cPickle from 2.2 to 2.1
remco
remco at localhost
Mon Dec 15 06:48:44 EST 2003
More information about the Python-list mailing list
Mon Dec 15 06:48:44 EST 2003
- Previous message (by thread): cPickle from 2.2 to 2.1
- Next message (by thread): cPickle from 2.2 to 2.1
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
paul <paul.koelle at archit.uni-weimar.de> wrote: > >Thanks for pointing that out (seems rather obvious now ;). I thought >pickle is a fast and convenient method to store/restore data. Now I see >the drawbacks. These drawbacks (backward/forward incompatibilities) occur nearly everywhere. I've used the Berkley db module, when it got updated, i had to deinstall Python, reinstall an old version, check out all my data to another format, deinstall the older python, and reinstall the latest version. Next i had to create a new Berkley database and put the data back in. And i thought Berkley would never change again :)) >BTW: It was a hack to avoid the xml.dom.minidom/unicode/non ASCII >charset/codec/writer mess. Seems like I have to look at that once more. Therefore i use libxml2 instead of minidom, have once written a nice function that will quote all the unicode stuff you're problems will be over. On the other hand, you can't edit pickled data with text editor. With xml you can, so it's an advantage. If this doesn't matter: why not UUencode any unicode string when writing it to the xml? It worked for me.. And last but not least; there's allways the old .ini format 8) Cheers Remco Boerma
- Previous message (by thread): cPickle from 2.2 to 2.1
- Next message (by thread): cPickle from 2.2 to 2.1
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Python-list mailing list