[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-nio-dev
Subject:    Re: RFR: 8229887: ZIP FS cannot replace an entry in a that was added via the STORED method
From:       Martin Buchholz <martinrb () google ! com>
Date:       2019-08-24 22:54:59
Message-ID: CA+kOe08pGT7-yKETms4he-KzZPDgYc2_ihnJ1OXh92yn_Zyyrw () mail ! gmail ! com
[Download RAW message or body]

On Sat, Aug 24, 2019 at 1:31 AM Alan Bateman <Alan.Bateman@oracle.com>
wrote:

> On 22/08/2019 16:49, Lance Andersen wrote:
> > Hi Martin,
> >
> > Yes there is additional work to do both WRT documentation and testing.
> >  We will continue to improve the documentation  and the test coverage
> > as we move forward.
> >
> Martin - are there folks in Google using zipfs in interesting ways that
> could be turned into tests? One long standing need, shared with the
> java.util.zip implementation, is automated interop testing with other
> zip tools where the spec might be interpreted differently (time stamps
> and zip64 are two areas where these type of issues can bite).
>

Zip implementations in the JDK:
ZipInputStream/ZipOutputStream
ZipFile
zipfs
libjli/parse_manifest.c
... and probably more

We don't encourage use of zipfs at Google, but people will use it when
convenient.
If all you want to do is replace a single entry in a zip file, it's hard to
beat zipfs for convenience.
It's hard work to write a program that makes small edits to a zip file
while leaving it mostly intact.
The recent undocumented switch of zipfs from STORED to DEFLATED is not nice
to users.

My test test/jdk/java/util/zip/EntryCount64k.java compares 3 different ways
to read a zip file, but zipfs is not one of them :-( . I'd  also like it to
try unzip or zipinfo if they appear to be installed on the system.

[Attachment #3 (text/html)]

<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" \
class="gmail_attr">On Sat, Aug 24, 2019 at 1:31 AM Alan Bateman &lt;<a \
href="mailto:Alan.Bateman@oracle.com">Alan.Bateman@oracle.com</a>&gt; \
wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 22/08/2019 16:49, \
Lance Andersen wrote:<br> &gt; Hi Martin,<br>
&gt;<br>
&gt; Yes there is additional work to do both WRT documentation and testing. <br>
&gt;   We will continue to improve the documentation   and the test coverage <br>
&gt; as we move forward.<br>
&gt;<br>
Martin - are there folks in Google using zipfs in interesting ways that <br>
could be turned into tests? One long standing need, shared with the <br>
java.util.zip implementation, is automated interop testing with other <br>
zip tools where the spec might be interpreted differently (time stamps <br>
and zip64 are two areas where these type of issues can \
bite).<br></blockquote><div><br></div><div>Zip implementations in the \
JDK:</div><div>ZipInputStream/ZipOutputStream</div><div>ZipFile</div><div>zipfs</div><div>libjli/parse_manifest.c</div><div>... \
and probably more</div><div><br></div><div>We don&#39;t encourage use of zipfs at \
Google, but people will use it when convenient.  </div><div>If all you want to do is \
replace a single entry in a zip file, it&#39;s hard to beat zipfs for \
convenience.</div><div>It&#39;s hard work to write a program that makes small edits \
to a zip file while leaving it mostly intact.</div><div>The recent undocumented \
switch of zipfs from STORED to DEFLATED is not nice to \
users.</div><div><br></div><div>My test  test/jdk/java/util/zip/EntryCount64k.java \
compares 3 different ways to read a zip file, but zipfs is not one of them :-( . \
I&#39;d   also like it to try unzip or zipinfo if they appear to be installed on the \
system.</div></div></div>



[prev in list] [next in list] [prev in thread] [next in thread]