Re: Deprecated functionality & fixes

On Sun, Aug 28, 2011 at 12:59 PM, Alexey Borzov <borz_off@cs.msu.su> wrote: > Hi Daniel, > > On 28.08.2011 14:09, Daniel O'Connor wrote: >> >> It's been a while since I looked at >> http://ollehost.dk/media/pear-bugs/ >> but >> Daniel C's recent emails to the pear-qa list prompted me to. >> >> Bascially; on the SVN versions of many packages we're in reasonably good >> shape when it comes to deprecated behaviour. We had several hundred uses >> of >> eregi related code; plus some magic quotes and other bits and pieces. >> >> When I looked again; I noticed HTTP_Request2 had re-adopted magic quotes >> functions - and for good reason when reviewing the bug ( >> http://pear.php.net/bugs/bug.php?id=16440). >> >> So... we've got 36 deprecated bits in a handful of packages - how to we >> hit >> 0? Are there some behaviours like what's seen in Bug #16440 we work around >> further; or draw the line and refuse to support? > > I have a feeling that "Deprecated stuff in HTTP_Request2" was extensively > discussed before (maybe it was off-list, though), but seems another round is > required. > >  1) get_magic_quotes_runtime() is NOT DEPRECATED. See here: > > http://www.php.net/manual/en/function.get-magic-quotes-runtime.php >  2) set_magic_quotes_runtime() is indeed deprecated and produces a warning > when called. HOWEVER, it is called in HTTP_Request2::send() if and only if > get_magic_quotes_runtime() returns true. >  3) THEREFORE, the only way to get a "deprecated" warning from HTTP_Request2 > is to use deprecated functionality beforehand (either enabling > magic_quotes_runtime in config file or calling > set_magic_quotes_runtime(true)). > > Frankly, I do not see the point in hiding this warning from the user with > either @ or using ini_set() instead (I also think the latter not showing a > warning is an oversight that may be fixed sometime). After all he is either > aware of the deprecated stuff and hides relevant errors or needs a heads up > with it. > > Of course, if some package uses deprecated functionality unconditionally > this should be fixed. > +1