Message 199480 - Python tracker

Message199480

Author	pitrou
Recipients	barry, christian.heimes, kristjan.jonsson, pitrou, ronaldoussoren, serhiy.storchaka, vstinner
Date	2013-10-11.14:34:55
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1919290567.69360662.1381502068060.JavaMail.root@zimbra10-e2.priv.proxad.net>
In-reply-to	<1381500899.49.0.0614432256474.issue19219@psf.upfronthosting.co.za>

Content
> I ran into this when developing version 3. The reason is that while > the marshal format includes the version information in its header, > it isn't actually verified on loading. IIRC. You specify the > expected format to the function, or something like that. So, if you > don't do this, you get errors when loading previously generated .pyc > files. Exactly (I also tried this :-)). The problem is the version number is outside of the marshal format: e.g. it's in the pyc file header. When freezing a module (see frozen.c), the version number isn't included. So our freedom is quite limited here: we have to support legacy frozen modules.

Content

> I ran into this when developing version 3.  The reason is that while
> the marshal format includes the version information in its header,
> it isn't actually verified on loading. IIRC.  You specify the
> expected format to the function, or something like that.  So, if you
> don't do this, you get errors when loading previously generated .pyc
> files.

Exactly (I also tried this :-)). The problem is the version number
is *outside* of the marshal format: e.g. it's in the pyc file header.
When freezing a module (see frozen.c), the version number isn't included.
So our freedom is quite limited here: we have to support legacy frozen
modules.

History
Date	User	Action	Args
2013-10-11 14:34:55	pitrou	set	recipients: + pitrou, barry, ronaldoussoren, kristjan.jonsson, vstinner, christian.heimes, serhiy.storchaka
2013-10-11 14:34:55	pitrou	link	issue19219 messages
2013-10-11 14:34:55	pitrou	create