CVS log for parser3/www/htdocs/Attic/_bug.html

Up to [parser3project] / parser3 / www / htdocs

Request diff between arbitrary revisions - Display revisions graphically

Keyword substitution: kv
Default branch: MAIN

cleanup

+ ^break[] ^continue[], in ^for, ^while, ^menu, ^hash/hashfile.foreach

! switched to VS2005 (projects and object<info explicitly typed>.for_each(, info))
+ $response:field[] setting void means removing
+ grammar $name` name stops now, for mysql `$field`
+ started ^break[] ^continue[], in ^for. TODO: to other iterators (while, menu, foreach)
+ all log messages += [uri=, method=, cl=]

. merged from 3.2.1

. win32: grammar is now compiled with bison 1.875b, it reports unexpected token names

! bugfix xnode now holds a link to xmlNode to prevent premature gc(xmlNode)

. removed .nodeValue from all node types other than 5 (grabbed piece from gdome)

. more tests

! fixed overoptimized ^call(false) case. (confused it with ^call[] case)

. started killing gdome

. less warnings

+ ^while(){}[SEPARATOR]

merged from HEAD:
! changed file/dir_readable to simple file/dir_exist, this would help in situations "class not found because .p file has bad rights" << in that case error would be explicit "access denied to 'this' file"

+! ^file::exec/cgi[script;$.charset[changed] $.QUERY_STRING[^untaint[URI]{aaa=$form:text} << now %HH would be encoded in $.charset charset

+ $exception.handled[cache] now reports original exception if we have no old cache

merged from HEAD:
+ ^taint/untaint[regex] << escapes these: \^$.[]|()?*+{}

+ ^taint/untaint[regex] << escapes these: \^$.[]|()?*+{}

+ $response:headers access to internal hash

+ GC_large_alloc_warn_suppressed=0 between requests [apache mod_, isapi]
this reduces number of "GC Warning: Repeated allocation of very large block" messages to only important onces

merged from HEAD:
+ ^file.base64[] encode
+ ^file::base64[encoded] decode

+ ^file.base64[] encode
+ ^file::base64[encoded] decode

+ ^string.base64[] encode
+ ^string:base64[encoded] decode

+ ^string:base64[in] encode
+ ^string.base64[] decode

no message

+ merged from HEAD: ^file::create[text;file.xml;^untaint[xml]{data}]

+ ^file::create[text;file.xml;^untaint[xml]{data}]

+ expression literals: true/false. ^format[$.indent(true)]

+ merged: $env:PARSER_VERSION reports "3.1.5beta (compiled on i386-pc-win32)"

+ $env:PARSER_VERSION reports "3.1.5beta (compiled on i386-pc-win32)"

+ ^cache[] << returns current cache expiration time

! merged: file:list now sees .xxx files, only . and .. now removed from list [were removed all .*]

merged to HEAD:
! table::save with enclosers now doubles them: "->""
! table::save/load do not remove elements from options hash

! table::save with enclosers now doubles them: "->""
! table::save/load do not remove elements from options hash

merged + date.week

+ date.week

! optional options were allowed to be empty. were checked if(is_string), changed to !defined||is_string to allow void

! many classes/*  used as_junction where they needed as_int/double, so failed with recent "(const) now no junction" optimization

no message

. file::exec $charset transcodes env fine

+ ^call[$void] param inside now is void (were: empty string)
! simplified (double) speedup consequent checks in vmethod_frame

merged
+ optimized constants in expressions vstring->vdouble
+ optimized ^call(vdouble), no junction creation, no calls later

. merged ! second hashfile::open would cause an exception

. merged VJunction(new Junction) optimization from 3.1.4

no message

+ removed limit "ctor must be declared in class being created"

+ 3.2.0 beta: merged from dynamic_fields_join

! new asserts were all false. regretfully reverted them all to checks

joined_dynamic_fields: dynamic get/set works with overriding props [alpha2]

joined_dynamic_fields: dynamic get/set works [alfa]

. object setters [only started, does not work yet]

. properties: set works [alpha2]. for classes. todo: for objects

. properties: set works [alpha1]

. properties: started, get works [alpha1, other approach]

