|
A workaround at present seems though to be to pass in "req.form" explicitly
through the "vars" dictionary of user settings. This will the take precedence over that set up by psp.PSP(). It is still wasteful for psp.PSP() to process the form parameters a second time. settings.update({"form":req.form}) template = psp.PSP(req,filename=path,vars=settings) template.run() Adding more to this and why PSP should check for req.form and also set req.form
if it is the first to create the form object, one should look at the set_error_page() mechanism in PSP pages. If setting an error page, and it gets triggered, an additional instance of the PSP object is created, it will yet again create a unique instance of the form object if a reference to it exists in the page. Again, POST form parameters will not be preserved. Related to the form object, one also needs to be careful with the session object which is created by PSP run() method. If PSP pages were being created explicitly from a basic content handler or publisher, and the parent was creating a session object and the PSP code also tried to reference a session object, the code would deadlock on itself. Even if the parent didn't create a session, if the main PSP page and an error page both tried to access a session object, it would again deadlock on itself. Code should perhaps be something like: # does this code use session? session = None if "session" in code.co_names: if not hasattr(req,"session"): session = Session.Session(req) req.session = session # does this code use form? if "form" in code.co_names: if not hasattr(req,"form"): req.form = util.FieldStorage(req, keep_blank_values=1) # create psp interface object psp = PSPInterface(req, self.filename, req.form) try: global_scope = globals().copy() global_scope.update({"req":req.form, "session":req.session, "form":form, "psp":psp}) global_scope.update(self.vars) # passed in __init__() global_scope.update(vars) # passed in run() try: exec code in global_scope req.flush() # the mere instantiation of a session changes it # (access time), so it *always* has to be saved if session is not None: session.save() except: et, ev, etb = sys.exc_info() if psp.error_page: # run error page psp.error_page.run({"exception": (et, ev, etb)}) else: raise et, ev, etb finally: if session is not None: session.unlock() Whoops, got the following code wrong in a couple of ways:
global_scope.update({"req":req.form, "session":req.session, "form":form, "psp":psp}) Should be something like: global_scope.update({"req":req,"psp":psp}) if hasattr(req,"form"): global_scope["form"] = req.form else: global_scope["form"] = None if hasattr(req,"session"): global_scope["session"] = req.session else: global_scope["session"] = None Another thing to be considered is that if PSP object is used from external handler and session object is created by external handler, should the PSP class still be responsible for saving the session. Ie., should it do: if session is not None: session.save() elif hasattr(req,"session"): req.session.save() We could also solve the problem by modifying the FieldStorage and Session constructors, so that if the req object already has a 'form' or 'session' attribute, respectively, we just return its value :
class FieldStorage(object): def __new__(cls,req,*args,**kw): try: return req.form except AttributeError: req.form = object.__new__(cls) return req.form # the rest of the object is as before This way, we can build util.FieldStorage(req) as many time we want, we will always get the same instance. The comment above is not really true. We have to modify __init__ as well so that the instance does not get initialized more than once.
Explicitly checking for req.session within PSP.run would also fix an unreported bug, where the calling code uses a Session class which is different from the default.
def handler(req): ... sess = Session.FileSession() template = psp.PSP(req,filename=path,vars=settings) template.run() And in class PSP we have: def run(self, vars={}): code, req = self.code, self.req # does this code use session? session = None if "session" in code.co_names: session = Session.Session(req) Currently Session.Session uses dbm as the default session class, so in the above code snippet we end up with 2 different persistent stores for the session data. Have something else to question in the run() code above. That is, I think
that doing: if session is not None: session.unlock() is questionable and probably shouldn't be done. It should be done in the cleanup handler registered by the session object itself when it is created if I remember things correctly. Especially if we start caching session object in the req object, by unlocking it, even if it was created by the PSP object, it makes the session object unusable in code that may follow explicit use of PSP object in a handler, or in a subsequent log handler. Some of the unfinished session object issues in this entry split out into
This entry should only be used for req.form/PSP interaction issue. Fixed, with PSP code now using FieldStorage object cached as req.form and left there by prior executing code. Allows for better interaction between publisher handler and PSP pages, plus PSP error pages.
All the session issues described in this issue are dealt with by other problem reports. |
||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Improvement
Closed
Minor
Passing req.form into psp.PSP().
for the presence of req.form as may have been left by a prior parsing
of the form parameters. This would be better than a "form" parameter
as it would be transparent.