CVS log for parser3/src/classes/xdoc.C

Up to [parser3project] / parser3 / src / classes

Request diff between arbitrary revisions - Display revisions graphically

Keyword substitution: kv
Default branch: MAIN

Copyright year updated, websites links changed to https://

params.as_file_name(N) added and used instead of params.as_string(1, FILE_NAME_MUST_BE_STRING) where appropriate (implements feature #1121)

cleanup after classes.awk deleted

Copyright year updated

optimization: Value::as() removed as dynamic_cast is faster, Value::is() is still required for is operator.

copyright, authors actualized

hash no longer is accepted as file parameter (fixes bug #1218)

Copyright year updated

xsltStylesheet_auto_ptr removed, as xsltFreeStylesheet has side effects on original doc.
compiled stylesheet no longer cached as with libxml2-2.9.9 it was not updated after
xdoc modification (part 3 of test 362). (related to bug #1108)

Stylesheet_connection_ptr changed a bit

r.absolute -> r.full_disk_path

xmlHashScanner declaration changed, using typecast for more compatibility (related to feature #1136)

XML_PARSE_OPTIONS (XML_PARSE_DTDLOAD | XML_PARSE_NOENT) added to fix entities load, (related to feature #1181)

xmlReadMemory/xmlReadFile now used with XML_PARSE_HUGE option passed by default (implements feature #1181)

regular new/delete no longer used in our code, stubs defend from accidental use (implements feature #1170)

Copyright year: 2015->2017

xsltParseStylesheetDoc result now cached as xdoc is modified after it (fixes bug #1108)

cleanup: r.write_pass_lang, r.write_no_lang -> r.write (implements feature #1091)

Temp_lang removed, write_no_lang now almost equal write_pass_lang (related to feature #1091)

charsets -> pa_charsets, vs2015 warning war

charsets.get now does str_upper inside (related to feature #1061)

xnode_class initialization moved to xdoc.C as under FreeBSD it was inited after and inheritance failed (related to issue #1051)

base -> methoded_donor for static classes, type() moved from VStateless_class to VClass and Methoded (implements feature #1051)

optimizaion: *_base_class removed (related to feature #1051)

Copyright year updated

explicit String::Body(char) calls removed; String::Body(String::C) constructor added (related to bug #957)

new(UseGC) replaced with new, malloc/strdup/free replaced with pa_ version and checks for accidental use of non-pa_ version added (related to issue #1018)

libxml 2.9.1 compatibility with define LIBXML2_NEW_BUFFER check (closes issue #943)

XmlException now calls fixUTF8 if source charset is UTF-8 (closes issue #938)

cstr_to_string_body_untaint should be called with charsets for possible uri language and connection for possible sql language.
related to issue #857

output_options returned and used (bugfix for feature #265)

feols_normalized -> fis_text_content
set_binary added

- ^file::create[text;content] EOLs in content now is normalized. ( new feature: #815 )

- MethodParams::as_hash is optimized and improved (whitespaces are allowed as empty options)
- MethodParams::as_table is added
- above methods are used for parsing methods' options ( new feature: #9 )

ident now works under Linux + ident displays filenames (closes issue #818)
Copyright updated

- constructor ^file::create[mode;filename;content[;options]] now accepts binary mode and file-content
- new constructor's format: ^file::create[string-or-file-content[;$.name[filename] $.mode[text|binary] $.content-type[...] $.charset[...]]] ( new feature: #65 )

- charset.isUTF8 is used instead of comparation charset name with string "UTF-8" ( bugfix: #759 )

- new option was added: ^xdoc.file[$.name[имя файла]] (new feature: #622)

- $.mode must be set for newly created file

- validation of some input params was added. it isn't possible not wo create xdoc with invalid tagName. ( bugfix: #160 )

- code cleanup (unused output_options were removed)

- now json:string can serialize xdoc-objects. options (the same as ^xdoc.string[]) could be specified in $.xdoc[] ( new feature: #265 )

- file_write accepts Request_charsets and asked charset and cound transcode content before writing

- new classes implementation

- some code cleanup

- cstr(params) and cstrm(params) were deleted. taint_cstr[m] and untaint_cstr[m] should be used

- optimizations and simplification of string.cstr(...)