. properties: started, get works [alpha2]

no message

! bugfix: $response:body[nonfile] caused gpf

! bugfix ^string:sql ^table:sql ^hash:sql now DO process $.bind option

. compiled on freebsd 5.3 with gcc 3.4

beauty: removed needless includes

version

sql connect string, rsplit @

win32: bugfix: file::cgi/exec now set current dir to directory of script (when shbang used), not to directory of its interpreter

new: @unhandled_exception can prevent exception from being logged by setting $exception.handled(1)

change: ^file::sql{query}[[options]]
options:
$.name[file name]
$.content-type[application/what]

beauty: hashfile::open now creates dir for new file

bugfix: table->hash now regards _default

bugfix: when method junction was created by accessing $name_of_method it acquired bad self=closest methodframe; instead of proper self (current class)

bugfix: reading empty string from hashfile produced bad cord

win32: http connection $.timeout option now works [setsockopt on send/receive]

bugfix: mail body now cstr-ed knowing mail charset, and untainting uri lang now knows proper charset

debugger help: PA_RELEASE_ASSERTS enables release asserts

bugfix: $s[+008] ^eval($s) now parsed OK [were as octal]

beauty: not reproduced $hashfile.key[$novalue] bug [reported by motorin],
inserted safety-check

bugfix: error numbers now from errno.h = strerror now returnes something and we can properly report that to client

change: ^xdoc.string now outputted as-is  [tainted as-is]
this helps //[space][newline] to remain as-is and not be unnecessary optimized away [ruining javascript]

new: ^mail:send[$.file[$.content-disposition can be overriden
new: empty mail header fields removed from letter

new: ^mail:send[
   $.file[
      $.any[header]

beauty: less warnings

bugfix: empty input variables can be replaced by output
[forgot to allocate proper buffer, oracle _server_ died with
kgepop: no error frame to pop to for error 21500
message]

bugfix: win32: system&parser exceptions in release mode reported properly
parser exception in exception handler WERE mistakenly reported as system exception, without details

bugfix: ^table.save[export.csv;$.separator[^;]] now works fine
[separator were ignored, and \t used unconditionally]

bugfix: isspace((unsigned char)c) everywhere. failed on russian letters

bugfix: win32: chdir not needed, dir passwed as 'currentDirectory' parameter to CreateProcess

new: $response:body/download[
	$.file[name on disk]
	$.name[of file for user]
	$.mdate[date of last-modified. default from directory]
]

bugfix: too long header values now splitted to several lines

were:
header: vaaaaaaaaaaaalue
now:
header: vaaaaaa
 aaalue

note: ms outlook[!express] shows only first 255 characters of subject

bufix: on unix AND win32 environment string now untainted according to their languages.
were: as-is. EVERYWHERE

convinience: transcodeFromUTF8 now never fails. in case on input appears nonutf,
those bytes will be printed in %HH form. that can be easily decoded/recovered.

this form is quite noticable, and hopefully would not go noticed

workaround kinda bug in libxml: life requires to do xdoc::create{invalid chars}
standard disables chars less then \x20, except tab, cr, lf.

changed tainting so that those become '!'

checked: oracle: various ways of returning error from stored proc.
bad: no way of knowing which exactly user-defined exception were thrown.
good: raise_application_error can return error message number and string FINE
good: PRAGMA EXCEPTION_INIT works good too

checked: in out variables OK

new:
	^void:sql{call paf(:a)}[
		$.bind[
			$.a[2]
		]
	]
output variables work.
todo: check in out variables

started:
	^void:sql{call paf(:a)}[
		$.bind[
			$.a[2]
		]
	]
input variables work.
todo:output

beauty: removed outdated level 'table'

cancel: more speed, less memory: CORD_chars_block originally intended to ...
it was good that CORD_chars_block were not used: it consumes more memory [and, might be slow too]
undoing that. [and removed that func so that it would not confuse parser developer in future]

bugfix: void now has vfile value, and $response:body[] works OK

beauty: ^if(def $bool) now equals ^if($bool)
and shaped up other sources to use is_defined() {return as_bool();} along VBool

bugfix: bcc with sendmail now left intact

bugfix: step3 to fix explicit result problem

bugfix: step2 to fix explicit result problem

bugfix: step1 to fix explicit result problem

attempt_check_call_in_explicit_result_mode: failed

reimplemented: @method[][result] means "no string output here"
moved to lexical level [on grammar level it were too difficult]

