These are chat archives for exceptionless/Discuss

19th
Mar 2016
Sharath Satish
@sharathsatish
Mar 19 2016 01:43
@niemyjski that's fine . I have that portion scripted. This secind atep is going to take longer. Perhaps we can merge just that first step so we don't have such a dependency.
Second step
Is that possible?
Blake Niemyjski
@niemyjski
Mar 19 2016 01:49
Be nice to have it as one so we can let it bake and test anything I can help with?
I don't see how deduping would be that much faster I think it's a bug or we should dig deeper with a profile and take a look
Sharath Satish
@sharathsatish
Mar 19 2016 03:10
It will probably mean a bit of rework
Since that first one is really tricky to maintain with other changes.
Since it changes the location of external dependencies
The next stage of moving to project.json / xproj instead of csproj is easier to do side by side like EF has done
The next set we can do all at once for sure
Sander Rijken
@srijken
Mar 19 2016 15:52
@niemyjski I’m running the profiler against both branches
it’s the JSON deserialize that’s causing it to be more expensive
we don’t run that in the deduplication branch
Sander Rijken
@srijken
Mar 19 2016 15:59
what’s strange is that GetHashCode() takes 2.7ms and GetError takes 2.3ms, so that doesn’t explain the perf difference
Sander Rijken
@srijken
Mar 19 2016 16:04
Screen Shot 2016-03-19 at 17.03.59.png
Screen Shot 2016-03-19 at 17.06.18.png
Sander Rijken
@srijken
Mar 19 2016 16:11
I’d say we’re good :)
Blake Niemyjski
@niemyjski
Mar 19 2016 18:34
nice find
that’s good to know
I’m starting to think that maybe we late bound these serializers from the client
but they should be really fast to get from the di system
so if that’s not the case
I guess we should look into why it’s slow
nice job!
Sander Rijken
@srijken
Mar 19 2016 18:41
it’s not getting the serializers what’s slow
it’s deserializing itself
Sander Rijken
@srijken
Mar 19 2016 19:05
running around in circles
Blake Niemyjski
@niemyjski
Mar 19 2016 19:50
Ah got ha
Sander Rijken
@srijken
Mar 19 2016 20:48
I’m in your branches breaking your builds :P
Eric J. Smith
@ejsmith
Mar 19 2016 20:49
Haha!
Blake Niemyjski
@niemyjski
Mar 19 2016 20:50
:)
Sander Rijken
@srijken
Mar 19 2016 20:50
not joking :)
build script ignored failed builds, so when you introduce a build error in the tests, the tests don’t run, and everything is well
Eric J. Smith
@ejsmith
Mar 19 2016 20:52
Doh!
Guess we need to get that fixed
Sander Rijken
@srijken
Mar 19 2016 20:53
it is
Eric J. Smith
@ejsmith
Mar 19 2016 20:53
Awesome. Thanks!
Sander Rijken
@srijken
Mar 19 2016 20:54
it isn’t fixed in the way I want yet though, it doesn’t make the lines with the errors red
most important thing is that it fails when it should fail
and this one is on the other branch (failed testcase that wasn’t noticed because tests didn’t run): https://ci.appveyor.com/project/Exceptionless/exceptionless-net/build/3.3.1618
Eric J. Smith
@ejsmith
Mar 19 2016 20:56
Dang out build output got ugly somewhere along the way. Gotta get those warnings cleaned up too.
@niemyjski we need to turn warnings as errors on.
Sander Rijken
@srijken
Mar 19 2016 21:01
@niemyjski I’m done now with benchmark branch/PR
Sander Rijken
@srijken
Mar 19 2016 21:07
Is the nuget badge broken for you guys too? for Exceptionless.Net
Eric J. Smith
@ejsmith
Mar 19 2016 21:08
Yeah I don't think the download count works any more
I removed it from Foundatio
Blake Niemyjski
@niemyjski
Mar 19 2016 21:08
Let's merge it and remove the badge
Haha c# 6 bites us again
Sander Rijken
@srijken
Mar 19 2016 21:11
yes