- create object for storing object's fields only for parser objects

no message

- result optimisation
- helper_length parameter removed from string constructors

- normalizing todo-comments: '@todo' now

- more duplicated exception text strings were removed

- some duplicate exceptions' text strings removed

- "parser.runtime" strings were removed

- another attempt

- roll back last changes for a while

- ^xdoc::create[$file] added.

+ ^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=]

. started killing gdome

2005

. 2005 ;)

+ 3.2.0 beta: merged from dynamic_fields_join

joined_dynamic_fields: started

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]

change: ^xdoc.string now outputted as-is  [tainted as-is]

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

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

beauty: ugly code removed, thanks to egr for pointing that out

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

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: xhtml doctype-public/system forcly turns off omit-xml-declaration

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

beauty: needless helper class removed

beauty: method=html outputxmlbuffer not copied extra time

bugfound: meta again, now somewhere in xhtml part of libxml

new: xhtml now works [libxslt prevented libxml to do the work]
<xsl:output
	method="html"
	doctype-public="-//W3C//DTD XHTML 1.0 Strict//EN"
	doctype-system="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"
	/>

beauty: no <meta ... charset generated

change: .xsl.stamp now not used, introducing dependencies timestamp check

2004

bugfix: libxml: FOR UTF-8 TOO russian letters in attributes or documents-results of transform now not &#04xx;

bugfix: libxml: russian letters in attributes or documents-results of transform now not &#04xx;

bugfix: when stylesheet contains error, gpf

bugfix: taint uri under ^xdoc::create{now works}

beauty cancel: gcc does not understand that

beauty: removed redundant ctor call [relying on implicit String::Body(cstr) call]

$status:sql.cache back operational

bugfix: typo

replaced original open mech of xdoc::create & co to use libxml open, so that this would work for root document
todo: safe mode checks to libxml: ideally to some 1 point

more warnings --

more warnings --

static const char * const IDENT

turned on warnings level4 on all projects (except libltdl=off & libsdbm=level3)
found several unitialized vars

resurrected: pool idea. now only for destructing objects at request processing end

XmlException changes

hashfile: started again with sdbm from apache

string_fragments_to_cord merged to HEAD

v() functions of String::Body, String::Languages and String itself
StringBody->String::Body

copy/paste comment bugfix

new DOM2 methods from Alexandr Egorov <egr> (all?)
xdoc:
createAttributeNS
createElementNS
xnode:
getAttributeNS
setAttributeNS
removeAttributeNS
getAttributeNodeNS
setAttributeNodeNS
hasAttribute
hasAttributeNS

moved tempate_gc to HEAD

merged fix for
found very old xml (dom) bug:
were passing domString objects and later ERROREOUSLY freed them

found very old xml (dom) bug:
were passing domString objects and later ERROREOUSLY freed them

started http://
  $.charset[] param
and
  http response charset detection
just compiled.
todo:test

memory allocation func bugfix
http://i2/tasks/edit/?id=4499303470368629745

MethodParams& now [methods without params receive zero reference, but they are expected not to look there]

MethodParams& now [methods without params receive zero reference, but they are expected not to look there]

removed ref leak in xdoc::create/load

memory bug debugged down: gdome uses glib memory, and stores last pointer to xmlDoc there, gc misses that and collects valid memory
fixed by remembering xmlDoc from dom object in xdoc.
todo: do something with premature free of xdoc with xnodes/node values referring into it

started xml lib memory debugging, simple checks failed:
it seems that library does realloc(bad ptr), and recording those ptrs in heap [for debugging] was bad:
heap situation changes = everything works fine

XML memory funcs to GC works in debug, but fails in release
todo: fix that

forgot to merge XML memory funcs replacement from gc branch

Array::put(index, T>>&<< removed
after Stack::pop wiping removed [moved to separate func]
other Array & removed

fixed transform params2

fixed transform params

removed StringBody(0) ambiguilty, introducting static StringBody::Format(int)

xdoc.C compiled
todo: xnode.C / finalizers

started XML

pa_vmail.C compiled

hash.C compiled

file.C
table.C compiled

pa_request.C compiled

introducing StringBody [C++ CORD wrapper]

more tests OK

untaint.C 30%,
pa_common.C [done, without http for now]

started test

lots of replacements,
todo:we can ignore lang in cmp and pos really,
		but would split properly!

