Todo List
core_util
- see http://www.phpclasses.org/browse.html/file/1366.html
- document the possible return values.
- i bet this method could be optimized alot using regexp.
- maybe preg would be faster here than converting. dunno. at the moment i'm happy.
- maybe make use of the new php function array_rand() internally.
- implement $startPos feature if needed.
- optimize this shit. imagine a really huge $haystack and a small $needle. all $haystack is converted to lowercase. only the strlen($needle) chars would need to be converted. and passed over. same for endsWithI. better use regexp here.
- this function is not tested enough/has some new code. don't use it in a production environment. test it before!
- make use of the param char $thousandsSeparateChar
- this function is not tested enough/has some new code. don't use it in a production environment. test it before!
core_net
- implement this, add option to send user-defined cookies
- i think the first element here (*) makes it accept everything anyway...
- implement options 2 and 3.
- implement this, add option to send them again, and to send user-defined cookies
- support address lists (as specified in the rfc).
- add HTTP_ACCEPT_ENCODING HTTP_ACCEPT
this is a multiline description: - problem: if you connect to a host, then grab a few files, then disconnect, only
the 1st fetch will succeed. for the 2nd, 3rd ... you can send your
request without errors, but there's no reply (eof or error) from the
server => connection is lost somehow?
- reason: probably this is a windows problem only. the manual for fsockopen() sais
param "double timeout" is not available on all systems, but doesn't mention
on which it's not. it if's ignored, this would be the reason for my problem.
- solution: the param $tryReconnect in fetchPage() helps. it reconnects once each time
a conn is fucked/lost somehow as in this problem.
- comments: of course this is not satisfactory, because if you connect, fetch a lot,
then close, the connection has to be reopened for each fetch request.
but at least the coder doesn't need to worry.
hint: maybe you are just looking for: $file = fopen("http://www.php.net/", "r");if you're looking for information about the http protocol, try these, have fun :) Hypertext Transfer Protocol HTTP/1.1 ftp://ftp.isi.edu/in-notes/rfc2616.txt Obsoletes: 2068 Hypertext Transfer Protocol HTTP/1.1 ftp://ftp.isi.edu/in-notes/rfc2068.txt Hypertext Transfer Protocol HTTP/1.0 ftp://ftp.isi.edu/in-notes/rfc1945.txt Upgrading to TLS Within HTTP/1.1 ftp://ftp.isi.edu/in-notes/rfc2817.txt Updates: 2616 dependencies: Net/Bs_NetApplication (which uses Net/Bs_SocketClient and Net/Bs_Url)
- maybe replace the 2 Bs_SocketClients with an array of 10? was a sugguestion from sam.
- Prevent SID hijacking. If we reuse a previously created session by getting the sid from the querystring (and not from a cookie), we have a possible hole. Someone might send his url (with sid included) to someone else, and this person pastes it into the browser. See the problem. So what we do is check the class c network of the ip address (proxies might switch ip's). This still leaves room for ppl trading url's in the same company/using the same isp, but what can we do? Check browser string? hmm maybe. Well this is internal anyway and you shouldn't care.
- Message-Id: <20010913092203.45C8.SAM@gibtsnicht.xx>
- MIME-Version: 1.0
- check from time to time setBlocking() (if nonblocking is possible on win now)
- fix the problems listed above
- ecg
- allow full url's (with http://www.domain.com part)
- finish the descriptions for all codes, or decide to kick them out.
- what happens with dir- and file links?
- what happens with dir- and file links?
- test this code
- check the syntax of $param, or offer other formats to be given and convert them.
- check if php's getMxRr() and checkDnsRr() OR exec() are fixed now for win and update the code/manual.
core_gfx
- maybe add a profileDir var?
- finish implementation, currently only works with dataCsv.
- code support for all options. currently only 'month' and 'day' are supported. see _groupDataByInterval() and _fillDataGaps().
core_auth
- when the user gets to the login form, include the url he's coming from so after a successful login we can redirect him to where he was. just like we do it in the browscap class.
- make use of numbers and special characters.
- check Bs_Cug->_validateLoginData(), not sure where to implement this functionality.
core_html
- add security stuff because of the eval.
- make use of this or kick it out.
- make use of this or kick it out.
- add security stuff because of the eval.
- add security stuff because of the eval.
- add security stuff because of the eval.
- enforce is the wrong name for this. change it to $overlook. update text. 2002/11/10 --andrej
- the gui only supports string, not vector.
- all (that won't work, as far as i see it. --andrej)
- rethink the available modes.
- not implemented in the gui. difficult.
- not implemented in the gui. difficult.
- the gui only supports string, not vector.
- not implemented in the gui.
- not implemented in the gui. difficult.
- not implemented yet. could be needed somewhen.
- i think we should use the $_SERVER['DOCUMENT_ROOT'] or so. at least that var should be used.
- name[] _enforce won't work with arrays.
- not satisfied with the params, how the information is passed.
- add/recode the style stuff.
- finish this method, especially implement xml.
- use the explodeKey param here. dunno why that's not implemented.
- only send the enhanced button html code if the client supports it (ie4+)
- use the explodeKey param here. dunno why that's not implemented.
