CVS log for sql/oracle/parser3oracle.C

Up to [parser3project] / sql / oracle

Request diff between arbitrary revisions - Display revisions graphically

Keyword substitution: kv
Default branch: MAIN

Copyright year updated

Oracle 11.2 client reads enviroment on library load, thus env should be set before (related to issue #929)

bugfix: lt_dlinit() can fail, but lt_dlerror return 0, and we count 0 for success.

compiler warnings fixed:
-format '%u' expects type 'unsigned int', but argument has type 'long unsigned int'

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

fixes issue #451 - x64 compile errors

new in-parser quoting implementation (fixes issue #1)

- if limit/offset specified modify query (use ROWNUM in two different ways)
- with new option DisableQueryModification=1 in connect string this query modification can be switched off

*** empty log message ***

no message

- $.limit(0) fix + some transcode calls optimisation

- bug with incorrect processing escaped symbols in clobs was fixed

sql connect string, rsplit @

bugfix: empty bind param now passed OK [were not changing bind buffer value]

bugfix: output value of bind var now cloned out from bind buffers

no message

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

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

change: "returning" placeholders now bound by name
[preparing to general placeholders]

bugfix: connection struct allocated from gc mem, so that pointers to parts of connect url would be known to libgc [they were mistakenly freed]

bugfix: of several clobs per insert/update were properly processed only last, others became empty

bugfix: connection struct were allocated in services memory, which were connected with request

merged bugfixes from 3.1.2

beauty: merged from HEAD

cstrClientCharset localized

beauty: needless temp vars removed

beauty: static to hide some minor helper funcs from linker

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

style:

beauty: renamed cs to connection & class too

comment: about not needing to free handles

bugfix: fetch_buffers allocated with ::malloc so that they would not be collected by gc [freed at disconnect]

better: column fetch buffer cached as allocated [per connection]

bugfix: clob now read fully [were truncated]. oracle api is somewhat tricky

bugfix: LocalCharset now case insensitive

bugfix: if ?ClientCharset not defined typo error caused no-letters in result

feature: introducing ?ClientCharset=parser_charset
charset in which parser thinks client works

two bob in one operation never worked?

potential-bug fix: ODIDefine def field were not cleared
http://i2/tasks/edit/?id=4584951251464180531

added option: &LowerCaseColumnNames=0 [deafult=1]

bugfix: lowercase zero terminated

simplified tolower,
naming changed a little

merged bugfix: name were not zeroterminated

bugfix: name were not zero-terminated

fix: very old bug: wrong quote escaping

'/**xxx' situation were not handled

malloc_atomic changed to malloc

moved tempate_gc to HEAD

updated to interface v8

queries with more then one /**xxx**/  had more problems. fixed.

typo error from very start fixed
were problems with queries with more then one /**xxx**/

mistakenly patched HEAD branch. copying changes to template_gc one

2002->2003

no message

changed exception handling mech in sql handlers #2

changed exception handling mech in sql handlers

changed exception handling mech in sql handlers

ifdef HAVE_SETENV const char bug fixed

OCI_SUCCESS_WITH_INFO considered SUCCESS

OCI_SUCCESS_WITH_INFO considered OK

OCI_SUCCESS_WITH_INFO described

style changes

name spelling

2002

isspace

removed leading whitespace before comparing to 'select...

introducing estimating String::cstr_bufsize, 0

removed \ unescaping

removed \ escaping, not needed actually

removed return [enabled actuall disconnect]

found quote bug - wrong result

+= env param types now. for win32 PATH+=

prefetching 100rows+100Ksize

forgot ! for not null :)

null lob never cleared ptr&size = were copy of prev column

@todo ?objects=1 which would turn on OCI_OBJECT init flag

@test

setenv for env strings

column count bug [+1]

conf

no message

Initial revision