm(dir+file); -  -  - o Calendar module -  The %S format is added to Calendar.parse() to parse seconds since -  the Epoch. The TimeRange class is extended with a format method -  format_commonlog(), a tzname_utc_offset (+HHMM) and -  tzname_location(), which, when available, returns zone names like -  "Europe/Stockholm". The database has been updated with Swedish -  pre-1901 namedays and Finnish namedays. -  - o SSL -  Large parts of the SSL module have been rewritten, either as part -  of the rewrite of the Crypto module or in an attempt to fix many -  issues regarding blocking/nonblocking files, threads et cetera. -  - o EXIF -  The EXIF module is now much more robust than before and can handle -  some of the erroneous EXIF tags some cameras procduce. Extended -  support for Nikon cameras. -  - o Standards.IDNA -  This module implements various algorithms specified by the -  Internationalizing Domain Names in Applications (IDNA) memo by the -  Internet Engineering Task Force (IETF), RFC 3490. -  -  +    Optimizations:   --------------    - o Instantiation and destruction of pike-classes is now significantly + o Instantiation and destruction of pike classes is now significantly    faster.      o Handling of bignums has been optimized, especially conversion
504:   Compatibility changes:   ----------------------    - o Incompatible changes has been done to ADT.Heap, Array.diff3_old, + o Incompatible changes has been made in ADT.Heap, Array.diff3_old,    Crypto, Locale.Charset, Protocols.SMTP, SSL, Sql.sql,    Standards.ASN1, Standards.PKCS, Stdio.File, Thread.Mutex and    rusage(). Please refer to the 7.4 namespace in the compatibility
519:      o Module ABI changed    The pike_frame struct has a new member, which makes modules compiled -  with earlier versions of Pike incompatible. +  with Pike 7.4, and earlier versions of Pike, incompatible with Pike 7.6.         C level/development changes:
527:      o The security.h include file is renamed to pike_security.h    - o The testsuite has three new, handy macros, cond_resolv, -  test_program_eq, test_program_equal, all defined in mktestsuite. A + o The testsuite has three new, handy macros: cond_resolv, +  test_program_eq and test_program_equal, all defined in mktestsuite. A    testsuite test that produces a warning, unless it is a test_warning    test, will now be considered an error. The test program can now run    testsuites in forked processes. By default every testsuite file will