The state of 0.5 version? #75

qrilka · 2016-11-04T13:34:48Z

I think it (version 0.5) is in the master for some time already but I have seen no plans or roadmap - are there any?

The text was updated successfully, but these errors were encountered:

ryantrinkle · 2016-11-04T13:40:46Z

@qrilka The basic plan is to get some of the branches that have been kicking around merged in, then some testing and release. The branches have fairly big performance improvements, so I don't want to pass them by. Also, I need to do a bit more to ensure that 0.5 is as compatible with 0.4 as possible - I don't want people to have to change their code too much to upgrade.

qrilka · 2016-11-04T13:59:34Z

Thanks for your quick response @ryantrinkle
Are there any estimates maybe?
Is it planned as a Christmas present? :)

lingnand · 2016-11-04T14:56:33Z

Joining the queue for the Christmas present :)

On Fri, 4 Nov 2016, 13:59 Kirill Zaborsky, notifications@github.com wrote:

Thanks for your quick response @ryantrinkle
https://github.com/ryantrinkle
Are there any estimates maybe?
Is it planned as a Christmas present? :)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#75 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AB5kSoxTdCiXrNaNXO-uakutoClb41tqks5q6zpIgaJpZM4KpiqP
.

Lingnan Dai

qrilka · 2016-11-29T15:30:20Z

@ryantrinkle is it expected for the current master to consume more RAM than reflex-0.4.0? In one of "bad" pages in our app we load quite a huge table with XHR and Chrome shows something like 500M consumed by the previous version while for the version ported to 0.5 it shows1500M.
Could it be related to the switch from String to Text ()?

ryantrinkle · 2016-11-29T20:01:30Z

@qrilka Are you referring to develop? master should be basically the same as 0.4.0 right now.

No, I wouldn't expect increased memory usage like that - it would be great to have an issue with the info to reproduce that. Would be very interesting to track down!

qrilka · 2016-11-29T20:39:25Z

oh, sorry...
develop sure
I'll try to reproduce it in a small example tomorrow.
And those number are for peak consumption though average value also seems to be higher for develop.

qrilka · 2016-11-30T08:33:16Z

Filed #81 regarding memory issues

qrilka · 2016-11-30T08:34:43Z

Also I've noticed that porting to 0.5.0 increased the size of our minified app js by 200k - is it expected?

deepfire · 2017-03-16T15:43:52Z

@ryantrinkle, maybe making a GitHub milestone for 0.5.0 release could help us?

sboosali · 2018-01-03T02:55:24Z

Any update on this? It'd be nice to get an official release.

ryantrinkle · 2018-01-03T05:30:52Z

@sboosali Agreed. @deepfire A github milestone is a great idea; I will try to prioritize creating one with everything that's needed for a release ASAP. Hopefully it will be fairly small things that are required. Off the top of my head:

There are some missing typeclass instances
ModalsT is commented out; it should probably just be removed for now, and reimplemented in terms of the new RequesterT stuff at some point
There are probably some docs missing

sboosali · 2018-01-03T09:36:15Z

while pretty trivial, w.r.t. missing instances, FunctorMaybe is satsified by several containers, anything that can be empty. Like:

instance FunctorMaybe Maybe
fmapMaybe :: (a -> Maybe b) -> Maybe a -> Maybe b   
fmapMaybe k m = m >>= k

also, the current Witherable seems to have a Filterable class with only a Functor super class.

https://hackage.haskell.org/package/witherable-0.2/docs/Data-Witherable.html

If it doesn't add any dependencies not already present, maybe it's worth it to just pull it in?

ryantrinkle · 2018-01-03T13:21:17Z

@sboosali Using something outside reflex for FunctorMaybe would be great; however, there are a few options, and I'm not quite sure which is best. What about Compactable? Do you have any sense of how it compares with witherable?

ryantrinkle · 2018-01-03T13:23:55Z

@deepfire @sboosali Added the new milestone here: https://github.com/reflex-frp/reflex/milestone/1

sboosali · 2018-01-03T16:34:59Z

I like that Witherable subclasses Functor; Compactable doesn't. It's more lawful, and will keep the constraints shorter. (But both provide default method definitions on Functors.) Also, Witherable has more downloads, and fewer dependencies. But, I haven't used the `fmapMaybe` concept beyond mapMaybe, so I can't say, they both look fine.

