Share
Embed
Kotlin: MAX PAYNE (BKUG, July 2017)
Kotlin gives us a lot of neat features and just general developer satisfaction, but little is known about its hidden costs and pain it causes. This talk is a almost a 2 year combination of small (and no so small) issues we encountered when developing a full-Kotlin application started from the scratch. Bugs, features, performance, testing — it’s all there with our answers, conclusions and advises.
Presented at Belarus Kotlin User Group meetup on July 19, 2017.
Artur Dryomov
July 19, 2017
Transcript
79 slides
-
K otlin: MA X PA Y NE A rtur D
ryomov, J uno L ab @arturdryomov -
J uno K otlin K otlin J uno W ell,
only sometimes, so no grudge here -
J uno R ider F u ll‑K ot lin +
R xJ av a fr om t he st art 2 y ears K ot lin 1 . 1 - E A P in pr odu ct ion for 6+ mont hs K ot lin 1 . 2 - E A P r ight now 99.9 % cr as h‑fr ee 2 w eek spr ints dev elop ers, QA -
J uno R ider c l o c m a
i n / k o t l i n — 40 K c l o c t e s t / k o t l i n — 38 K c l o c a n d r o i d T e s t / k o t l i n — 12 K -
C heckstyle and S tatic A nalysis S o yeah,
forget about it -
S pecification mumbo‑jumbo J av a AST — A bstr
act S y nt ax T r ee O nly J av a PSI — P r ogr am S tru ctur e I nt er face O nly IJ UAST — U niv ers al AST J av a and K ot lin A ndr oid L int is migr at ing t o it! -
C heckstyle H ow about not mapp ing it t
o AST and inst ead map it t o UAST (U niv ers al A bstr act S y nt ax T r ee)? T hat w ay als o adding supp ort for K ot lin, w hich is an official t hing for A ndr oid now, can be supp ort ed. C hecksty le w or ks only w it h AST as w e dep end on ANTLR and J av a gr ammar. -
C heckstyle and S tatic A nalysis P ut your
IJ config into repository You can somewhat run IJ headless D etekt works with PSI , so take it for a spin -
∞ paths L et’s bu ild s omet hing! v
a l g a m e B u i l d e r = S t r i n g B u i l d e r ( ) g a m e B u i l d e r . a p p e n d ( " t i c " ) g a m e B u i l d e r . a p p e n d ( " t a c " ) g a m e B u i l d e r . a p p e n d ( " t o e " ) v a l g a m e = g a m e B u i l d e r . t o S t r i n g ( ) -
∞ paths A ls o, a l s o .
v a l g a m e : S t r i n g = S t r i n g B u i l d e r ( ) . a l s o { i t . a p p e n d ( " t i c " ) i t . a p p e n d ( " t a c " ) i t . a p p e n d ( " t o e " ) } . t o S t r i n g ( ) -
∞ paths H mm, l e t it be? v
a l g a m e : S t r i n g = S t r i n g B u i l d e r ( ) . l e t { i t . a p p e n d ( " t i c " ) i t . a p p e n d ( " t a c " ) i t . a p p e n d ( " t o e " ) i t . t o S t r i n g ( ) } -
∞ paths a p p l y it! v a
l g a m e : S t r i n g = S t r i n g B u i l d e r ( ) . a p p l y { a p p e n d ( " t i c " ) a p p e n d ( " t a c " ) a p p e n d ( " t o e " ) } . t o S t r i n g ( ) -
∞ paths r u n s eems nice. v a
l g a m e : S t r i n g = S t r i n g B u i l d e r ( ) . r u n { a p p e n d ( " t i c " ) a p p e n d ( " t a c " ) a p p e n d ( " t o e " ) t o S t r i n g ( ) } -
∞ paths H ow about w i t h ?
v a l g a m e : S t r i n g = w i t h ( S t r i n g B u i l d e r ( ) ) { a p p e n d ( " t i c " ) a p p e n d ( " t a c " ) a p p e n d ( " t o e " ) t o S t r i n g ( ) } -
∞ paths B ett er br ew t hat coffee
for long PR s ? : vs. l e t vs. i f ( i t = = n u l l ) -
∞ paths C O D E _ S T Y
L E . m d I t does not r eally matt er, just go w it h it -
D estructuring N o joke, actu ally us efu l.