- i'm sure this won't work for select fields with a multiple selection. need to overwrite it. also the value has to be tag-encoded, i bet.
- i have moved this method to bs_textutil. we should use it instead to avoid double-maintenance.
- check from time to time if this method is implemented in php now. php4.3: not yet. --andrej
- use $paramValue, see Bs_FormField->inputManipulate()
- use ini-files instead of hardcoded php code. makes it much easier to extend.
- finish this code. it works fine but only if everything's right the first time.
- It works, but not very efficent.
- why is the inihandler not used here? i think this code could be replaced. i think the inihandler did not exist at that time.
- only works with insert yet, not update.
- this method is called from somewhere when the elementcontainers are not ready yet. maybe fix this, altough it doesn't break the code, it's just unclean.
- i have to do something with the return if it's false. log, whatever. don't know what's best at the moment, leaving it for later.
- finish this code. it already adds missing fields, but doesn't alter fields that need to be changed.
- security hole here. we should use the concept of evalWrap(). i want to have it executed here and not there so we're in this object's scope.
- security hole here. we should use the concept of evalWrap(). i want to have it executed here and not there so we're in this object's scope.
core_date
- maybe we better use the bs_kb.LangMonth table?
core_db
- take care, php's dl() function has been deprecated and is likely to be removed in php 4.1 or 5.
- his class is accurate for MsSQL __version__.
- please check the following methods to be up to date:
2do > fieldFlags(), hasFieldFlag(), fieldType(), getTableProperties(), getTableInfo(), getTableType(),
2do > getIniVar(), getStatusVar()
Make sure the existing code doesn't break because of your changes! thanx. - recode some functions more down the file for mssql (taken from mysql).
- code the transaction stuff
- implement the stored procedures functionality.
- i wonder if we should move more logic into bs_db.class (that is the same in the mssql and mysql implementation).
i don't want to copy the same code blocks over and over again for every db vendor. but then how many impl. will
we write?
dependencies: Bs_Db
- we could optimize this, modifying the query to do "SELECT COUNT(*)".
- check if that really works. and what about \n etc? see http://www.zend.com/codex.php?id=204&single=1
- check documentation above and the return values. maybe test a view.
- should we rename this to freeStatement here? think so.
- code and test this as soon as php405 is out.
- rewrite code for mssql.
- code and test this as soon as php405 is out.
- code and test this as soon as php405 is out.
- code and test this as soon as php405 is out.
- rewrite code for mssql.
- code and test this as soon as php405 is out.
- rewrite code for mssql.
- code and test this as soon as php405 is out.
- code and test this as soon as php405 is out.
- rewrite code for mssql.
- code and test this as soon as php405 is out.
- i bet there are a few bugs in the way the data array gets built. have to test this one. need to create a few test-db-tables for that to see how mysql returns the values.
- decide about return value 0/false/null/exception.
- implement this. currently returns a not-capable exception.
- add regexp for syntax check.
- add regexp for syntax check.
- add regexp for syntax check.
- rewrite code for mssql.
- implement this. currently returns a not-capable exception.
- code is not done, see yourself. currently only adds fields. maybe we should also create the table if it does not exist at all.
applications_debedoo
- it's not really an overview setting, it's used for everything. rename it? recode stuff? i leave it by now. --andrej
- all (not sure if this one gets used at all in the future.)
core_file
- Things like open, close, readLine, read, create, write, append
- Slash-handling (espacilly the '/' and '\' distinction)
- remove empty subdirs afterwards? a prune() method would be neat (that dives into a dir and removes empty subdirs).
- should we return '6' instead of 'word6', the number only?
core_lang
- make use of the $security param.
- return a bs_exception if the security check complains or a parse error occures while evaluating $phpCode.
- add something to disallow output (echo, print, escaping code).
- if we get a relative url we have to make it absolute, otherwise some clients may not understand it.
plugins_instanthelp
- implement language handling
- change the default lang handling, i think.
applications_imagearchive
- allow multiple file uploads at once as defined in getFileUploadForm().
plugins_indexserver
- support "*.domain.com" (or is this supported using "domain.com"?)
- check if profile exists, and return FALSE if it does not instead of throwing an exception.
- if $lang is an array then decide which to use, or assign more than one. do something.
applications_cms
- make use of the $highlightCurrent param
- implement language dependency somehow
- add the language code to each fucking url. 3rd level domains would be easier like fr.domain.com/ so you could just link to /someshit/ instead of /fr/someshit/ ...
- implement cache (make use of $useCache param)
- finish code
- add support to check versions like for javascript 1.2+
- add support to check versions like for pdf5+
core_text
- document this data structure
- currently only reads the languages, should be able to write also. (implement in Bs_IniHandler)
- finish code with char replacements.
core_storage
- finish this code. does not work yet, i think. or use the mysql implementation instead.
- finish this code. it already adds missing fields, but doesn't alter fields that need to be changed.
plugins_onomastics
applications_simplequiz
applications_smartshop
- implement $format support.
applications_websearchengine
- support "*.domain.com" (or is this supported using "domain.com"?)
- this seems to not work. check this.
- finish code with 'preg' and 'ereg'
- check if profile exists, and return FALSE if it does not instead of throwing an exception.
|
|