started porting to gc:
PA_Object done
lots of replacements also

libgc attempt

forgotten: global xdoc when ndef XML

xml: test 108 transform with params [bugs fixed]

xml: test: 107 bug fixes

test: 107 bug fixes

xml: linked

xdoc.C compiled

xdoc.C 50%

methoded_array now contains all Mxxx classes
SAPI::get_env now returns info on pool
more tests

fixed all cstr's that must use pool [as the one in ^process]

string cmp fixed

fixed all remained bugs of unitialized simple-typed field

fixed more leaks resulted from pooled::malloc calls

parser3.exe - 0 error(s), 0 warning(s)

image compiled

date compiled

int compiled

void  compiled :)

grammar compiled

started main: compile_tools

2002->2003

doc->URL on xdoc.load set correctly now

checked http options [invalid onces now reported]
made default
user-agent: paf

table/xdoc/file::load now understand http:// prefix and additional params, sample:
$rates[^xdoc::load[http://www.cbr.ru/scripts/XML_daily.asp?date_req=02/03/2002;
   $.USER-AGENT[parser3]
]]

xml->parser strings now have origin, which points to place where value left xml library and came to parser: place of dom field extraction/call

xdoc::create[name] now sets $request:charset as internal xdoc encoding, and after decoding attributes set by dom functions now encoded OK, not as digital entities

empty transform result, being taken as file now returns empty file, not raises stupid error about "stat-ed file"

pa_request contexts made privated, plus get_{self/method_frame} to read.
Request_context_saver used in ^try to save flang too.
[were not saved]

table::sql  options
table::create  copy options
options checked, wrong option now fatal error

xdoc getElementsByTagName, ...NS
overriden, work now

instead of type() checking everywhere used Value.as
now user descendants can be used in params.

VObject::as_*, is_defined now taken from bases.
xtable(table)
^if($xtable) now OK

$Date: now$

ident.C* removed

merged latest updates to head

^xdoc::create[[uri]]... base uri for document being created, imports and other relative file names would be relative to this uri
default uri=path_translated

xsltParseStylesheetDoc bug workaround

^xdoc.transform[xdoc <<can be now

xdoc(xnode) now fully - fields&methods

killed Request::process() wrapping VString(String) in case we need only String

xslt params made literal

xslt params fixed

removed redundant pool param

assigned exception types

getElementsByTagName and *NS moved to node [element]

name spelling

2002

xdoc::set obsolete now, now xdoc::create, with both sences

now compiles under win32

acconfig.h changed [comments before #undefs removed]
those comments were a mistake, and prevented libltdl from function properly

few forgotten ifdef XML added

xdoc.importNode 0

xdoc::set now untaints OK

removed redundant pool param to some pa_common funcs

gdome_xml_doc_get_xmlDoc handy macro, which is strangly abscent in dome lib

xsltSaveResultTo checked <0 response

checked empty response

xdoc::load/set entities substituted.
	// 2. when dom tree with entites goes under transform text nodes
	// got [erroreosly] cut on first entity occurance

%.20E now default float format
this is for regretful twice-converting here:
$a(double value) when they write double falue  they convert it to string first,
thus %g were losing precesion.

todo: think up some way to remove double->string string->double twice-converting

xmlSubstituteEntitiesDefault(1) now global, reasons in comment inside

xsl stylesheet load: parsed entities, this helps compiling stylesheet properly.
example:
this refused to work,
            <b>&laquo;<xsl:value-of select="@title" />&raquo;</b>
stranly worked only this way:
            <b>&#171;<xsl:value-of select="@title" />&raquo;</b>

docs says "set it to 1", never going into details, so I decided to try NOT to do that

xdoc::load error source = filespec

transform error source now stylesheet_filespec

autoconf gnome

xslt generic error 1

auto closers to sql_connection, stylesheet_connection
auto destroyers to remaining gnome objects