v a l ( f i r s t , s e c o n d ) = p a i r v a l ( f i r s t , s e c o n d , t h i r d ) = t r i p l e O b s e r v a b l e . f r o m C a l l a b l e { c r e a t e P a i r ( ) } . s u b s c r i b e { ( f i r s t , s e c o n d ) - > p r i n t l n ( " 1 s t i s $ f i r s t , 2 n d i s $ s e c o n d " ) } -
D estructuring S upp orts d a t a c
l a s s and su ggest ed by IJ . d a t a c l a s s P e r s o n ( f i r s t N a m e : S t r i n g , l a s t N a m e : S t r i n g ) v a l ( f i r s t N a m e , l a s t N a m e ) = p e r s o n -
D estructuring S omebody sw aps fields C omp iler
is happy! d a t a c l a s s P e r s o n ( l a s t N a m e : S t r i n g , f i r s t N a m e : S t r i n g ) v a l ( f i r s t N a m e , l a s t N a m e ) = p e r s o n -
D estructuring W e just ban it unless P a
i r or T r i p l e -
n u l l safety? T his w or ks
just fine! f u n c r e a t e O b s e r v a b l e ( ) : O b s e r v a b l e < I n t > { r e t u r n O b s e r v a b l e . j u s t ( n u l l ) } OH HAI R xJ av a 2 R epro du ced on RxJ av a 2.1.1. -
n u l l safety!! T he langu age r
elies on @ N o n N u l l annot at ions. Your code is fu ll of u nicor ns J av a code is fu ll of n u l l . -
n u l l safety!! R ead docs and s
our ces car efu lly. P r omot e @ N u l l a b i l i t y t o y our fav or it e libr ar ies. H op efu lly A ndr oid SDK w ill cat ch up aft er annou ncements. -
R xJ ava 2 S tr eams p anic on
n u l l v alu es. K ot lin has t o be fr iends w it h J av a. S ide r oad — S w ift vs. K ot lin S w ift: ? and n i l just su gar for e n u m O p t i o n a l . K ot lin: ? is a langu age help er, no gu ar ant ees t hou gh. -
K optional https://github.com/gojuno/koptional v a l s o m e
= S o m e ( v a l u e ) v a l n o n e = N o n e v a l o p t i o n a l = n u l l a b l e V a l u e . t o O p t i o n a l ( ) v a l n u l l a b l e = o . t o N u l l a b l e ( ) v a l v a l u e S t r e a m = o p t i o n a l S t r e a m . f i l t e r S o m e ( ) v a l s i g n a l S t r e a m = o p t i o n a l S t r e a m . f i l t e r N o n e ( ) U s e ? : and l e t — embr ace t he langu age. -
SAM C onversion Singl e A bstr act M etho
d / / R x J a v a 1 O b s e r v a b l e . z i p ( a , b ) { a , b - > a t o b } / / R x J a v a 2 O b s e r v a b l e . z i p ( a , b , B i F u n c t i o n < I n t , I n t > { a , b - > a t o b } ) -
SAM C onversion / / R x J a v
a 1 p u b l i c s t a t i c < T 1 , T 2 , R > O b s e r v a b l e < R > z i p ( O b s e r v a b l e < T 1 > o 1 , O b s e r v a b l e < T 2 > o 2 , F u n c 2 < T 1 , T 2 , R > z i p p e r ) / / R x J a v a 2 p u b l i c s t a t i c < T 1 , T 2 , R > O b s e r v a b l e < R > z i p ( O b s e r v a b l e S o u r c e < T 1 > o 1 , O b s e r v a b l e S o u r c e < T 2 > o 2 , B i F u n c t i o n < T 1 , T 2 , R > z i p p e r ) -
SAM C onversion / / R x J a v
a 1 p u b l i c i n t e r f a c e F u n c 2 < T 1 , T 2 , R > e x t e n d s F u n c t i o n { R c a l l ( T 1 t 1 , T 2 t 2 ) ; } / / R x J a v a 2 p u b l i c i n t e r f a c e B i F u n c t i o n < T 1 , T 2 , R > { R a p p l y ( T 1 t 1 , T 2 t 2 ) ; } -
SAM C onversion / / R x J a v
a 1 p u b l i c c l a s s O b s e r v a b l e < T > { / / R e d a c t e d } / / R x J a v a 2 p u b l i c i n t e r f a c e O b s e r v a b l e S o u r c e < T > { v o i d s u b s c r i b e ( O b s e r v e r < T > o b s e r v e r ) ; } -
SAM C onversion https://youtrack.jetbrains.com/issue/KT ‑14984 Impossible to pass not all
SAM argument as function -
