[Issue 1832] reading/writing an archive causes data loss; std.zip horribly broken

[d-bugmail at puremagic.com]

https://d.puremagic.com/issues/show_bug.cgi?id=1832



--- Comment #10 from Andrea Fontana <advmail at katamail.com> 2013-12-06 01:22:46 PST ---
(In reply to comment #9)
> (In reply to comment #8)
> > This is another bug. m1 is owned by a1. m1._compressedData refers to data
> > inside a1. So you're mixing data from two different objects/archives. I don't
> > think it's a good idea. When you build a2, it tries to read compressed data
> > from a2 data, but it should read from a1.
> >
> I see.
> 
> > There's a simply workaround, but again, there's a problem on library design. If
> > m1 is just compressed, we can read content it has inside _compressedData, if
> > any. It filled it before with right reference to a1 data. This solve your bug,
> > but probalby don't solve the mixing problem and of course it's not an elegant
> > solution!
> >
> I think m1._compressedData is simply a slice to (constant) GC memory which was
> allocated when reading a1. So here shouldn't be any problem to copy this data
> into a new zip archive.

Yes it is. But we're going to use data compressed/stored in/owned by another
archive. That seems a mistake to me. I know we can do it, of course. Tonight
I'll fix it then.

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------