few comments

xdoc
  file
  save
  string

ready.
todo: error handling

xdoc save rewritten using xmlAllocOutputBuffer
less mallocs, more reallocs. less fragmented result

started output options parsing

method

xslt transform0.
TODO:
  use output options
  handle errors

^xdoc.file

GdomeDOMString_auto_ptr  c++ wrapper [calls refcounter]

xdoc
^set
^string with glib works 0

started glib transcoders

gnome libs just compiled in.
no refcounting
no xpath
no xslt
  yet

introducing Charset

rewritten auto_ptr part other way2, because of stupid gcc 2.96
/usr/include/g++-3/memory:40: candidates are: auto_ptr<_Tp>
&auto_ptr<_Tp>::operator= (auto_ptr<_Tp> &) [with _Tp =
FormatterListener]
/usr/include/g++-3/memory:48:                 auto_ptr<_Tp>
&auto_ptr<_Tp>::operator= (auto_ptr<_Tp1> &) [with _Tp1 =
FormatterListener, _Tp = FormatterListener]

rewritten auto_ptr part other way, because of stupid gcc 2.96
/usr/include/g++-3/memory:40: candidates are: auto_ptr<_Tp>
&auto_ptr<_Tp>::operator= (auto_ptr<_Tp> &) [with _Tp =
FormatterListener]
/usr/include/g++-3/memory:48:                 auto_ptr<_Tp>
&auto_ptr<_Tp>::operator= (auto_ptr<_Tp1> &) [with _Tp1 =
FormatterListener, _Tp = FormatterListener]

removed extra inc

freed up listener

lang list updated

cached xslt [removed nocache forcing]

<xsl:output
output options made identical to xsl:output attributes
exclusion:
   cdata-section-elements
what xalan can do but that not used now:
   it can escape and remove CDATA sections

hash now does not containt 0 values. put 0 actullay removes.
no check on 0 value in foreach now required

xdoc::set{code} now untaints param properly

$status:db
fixed used not initialized in db_connection & db_table

$status:stylesheet

hash locking
disabled changing of hash inside of foreach

rolled back {} in xdoc.createTextNode for now

rolled back {} in xdoc.createElement for now

debuggging memory.
string::as_int/double now usually uses stack, not heap [200K out of 17M :)]
VString(String&) not copies reference [29M -> 17M]

xdoc createTextNode createElement UL_XML default. but entities still reparsed. todo: tothink

transform2 for parsed_source

transform2 for parsed_source

^hashfile.hash[]

ifdefs so it compiled without db

hashfile 0

xnode clone got freed

first c++ exceptions result: can free up xalandomstring resulting from pool::transcode

switched to c++ exceptions 1xml
configure fixed to exclude xalan-patch from non-xml compile

vxdoc set_document & ctor question "who owns document" solved

extra free removed

encoding in dom creating funcs

dom created by create can be transformed now

dom created by create can be transformed now

z

comment

xdoc create now uses XercesDocumentBridge, appendChild&co now works

xdoc::create[] now. and all dom tag names forced to XML lang

fiew leechy-found bugs in code&doc fixed

    !::create{qualifiedName}

    !::create{qualifiedName}

DOM1 major addtion, only few attributes/methods left

xalan&xerces multithread bug fixed. initialization&free on each thread!

moved xslt exceptions convertors out of Pool into Exception

xalan patches starting

xdoc:load exceptions intercepted [like in set]

parsedSource produced by nonstandard liaison freed up

changed xdoc:set to prevent SAXParseException from outputing to cerr.
lots of "todos"

started xml errorhandler.
fixed string.replace a little, read @test

xdoc.transform now

ctype with unicode 1. todo: no П entities on output somehow

letter body string passed AS-IS now

xdoc addencoding works! invented a way of user-configuring... todo: implement

X^.getElementsByTagName[tagname] = array of xnode
X^.getElementsByTagNameNS[namespaceURI;localName] = array of nodes
removed - regretfully not supported @ xalan. maybe someday would reimplement, if needed.

win32build system: solved long waiting anti-non-started-apache problem

win32 ident make system so not to rebuild library [ident.C.new]

z

xdoc&xnode now