|
Nick
nick at dd.revealed.net
Tue Aug 17 11:14:39 EDT 2004
I don't know how mutable req is, but couldn't you replace write with a StringIO write function early in the execution? Nick Gregory (Grisha) Trubetskoy wrote: > > Yes, but Iker's problem is that it's not that simple to use with PSP, > which uses req.write()... > > On Tue, 17 Aug 2004, Nick wrote: > >> It is simple enough to do your own buffering with a CStringIO, which >> is relatively fast and file-like, and then sending the resulting >> buffer at the end of your handler to req.write. >> >> Nick >> >> Gregory (Grisha) Trubetskoy wrote: >> >>> >>> By "getting back" I meant having any access to it. Once it's written >>> out, the data becomes owned by the next module/filter in the filter >>> chain. The only way to do this would be for mod_python to do its own >>> buffering, which it does not. >>> >>> Grisha >>> >>> On Mon, 16 Aug 2004, Iker Arizmendi wrote: >>> >>>> I don't need to get it back, I just need to clear it so >>>> I can send just the error page without any dangling output >>>> from the faulty page. I guess one solution is to write the >>>> output to a buffer in Python code and only call req.write >>>> once all processing is complete. But this would eliminate >>>> the need for PSP (unless PSP offers a way to direct it's >>>> output to my own buffer). >>>> >>>> Iker >>>> >>>> >>>> Gregory (Grisha) Trubetskoy wrote: >>>> >>>>> >>>>> Hmm... The buffer that buffers output is controlled by Apache, and >>>>> AFAIK once something's been written to it, you cannot get it back. >>>>> >>>>> Grisha >>>>> >>>>> On Mon, 16 Aug 2004, Iker Arizmendi wrote: >>>>> >>>>>> One problem I've encountered with using set_error_page() >>>>>> for error handling is that often the faulty page has already >>>>>> generated some output when the error is encountered. This >>>>>> results in garbled pages with bits from both the error-raising >>>>>> page and the error-catching page. >>>>>> >>>>>> Is there some way to get around this? >>>>>> >>>>>> Regards, >>>>>> Iker >>>>>> >>>>>> >>>>>> Iker Arizmendi wrote: >>>>>> >>>>>>> Thanks Greg. Is there a method that clears the >>>>>>> current output buffer as well? >>>>>>> >>>>>>> Gregory (Grisha) Trubetskoy wrote: >>>>>>> >>>>>>>> >>>>>>>> Yes, by default PSP does not flush the output buffer until the >>>>>>>> very end which looks like what PHP ob_start func enables. If you >>>>>>>> want to force a flush from within PSP, you can always call >>>>>>>> req.flush(). >>>>>>>> >>>>>>>> Grisha >>>>>>>> >>>>>>>> On Fri, 13 Aug 2004, Iker Arizmendi wrote: >>>>>>>> >>>>>>>>> Does the PSP handler support output buffering? >>>>>>>>> Something along the lines of ob_start under PHP? If >>>>>>>>> explicit support is not there, is there some other >>>>>>>>> way to accomplish the same thing? >>>>>>>>> >>>>>>>>> Thanks, >>>>>>>>> Iker >>>>>>>>> >>>>>>> >>>>>> >>>> >>>> -- >>>> Iker Arizmendi >>>> AT&T Labs - Research >>>> Speech and Image Processing Lab >>>> e: iker at research.att.com >>>> w: http://research.att.com >>>> >>>> >>>> _______________________________________________ >>>> Mod_python mailing list >>>> Mod_python at modpython.org >>>> http://mailman.modpython.org/mailman/listinfo/mod_python >>>> >>> _______________________________________________ >>> Mod_python mailing list >>> Mod_python at modpython.org >>> http://mailman.modpython.org/mailman/listinfo/mod_python >> >> >> _______________________________________________ >> Mod_python mailing list >> Mod_python at modpython.org >> http://mailman.modpython.org/mailman/listinfo/mod_python >> > _______________________________________________ > Mod_python mailing list > Mod_python at modpython.org > http://mailman.modpython.org/mailman/listinfo/mod_python
|