dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688666744
> I don't think that's the case. Running the whole thing a second time is
quite fast.
Because it's cached in the daemon? No way of telling without actually
digging at
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688661171
I'm committing it then, ok Uwe?
This is an automated message from the Apache Git Service.
To respond to the m
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688650077
I have a suspicion those filetree filters (exclusion filters) are fairly
dumb and scan paths that can't be traversed anyway. I'd have to confirm to be
sure (which I can't do
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688642607
bq. 32s before, 24s with this PR
Well, that's 25% faster. The biggest gain should come from running
incrementally? If you modify a file in just one of the projects, o
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688492819
You're right, Uwe - the subproject scan inclusion sucks. I don't know how to
avoid it in an elegant way. This could be done with worker API as well but it's
not the same - pe
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688210409
I'll take a look, but not now - working.
On Mon, Sep 7, 2020 at 11:38 AM Uwe Schindler
wrote:
> I committed my changes. You see the duplicate scanning now, be
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-688201937
bq. Actually the whole task does not scan any files
Ah, so this would explain the speedup. :)
Let me reiterate, sorry about the noise.
-
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-687320373
No worries. Not very urgent.
This is an automated message from the Apache Git Service.
To respond to the mess
dweiss commented on pull request #1830:
URL: https://github.com/apache/lucene-solr/pull/1830#issuecomment-687038375
@uschindler you know these checks better - to me I refactored this stuff in
the same way as before (and they're running much faster now since they can run
in parallel). If yo