On 17/02/12 15:00, Lewis John Mcgibbney wrote:
Hi Andy,
I'm kinda loosing the track with this one. Working on this comment
On Thu, Feb 16, 2012 at 10:57 PM, Andy Seaborne<[email protected]> wrote:
It seems to be used by edu.uci.ics.crawler4j v2.6.1
I see on maven central there is a v3.0/3.1 (no 2.6.1).
3.1 does not have a declared dsiutils dependency - maybe it does not need
it now.
I've patched any23 to nearly work with Crawler4j 3.3. I get one compile
error but I think submitting the patch now under ANY23-27 is the best thing
as it gives us an opportunity to strike whilst the iron is hot. So please
see the patch on the issue.
As I've not been able to get this to compile yet, I can't declare whether
we can completely drop the offending LGPL library.
One thing which I think it worth noting, is that when I contacted
Sebastiano yday he mentioned
{qt}
My guess is Bloom filters. You can reimplement them in a day, or substitute
then with a LongOpenHashSet of 64-bit signatures.
{qt}
Is this possibly why we rely upon DSIutils?
Hi Lewis,
DSIUtils is included in the POM because the entry in
any23-repository-external for crawler4j is very basic and does not
specific it's dependencies.
You've replaced it with v3.3 and that has no declared dependency on
DSIutils.
So dsiutils can be removed from any and all any23 poms.
And can we remove:
<repositories>
<!-- Specific repository for Any23 dependencies without a dedicated
repository. -->
<repository>
<id>any23-repository-external</id>
<url>http://any23.googlecode.com/svn/repo-ext</url>
</repository>
This is going to go (I presume - right? has to go?) - and at the moment
is confusing the issue a bit by putting stuff that's project-specific
into the POM path.
Its got in it:
edu.uci.ics.crawler4j v2.1
it.unimi.dsi.dsiutils v2.0.1
org.apache.commons.commons-csv v1.0-SNAPSHOT-rev1148315
net.xeoh.jspf v0.9.0
You're replacing crawler4j with v3.3 so (1) and (2) can go.
(3) is easy :-)
(4) may need a project-local copy and ship in source-release else
someone else can't build any23 from the release bits + maven central
repositories.
Anyway, hopefully the former comes to fruition and crawler4j doesn't need
to rely upon DSIutils anymore... (ahem).
Here's to hoping :0)
Yes!
Lewis
Andy