SAM C onversion W e’ve made a bunch of extensions
for now R xK otlin has some -
G son and R eflection d a t a c
l a s s M o d e l { @ S e r i a l i z e d N a m e ( " m o d e l _ f i e l d " ) v a l f i e l d : S t r i n g } m o d e l . f i e l d . t o S t r i n g ( ) / / C r a s h ! -
G son and R eflection G son is happy to
write n u l l to properties R eflection is a thing in the J ava world... -
G son and R eflection F ail fast d a
t a c l a s s M o d e l { @ S e r i a l i z e d N a m e ( " m o d e l _ f i e l d " ) v a l f i e l d : S t r i n g } : V a l i d a t a b l e { o v e r r i d e f u n v a l i d a t e ( ) { i f ( f i e l d = = n u l l ) t h r o w P a r s e E x c e p t i o n ( ) } } -
G son and R eflection T ake a look at
M oshi or J ackson U ses K otlin native reflection -
i n l i n e or n o i
n l i n e N eat, r edu ces met hods cou nt. OH HAI A ndr oid F eels like ker nel dev elop ment. N ot emr aced at all. M ost ly for ces @ S u p p r e s s ( " n o t h i n g _ t o _ i n l i n e " ) . -
i n l i n e or n o i
n l i n e https://youtrack.jetbrains.com/issue/KT ‑16769 R evisit compiler and IDE warning about inline functions -
i n l i n e or n o i
n l i n e K now about it I ’m looking at you, p r i v a t e f u n -
c o m p a n i o n o
b j e c t and c o n s t / / K o t l i n c o m p a n i o n o b j e c t { p r i v a t e v a l N U M B E R = 4 2 } / / J a v a p u b l i c s t a t i c f i n a l c l a s s C o m p a n i o n { p r i v a t e f i n a l i n t N U M B E R = 4 2 ; p r i v a t e f i n a l i n t g e t N U M B E R ( ) { r e t u r n N U M B E R ; } } -
c o m p a n i o n o
b j e c t and c o n s t S o y eah, + met hod w it hout c o n s t . M et hods w ill not cou nt t hems elv es! I nv ocat ion costs ar e r eal -
c o m p a n i o n o
b j e c t and c o n s t / / K o t l i n c o m p a n i o n o b j e c t { p r i v a t e c o n s t v a l N U M B E R = 4 2 } / / J a v a p u b l i c s t a t i c f i n a l c l a s s C o m p a n i o n { p r i v a t e s t a t i c f i n a l i n t N U M B E R = 4 2 ; / / W i l l a c t u a l l y i n l i n e t h e v a l u e . } -
c o m p a n i o n o
b j e c t and c o n s t N ot adv ocat ed enou gh. O nly pr imit iv es and S t r i n g . S hou ldn’t comp iler do it its elf? -
c o m p a n i o n o
b j e c t and c o n s t K now about it -
H idden C osts IJ → T ools → S
how K otlin B ytecode L ots of fun, also — scary! -
B ugs f u n m u l t i
p l y ( d o u b l e : D o u b l e ) = d o u b l e * d o u b l e U p dat e K ot lin 1.1.2 1.1.3. j a v a . l a n g . I l l e g a l S t a t e E x c e p t i o n o r g . j e t b r a i n s . k o t l i n . k a p t 3 . d i a g n o s t i c . K a p t E r r o r E r r o r w h i l e a n n o t a t i o n p r o c e s s i n g J av a keyw or ds ar e su ddenly banned. Ch ang elog? Wh at ch ang elog? -
A ssemble C oming fr om J av a? +
bu ild st ep. U s e D agger? + bu ild st ep. k a p t incr ement al comp ilat ion didn’t w or k for a long t ime... H av e a change? R E B U I L D . J ust r emember y ou ar e doing it t imes p er day. -
A ssemble C omp ilat ion is CPU int ens
iv e. L apt ops hav e mobile CPU s. PC s ar e cheap, easy t o up gr ade... -
M ainframer https://github.com/gojuno/mainframer https://github.com/elpassion/mainframer‑intellij‑plugin 1. L apt op → Sour
ce Co de → B u ild M achine. 2. B u ild M achine → Buil d R esults → L apt op. -
M ainframer $ . / g r a d l
e w c l e a n a s s e m b l e D e b u g - - n o - d a e m o n W it hout MF : 4 minut es 10 s econds. L apt op ov er heat B as ically cannot us e y our comput er W it h MF : 1 minut e 10 s econds. -
