Re: HTTP_Request cookies

On 6/6/05, Andrei Railean <arailean@squiz.net> wrote: > Hi All, > This is my first post to this list, so please don't be too rough. > > I've proposed a slight mod to HTTP_Request such that the cookies set > between redirects are remembered. > You can see the proposed fix in the bug report: > http://pear.php.net/bugs/bug.php?id=4505 > > Currently, if redirects are allowed by HTTP_Request, the redirects will > happen up to the maximum number allowed. The issue here is that each of > those redirects can set cookies, which are silently ignored by > HTTP_Request. This means that the final destination (redirect) might > not be correct because the intermediate redirects might base their > redirect header location on the cookies from previous redirects. Since > so many packages depend on HTTP_Request, I thought that it is important > to rectify this, even though the multiple redirect scenario with > cookies might not occur very frequently. > > I suspect package maintainers are probably busy so I submitted a > proposed fix. The fix uses Cookie_Manager - which is a part of > HTTP_Client package, which depends on HTTP_Request, so that introduces > a circular reference (somewhat). Don't know the politics of using code > from child packages used in parent packages, but that suggest that > Cookie_Manager might need to be independent of HTTP_Client and possibly > reside under the HTTP root. > Well....this is precisely what HTTP_Client is for. To simulate a normal browser (with regard to cookies and session). HTTP_Request isn't supposed to have any extra cookie ro session logic. If you move this from HTTP_Client to HTTP_Request then you might as well just merge the two packages (which is IMHO a bad idea). What happens if you use HTTP_Client to make this request? Does it deal with the cookies ok? If yes, then I suggest you use HTTP_Client for this (as you should be) and that the HTTP_Request maintainers may want to add a note to their docs about this type of situation. -- Justin Patrin