memory leaks. use string object for buffering instead.
[ruby-dev:24738]
* dir.c (join_path): ditto.
* io.c (io_read): external input buffer may be modified even after
rb_str_locktmp(). [ruby-dev:24735]
* dir.c (fnmatch): p or s may be NULL. [ruby-dev:24749]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7242 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
or continuation is created. [ruby-dev:24671]
* io.c (rb_io_s_popen): do not expand argv array. [ruby-dev:24670]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7174 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[ruby-talk:117841]
* ruby.h (FL_ABLE): nodes are not subject for flag operations.
* io.c (ARGF_FORWARD): should have specified argv explicitly,
since we no longer have frame->argv saved. [ruby-dev:24602]
* string.c (RESIZE_CAPA): check string attribute before modifying
capacity member of string structure. [ruby-dev:24594]
* ext/zlib/zlib.c (gzreader_gets): use memchr() to to gain
performance. [ruby-talk:117701]
* sprintf.c (rb_f_sprintf): raise ArgumentError for extra
arguments, unless (digit)$ style used.
* io.c (rb_io_fptr_finalize): leave stdin/stdout/stderr open in
interpreter termination. [ruby-dev:24579]
* eval.c (frame_free): Guy Decoux solved the leak problem.
Thanks. [ruby-core:03549]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7121 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
the class of right operand. [ruby-dev:24526]
* ext/zlib/zlib.c (zstream_detach_buffer): should not expose
class-less object to Ruby world. [ruby-dev:24530]
* eval.c (proc_dup): provide Proc#dup as well. [ruby-talk:116915]
* eval.c (ruby_exec): stack marking position may be higher than
expected. thanks to Guy Decoux. [ruby-core:03527]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7077 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
rb_io_fread() by freezing it temporarily. [ruby-dev:24479]
* dir.c (rb_push_glob): block call at once the end of method.
[ruby-dev:24487]
* ext/enumerator/enumerator.c (enum_each_slice): remove
rb_gc_force_recycle() to prevent potential SEGV.
[ruby-dev:24499]
* ext/zlib/zlib.c (zstream_expand_buffer): hide internal string
buffer by clearing klass. [ruby-dev:24510]
* ext/socket/socket.c (sock_s_getservbyaname): protocol string
might be altered. [ruby-dev:24503]
* string.c (rb_str_upto): check if return value from succ is a
string. [ruby-dev:24504]
* io.c (rb_io_popen): get mode string via rb_io_flags_mode() to
avoid mode string modification. [ruby-dev:24454]
* io.c (rb_io_getline_fast): should take delim as unsigned char to
distinguish EOF and '\377'. [ruby-dev:24460]
* io.c (rb_io_getline): add check for RS modification.
[ruby-dev:24461]
* enum.c (enum_sort_by): use qsort() directly instead using
rb_iterate(). [ruby-dev:24462]
* enum.c (enum_each_with_index): remove rb_gc_force_recycle() to
prevent access to recycled object (via continuation for
example). [ruby-dev:24463]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7071 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
add WEBrick::Config::FileHandler[:AcceptableLanguages].
* lib/webrick/httpservlet/filehandler.rb
(WEBrick::HTTPServlet::FileHandler#set_filename): search files
having suffix of language-name which Accept-Language header field
includes if :AcceptableLanguages options is present.
* lib/webrick/httpservlet/filehandler.rb
(WEBrick::HTTPServlet::FileHandler#get_servlet): new method to
search servlet correspond to the suffix of filename.
* lib/webrick/httprequest.rb: add attributes access methods: accept,
accept_charset, accept_encoding, accept_language, content_length
and content_type.
* lib/webrick/httpresponse.rb: add attribute access methods:
content_length, content_length=, content_type and content_type=.
* lib/webrick/httputils.rb (WEBrick::HTTPUtils.mime_types):
use the second suffix to detect media type. (the first suffix
may be a language name.)
* lib/webrick/httputils.rb (WEBrick::HTTPUtils.parse_qvalues):
add method to parse Accept header field. it returns an Array of
values sorted by the qvalues.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7033 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/fileutils.rb (remove_dir): return value of Dir.foreach is reliable. (pass $SAFE=1)
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@7017 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* io.c (rb_io_flags_mode): requires output buffer no more. no
allocation needed.
* array.c (rb_ary_index): takes a block to compare items in an
array. [ruby-talk:113069] [Ruby2]
* array.c (rb_ary_rindex): ditto.
* marshal.c (r_byte): retrieve pointer from string value for each
time. [ruby-dev:24404]
* marshal.c (r_bytes0): ditto.
* enum.c (sort_by_i): re-entrance check added. [ruby-dev:24399]
* io.c (io_read): should freeze all reading buffer.
[ruby-dev:24400]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@6996 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
size. [ruby-dev:24273]
* parse.y: make ruby parser reentrant. merge ripper parser to the
real one. this change makes ruby require bison.
* file.c (rb_file_truncate): clear stdio buffer before truncating
the file. [ruby-dev:24191]
* ext/digest/digest.c: use rb_obj_class() instead of CLASS_OF
which might return singleton class. [ruby-dev:24202]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@6919 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to request.user attribute.
* lib/webrick/httpservlet/filehandler.rb
(WEBrick::HTTPServlet::FileHandler#initialize): should expand
the pathname of document root directory.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@6913 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[ruby-talk:110664]
* error.c (exc_equal): exceptions are equal if they share same
class, message and backtrace. [ruby-talk:110354]
* error.c (name_err_mesg_equal): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@6826 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