…

On Jan 3, 2018 5:23 AM, "Ryan Trinkle" ***@***.***> wrote: @deepfire <https://github.com/deepfire> @sboosali <https://github.com/sboosali> Added the new milestone here: https://github.com/reflex-frp/reflex/milestone/1 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#75 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACNoMXC-xVerDi6Q6L61p0z-Wfza_LHqks5tG39ugaJpZM4KpiqP> .

Fresheyeball · 2018-01-03T18:30:38Z

@sboosali Compactable subsumes Witherable. So I recommend Compactable (not just because I wrote it), but because you can't use Witherable with Event t, since it has extraneous constraints. Compactable also has a better structure for providing performance specific function instances where appropriate, for example, fmapMaybe can have a higher performance instance for list than catMaybes . fmap.

sboosali · 2018-01-03T18:32:21Z

Event is not Witherable, just Filterable.

…

On Jan 3, 2018 10:30 AM, "Isaac Shapira" ***@***.***> wrote: @sboosali <https://github.com/sboosali> Compactable subsumes Witherable. So I recommend Compactable (not just because I wrote it), but because you can't use Witherable with Event t, since it has extraneous constraints. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#75 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACNoMRGKXr_jWOoZW7p_Ex0etWo5Mfg8ks5tG8dSgaJpZM4KpiqP> .

Fresheyeball · 2018-01-03T18:34:40Z

@sboosali right. But if you use Compactable, Event t is Compactable, as well as can have a Compact Functor instance, and filter comes for free. Imho Filterable is a broken abstraction, compact :: f (Maybe a) -> f a is the core abstraction. All the Witherable abstractions can be written in terms of compact, but not vice versa.

ElvishJerricco · 2018-01-03T18:40:26Z

I agree with @sboosali. The Functor superclass is important. I do not see anything broken about Filterable. The only advantage of Compactable is that it works for non-functor types, which isn't really the semantics anyone is looking for. Plus, Compactable's weird use of inner constraints pose a couple of problems.

Specializing such constraints is very hard, so it could be a major performance issue.
As long as we're focusing on simplifying Reflex (and I do think that's a major focus right now), this use of inner constraints on outer types is probably unnecessarily complicated for newcomers (who have enough trouble just figuring out basic use of inner constraints in traverse).

sboosali · 2018-01-03T18:44:03Z

yeah, I agree about ` f (Maybe a) -> f a ` :-) On Jan 3, 2018 10:34 AM, "Isaac Shapira" <notifications@github.com> wrote: @sboosali <https://github.com/sboosali> right. But if you use Compactable, Event t is Compactable, as well as can have a Compact Functor instance, and filter comes for free. Imho Filterable is a broken abstraction, compact :: f (Maybe a) -> f a is the core abstraction. All the Witherable abstractions can be written in terms of compact, but not vice versa. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#75 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACNoMaDXlBTuuhhi94ib5U-_Tp1Njkoiks5tG8hEgaJpZM4KpiqP> .

ryantrinkle · 2018-01-03T18:45:31Z

@ElvishJerricco Very good point about specialization. Since ghcjs is still on 8.0, and 8.0 doesn't seem to ever specialize code when a method has a constraint, I think we should defer this discussion until ghcjs moves to 8.2 and we re-test. Unspecialized code is insanely slow.

ryantrinkle · 2018-01-04T03:36:47Z

I got rid of one of the missing method implementations today, leaving just traverseIntMapWithKeyWithAdust for QueryT. I'm also most of the way through assigning issues to milestones. Progress!

ryantrinkle · 2018-01-04T04:58:55Z

Alright, every issue now has a milestone: either v0.5 or Later.

