Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jan 31 21:39

    gafter on master

    Propose Feb 11 agenda (#2187) (compare)

  • Jan 31 21:39
    gafter closed #2187
  • Jan 31 19:42
    gafter edited #2115
  • Jan 31 15:46
    jcouv edited #1565
  • Jan 31 14:51
    yigitgursoy closed #1958
  • Jan 31 13:43
    ichensky closed #2190
  • Jan 31 12:17
    ichensky opened #2190
  • Jan 31 04:38
    MarkPflug edited #2189
  • Jan 31 04:37
    MarkPflug edited #2189
  • Jan 31 04:37
    MarkPflug edited #2189
  • Jan 31 04:36
    MarkPflug opened #2189
  • Jan 30 23:15
    MohammadHamdyGhanem opened #2188
  • Jan 30 20:01
    gafter review_requested #2187
  • Jan 30 20:01
    gafter review_requested #2187
  • Jan 30 20:01
    gafter review_requested #2187
  • Jan 30 20:01
    gafter assigned #2187
  • Jan 30 20:01
    gafter opened #2187
  • Jan 30 20:00

    gafter on gafter-2019-02-11-agenda

    Propose Feb 11 agenda (compare)

  • Jan 30 17:11
    tannergooding opened #2186
  • Jan 30 15:52
    lust666 closed #2185
Yair Halberstadt
@YairHalberstadt
Well I'm trying to move our codebase to that sort of model, but it's an old codebase, with millions of lines of code, and a slow process. Currently we tend to have one process for data access, one process for logic, and one process for the GUI. That's it. And each of those processes are huge. We definitely can't afford to kill one every time an error occurs. On the other hand, I can't help thinking that if we had killed the process every time an exception occurred to start off with, we wouldn't be having as many exceptions thrown which we can't fix now because the person who wrote that bit left, and we can't risk changing behaviour.
FailFasts also don't run finally blocks. Do you think in general it's more useful to run finally blocks when failfasting, or not to?
CyrusNajmabadi
@CyrusNajmabadi
depends on yoru architecture
basically, "it depends"
but you should have a design, document it, then follow it
define your isolation boundaries
define your critical regions
i'm a huge fan of process isolation
but i come from the unix world, where this is the norm
i also prefer it for things like security
because you can very clearly and easily define security boundaries at the process level
and because it's process level, no one is "accidentally" sharing state unintionally
it's very very easy to make it clear what the responsibilities are
qmail is a good example of this sort of design philosophy.
Shimmy
@weitzhandler
Is C# 8 already RTM?
Robin Sue
@Suchiman
no
VS2019 will have C# 8 (most likely)
Yaakov
@yaakov-h
gafter said not to infer a release date, but a few features have been merged into 2019 preview 1 branch. Not sure if the other features will even make the cut for 2019 rtm.
Shimmy
@weitzhandler
Thanks!
Joseph Musser
@jnm2
Accidentally did char + int + string.
Does not go well.
Yair Halberstadt
@YairHalberstadt
😆 That's almost JavaScript worthy
David Arno
@DavidArno
Of course, if it had been JS, you'd have likely ended up with an array at the end of it or some such. :innocent:
HaloFour
@HaloFour
a "falsey" array, somehow
Julien Couvreur
@jcouv
@Suchiman @weitzhandler @yaakov-h The C# 8 language support will not RTM with initial VS2019 release. The features and the language version will be there but as "beta", meaning some breaking language changes may still occur.
C# 8 will RTM in an update of VS2019, aligned with .NET Core 3.
HaloFour
@HaloFour
That is interesting.
Robin Sue
@Suchiman
hm so my tummy was right that C# 8 seems to late to finish up for VS2019
guess that means VS2019 will RTM quite soon (like sooner than my estimate of mid 2019)
HaloFour
@HaloFour
Have VS and C# ever been out of sync like that?
Robin Sue
@Suchiman
i don't think so, VS2017 is the first VS that got new language versions in minor updates and the first release with C# minor versions
CyrusNajmabadi
@CyrusNajmabadi
nope. but it def makes me happy to see
Robin Sue
@Suchiman
indeed, that flexibility is nice
although almost happened for... 2013? where roslyn almost shipped as an extension but eventually didn't
CyrusNajmabadi
@CyrusNajmabadi
this has been the case for thigns like TS.
have the release of htat decouples from the release of surrounding IDEs
Robin Sue
@Suchiman
more presents comming tomorrow :smile:
HaloFour
@HaloFour
Makes me wonder if VS will move to a subscription model like IntelliJ. Although, with MSDN, you could argue that's already the case.
Robin Sue
@Suchiman
@HaloFour it absolutely is, bothers you to login or expires
HaloFour
@HaloFour
eh, kinda, not quite the same. IntelliJ continually rolls out new releases, VS is still the big bang product with a long release cycle
Yaakov
@yaakov-h
VS 365
Joseph Musser
@jnm2
4ish hours until I should start looking for 2019-p1?
David Arno
@DavidArno
Guess it depends whether that "2019-p1" (not sure what the "p" means there) contains a beta of C# 8.
Sören Nils Kuklau
@chucker
(preview?)
David Arno
@DavidArno
Ah, of course. Thanks. :smile:
Joseph Musser
@jnm2
Sorry, silly shorthand.
Keynote starts at 11:30 am EST.
It ends at 1 pm, and it's usually sometime during the keynote IIRC?
(If it's happening)
David Arno
@DavidArno
Which time zone is that 1pm?