[reportlab-users] problem with PIL 1.1.7 and ReportLab 2.2

[Bill Janssen]

Adrian Klaver <aklaver at comcast.net> wrote:


> On Friday 11 December 2009 2:58:05 am Robin Becker wrote:

> > On 10/12/2009 18:57, Andy Robinson wrote:

> > > 2009/12/10 Bill Janssen<janssen at parc.com>:

> > >> Apparently in 1.1.6, PIL used to do a "load()" on the Image before

> > >> "split()", and in 1.1.7 it doesn't.  Looks like the workaround is to

> > >> call load() on the image before using it with ReportLab.  Or RL could

> > >> do this, too, just to be safe.

> > >

> > > Does anyone happen to know if calling load() before split() works in

> > > all the older versions of PIL out there?  I don't want to fix 1.1.7

> > > and simultaneously break support for some earlier 1.1.x...

> >

> > perhaps we should just check the version or for actual existence of the

> > loaded part before issuing an unwanted load.

> 

> Not to steal Bills thunder but it might be advisable to wait, see here:

> http://mail.python.org/pipermail/image-sig/2009-December/006023.html


Unfortunately, PIL 1.1.7 final is out, and widely distributed.  I'm sure
1.1.8 will eventually have a fix for this, but ReportLab will still
continue to break when combined with the buggy 1.1.7.  Just calling
load() in the ReportLab code before calling split() should work fine for
all versions of PIL.

Bill