commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Christian Grobmeier" <>
Subject Re: Compress vs. VFS
Date Sun, 06 Jul 2008 08:20:52 GMT
Hi Samuel,

> * I see that the
> Archive<
> >
> interface and the
> ArchiverFactory<
> >
> deal a lot with files.
> Wouldn't it be better for those APIs to only (mostly) rely on I/O
> streams? I believe the API shouldn't assume that a file system is
> available (when it makes sense). To cite one of my colleagues "files
> are evil" ;-) For example, I have a particular use case where I would
> like to extract a zip file from a within a war file. I obviously can't
> rely on files being available as nothing tells me whether the war is
> deployed exploded or not. Getting a stream to the given zip file
> however wouldn't be an issue?
> What do you think?
> * The Archive interface has the
> "getEntryIterator"<
> >
> method. However this method will not iterate through entries of an
> existing archive but only through added entries. I believe this should
> somehow allow me to iterate through all entries, existing and added.
> Again what do you think?

we allready thought about that too and started to redesign that api.
You can take a look at the new api with GIT here:
This API only uses streams at the moment. Later it may be discussed to
add the filebased classes as some kind of a helper somewhere.
If so, it's worth to think about your getEntryIterator idea. I will take it
my TODO list :-)

Best regards and thanks for you comments,

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message