Hi Tamas!
great to see you here.
On 18.04.23 10:57, Tamás Cservenák wrote:
Howdy,
just to chime in as a developer of maven-indexer:
1, Regarding server side processing we are shot in our foot, as index is
provided (is produced and hosted) by Sonatype, and we have no much impact,
nor is there any server side available. They just produce the GZIP files we
can download and munge as much as we like on them.
2. Am unsure about exact indexer use cases within NB, but indexer offers a
"remote client" for SMO as well (via HTTP to search web service),
https://github.com/apache/maven-indexer/tree/master/search-backend-smo
still, is less powerful than full indexing context (that one that is GB in
size), see tests
I didn't know that this existed. Awesome!
3. maven-indexer (the classic) could be combined with smo search: use
classic to index local repository while for remote use SMO client, etc.or
some sort of "overlay" could be possible as well...
I have a draft PR here which uses the web api directly for selected
search queries:
https://github.com/apache/netbeans/pull/4971
will take a look at search-backend-smo - since this sounds like this
would replace all the json parsing of that PR.
Regarding using web services: I do think they have potential for
selective queries. But there will be features which likely won't work.
For example editor hints which scan the pom for version upgrades. I got
latencies between 200ms and 4s (!) per query (see PR for log). Right now
hints can run those checks instantly, with a web service this could take
minutes (assuming we don't get throttled).
NB does have a mechanism for returning partial results for queries
(instantly) and optionally blocking for full results. The web API would
never return anything instantly unless we would start to cache a search
engine locally which probably wouldn't work that well.
best regards,
michael
In the end, I want to thank Michael Bien of NB, he made a great deal of
contributions to indexers so far. Finally, there is one related Michael PR
still waiting for me...
https://github.com/apache/maven-indexer/pull/302
Thanks
T
On Tue, Apr 18, 2023 at 10:44 AM Neil C Smith <neilcsm...@apache.org> wrote:
On Tue, 18 Apr 2023 at 06:49, Jan Lahoda <lah...@gmail.com> wrote:
I apologize for being contrarian, but since the index download started
for
me (again) while on a bus with very poor internet connection, I guess I
should tell you my view.
...
To be honest, it never seemed very polite to me to download and process
so
much without asking.
+1 - I do think that more control and explicit information before
first run would be good. I do a lot of testing during releases with
userdirs in /tmp - I hate to think how many GBs I've downloaded there.
If I consider there would be (only) 10 000
installations of NetBeans,
Let's not be quite that pessimistic! ;-) There have been 1.5
million* downloads of various NetBeans 17 binaries so far from ASF
infrastructure, never mind other sources, community installers, etc.
Of course, that increases your other figures too!
* or 0.5 million unique IPs, so probably somewhere in between - access
to stats for people with Apache account at
https://logging1-he-de.apache.org/stats/
Best wishes,
Neil
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org
For additional commands, e-mail: dev-h...@netbeans.apache.org
For further information about the NetBeans mailing lists, visit:
https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org
For additional commands, e-mail: dev-h...@netbeans.apache.org
For further information about the NetBeans mailing lists, visit:
https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