new: @method[][result] means "no string output here",
implemented part of that idea: compiler throws away string literal generation code. and barks nonwhitespace chars

merged with 3.1.3

merged bugfixes from 3.1.2, changed version to 3.1.3beta

bugfix: ^hash::sql{one column} now produces $.hash[$.column_value1(true) ... ]
[were producing some strang thing]

bugfix: ^image.polyline and http status line parsing
checked number of columns in table [user reported an assert]

bugfix: junction reattach now actually reattaches a junction to new wcontext [were only assigning it to junction,
which caused junction tracking to loose junction, and it's context were not killed, and it tried to process in invalid context, gpf]

beauty: error messages removed duplicate "type(type)" in braces

new: refined solution to search-in-namespaces problem:
$xdoc.search-namespaces.x[http://pif.design.ru/]
$nodes[^xdoc.select[//x:second]]

new: xnode.select*[xpath expression][[NAMESPACES HASH]]
allows to search for info in namespaces

$nodes[^xdoc.select[//x:second][
	$.x[http://pif.design.ru/]
]]

new: ^file::load[...][options]
        $.form[
            !$.field1[string]
            !$.field2[^table::create{one_column_only^#0Avalue1^#0Avalue2}]
        ]
        !$.body[string]

GET ?here
notGET(POST, HEAD, others) in content-type: application/x-www-form-urlencoded

todo: upload files ability

new: http:// options $.user $.password basic authorization

new: odbc option [like that recently added to oracle driver]
            ClientCharset=parser-charset << charset in which parser thinks client works

new: mysql option [like that recently added to oracle driver]
            ClientCharset=parser-charset << charset in which parser thinks client works

new:
    !^void.left(n)
       nothing
    !^void.right(n)
       nothing
    !^void.pos(p[;n])
       nothing

new: ^file::sql[[name_to_become_$.name]]{}
query result must be one row with columns:
  first: data
  second: file name
  third: content-type

^file.sql-string[] inside ^connect gets properly escaped string, which can be passed to request
now this for mysql only. it's up to parser sql driver to fix zeros properly

feature: oracle?..&LowerCaseColumnNames=0&ClientCharset=something column names got transcoded to $request:charset too [were only column data]

new: ^hash._keys[>>name<<] to call sole column of result

new: ^image.pixel(x;y)[(color)]
        get/set pixel color

beauty: image -- less checks [were ugly impl]

bugfix: ^string.left/right/mid with negative values now considered bad

bugfix: boolean output options now have unified defaults [after transform indent/omit-xml-declaration/standalone 'default' erroreously meant 'true']

bugfix: http://i.p.a.ddress gethostbyaddr added [on some platforms gethostbyname failed with such 'domains']

new method: ^string.trim[start|both|end[;chars]]        default 'chars' -- whitespace chars
finished

new method: ^string.trim[start|both|end[;chars]]        default 'chars' -- whitespace chars
  started, only start works

bug discovered: todo! xdoc::create+save != xdoc.transform&no<xsl:output>+save

just test:  file::load[text;http:// $.text utf-8 signature, line ends OK
http://i2/tasks/edit/?id=4629476859594276758

feature: ^process...{CODE}[now has options]
which are
        $.main[to what to rename @main]
        $.file[file, from which (user says) goes that CODE]
        $.lineno(line number in that file, where CODE starts. may be negative)

xhtml: &#9; as resutl of <xsl:attribute>[tab]  is OK

bugfix: <tag attr="&#digital_entity;" bug fixed again [same solution, mistanenly removed one line]

beauty: date::create[bad date time in daylightsaving hole] now throws error

beauty: coder can override save/file encoding by specifying it in xsl:output or in save/file hash-options

beauty: coder can override save/file encoding by specifying it in xsl:output or in save/file hash-options

beauty: xdoc.string now renderes in $request:charset, but in header puts $response:charset
todo: make that default only so that coder could override that

beauty: xdoc.string now renderes in $request:charset, but in header puts $response:charset
todo: make that default only so that coder could override that

beauty: xdoc.string now renderes in $request:charset, but in header puts $response:charset
todo: make that default only so that coder could override that

beauty: xdoc.string now renderes in $request:charset, but in header puts $response:charset
todo: make that default only so that coder could override that

feature: reimplemented xhtml output, see /.xsl for sample