jappeace · 2018-10-24T18:51:05Z

You could just push out a release and set the milestones for v0.6. Looking at the haddock difference 0.4 is hopelessly out of date.

ali-abrar · 2019-01-31T13:27:48Z

Hi everyone - closing this as 0.5 has hit hackage. Sorry for the wait.

deepfire · 2019-01-31T13:45:14Z

Congratulations, gentlemen, and thank you for your great work! : -)

deepfire mentioned this issue Mar 14, 2017

reflex-0.5.0 is not on Hackage deepfire/reflex-glfw#5

Open

qrilka mentioned this issue Mar 20, 2017

HasPostGui Equivalent in Reflex-Dom 0.4? reflex-frp/reflex-dom#139

Open

ryantrinkle added this to the v0.5 milestone Jan 4, 2018

dfordivam mentioned this issue Mar 16, 2018

Haddock+Hoogle for Reflex reflex-frp/reflex-frp.org#9

Closed

dfordivam added the question label Jan 25, 2019

ali-abrar closed this as completed Jan 31, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

The state of 0.5 version? #75

The state of 0.5 version? #75

qrilka commented Nov 4, 2016

ryantrinkle commented Nov 4, 2016

qrilka commented Nov 4, 2016

lingnand commented Nov 4, 2016

qrilka commented Nov 29, 2016

ryantrinkle commented Nov 29, 2016

qrilka commented Nov 29, 2016

qrilka commented Nov 30, 2016

qrilka commented Nov 30, 2016

deepfire commented Mar 16, 2017

sboosali commented Jan 3, 2018

ryantrinkle commented Jan 3, 2018 •

edited

Loading

sboosali commented Jan 3, 2018 •

edited

Loading

ryantrinkle commented Jan 3, 2018

ryantrinkle commented Jan 3, 2018

sboosali commented Jan 3, 2018 via email •

edited

Loading

Fresheyeball commented Jan 3, 2018 •

edited

Loading

sboosali commented Jan 3, 2018 via email

Fresheyeball commented Jan 3, 2018

ElvishJerricco commented Jan 3, 2018

sboosali commented Jan 3, 2018 via email

ryantrinkle commented Jan 3, 2018

ryantrinkle commented Jan 4, 2018

ryantrinkle commented Jan 4, 2018

jappeace commented Oct 24, 2018

ali-abrar commented Jan 31, 2019

deepfire commented Jan 31, 2019

The state of 0.5 version? #75

The state of 0.5 version? #75

Comments

qrilka commented Nov 4, 2016

ryantrinkle commented Nov 4, 2016

qrilka commented Nov 4, 2016

lingnand commented Nov 4, 2016

qrilka commented Nov 29, 2016

ryantrinkle commented Nov 29, 2016

qrilka commented Nov 29, 2016

qrilka commented Nov 30, 2016

qrilka commented Nov 30, 2016

deepfire commented Mar 16, 2017

sboosali commented Jan 3, 2018

ryantrinkle commented Jan 3, 2018 • edited Loading

sboosali commented Jan 3, 2018 • edited Loading

ryantrinkle commented Jan 3, 2018

ryantrinkle commented Jan 3, 2018

sboosali commented Jan 3, 2018 via email • edited Loading

Fresheyeball commented Jan 3, 2018 • edited Loading

sboosali commented Jan 3, 2018 via email

Fresheyeball commented Jan 3, 2018

ElvishJerricco commented Jan 3, 2018

sboosali commented Jan 3, 2018 via email

ryantrinkle commented Jan 3, 2018

ryantrinkle commented Jan 4, 2018

ryantrinkle commented Jan 4, 2018

jappeace commented Oct 24, 2018

ali-abrar commented Jan 31, 2019

deepfire commented Jan 31, 2019

ryantrinkle commented Jan 3, 2018 •

edited

Loading

sboosali commented Jan 3, 2018 •

edited

Loading

sboosali commented Jan 3, 2018 via email •

edited

Loading

Fresheyeball commented Jan 3, 2018 •

edited

Loading