A ssemble D o all optimizations you can M inimu
m SDK t o L ollip op. G r adle bu ild cache. G r adle A ndr oid P lu gin DE X bu ild cache. M ainfr amer -
S pek @ R u n W i t h
( J U n i t P l a t f o r m : : c l a s s ) c l a s s S p e c : S p e k ( { v a l s y s t e m b y m e m o i z e d { S y s t e m ( ) } c o n t e x t ( " b o o t " ) { b e f o r e E a c h T e s t { s y s t e m . b o o t ( ) } i t ( " r e p o r t s b o o t s t a t u s " ) { a s s e r t T h a t ( s y s t e m . s t a t u s ( ) ) . i s E q u a l T o ( B O O T E D ) } } } ) -
S pek: versions 1 . 0 . 8 9 D
e s c r i b e B o d y → D s l 1 . 1 . 0 - b e t a 3 D s l → S p e c B o d y D ocs ment ion 1 . 0 , 1 . 1 . 1 9 . D o not ex ist. M ov ed fr om JU nit 4 t o JU nit 5 at s ome p oint -
S pek: versions https://github.com/J etB rains/spek/issues/170 Please fix Spek versioning,
compatibility and publishing -
S pek: m e m o i z e d
caching @ R u n W i t h ( J U n i t P l a t f o r m : : c l a s s ) c l a s s O o p s S p e c : S p e k ( { v a l s y s t e m b y m e m o i z e d { S y s t e m ( ) } l i s t O f ( 1 , 2 , 3 ) . f o r E a c h { i n d e x - > c o n t e x t ( " b o o t " ) { i t ( " b o o t s " ) { } } } } S ame s y s t e m object for each ru n -
S pek: m e m o i z e d
caching @ R u n W i t h ( J U n i t P l a t f o r m : : c l a s s ) c l a s s O k S p e c : S p e k ( { v a l s y s t e m b y m e m o i z e d { S y s t e m ( ) } l i s t O f ( 1 , 2 , 3 ) . f o r E a c h { i n d e x - > c o n t e x t ( " b o o t $ i n d e x " ) { i t ( " b o o t s " ) { } } } } U niqu e s y s t e m object for each ru n -
S pek: m e m o i z e d
caching https://github.com/J etB rains/spek/issues/210 M emoized and same test names -
S pek: m e m o i z e d
before and after @ R u n W i t h ( J U n i t P l a t f o r m : : c l a s s ) c l a s s J u s t S p e c : S p e k ( { v a l s y s t e m b y m e m o i z e d { S y s t e m ( ) } b e f o r e E a c h T e s t { p r i n t l n ( " b e f o r e $ s y s t e m " ) } a f t e r E a c h T e s t { p r i n t l n ( " a f t e r $ s y s t e m " ) } } ) B efor e and aft er s y s t e m ar e differ ent. -
S pek: m e m o i z e d
before and after https://github.com/J etB rains/spek/issues/213 M emoized value is not the same before and after test -
S pek: JU nit A ctu ally JU nit 5
fr om now on. U s e @ R u n W i t h ( J U n i t P l a t f o r m : : c l a s s ) t o w or k w it h JU nit 4. JU nit 5 is not r eleas ed, s o IJ may not w or k. OH HAI An droi d Stu dio 3 B et a. f o n and f i t w er e r emov ed, us e S p ek IJ p lu gin. Do es not work with J Unit 4. T his is fu n -
S pek K ind of adv ert iz ed by
J etB r ains, but... C ommits ar e tw o t imes a mont h on av er age B r eaks comp at ibility eas ily D ocs feel out dat ed all t he t ime. T her e ar e bu gs. R emember v anilla JU nit st ability? A ls o, hav e 100 K t ests? H av e t o maint ain it -
B onus round You are on your own P repare
that YouT rack account... -
B onus round ... K ot lin is v ery
r efer ence s emant ics, it’s a t hin lay er on t op of J av a, and s o it p erp etu at es t hr ou gh a lot of t he J av ais ms in its model. ... I f w e had done an analog t o t hat for O bject iv e‑C it w ou ld be like, ev eryt hing is an NSO bject and it’s objc_ms g S end ev eryw her e, just w it h p ar ent hes es inst ead of squ ar e br ackets. ... Chris L attn er, Swift -
B onus round J av a: O r acle, R
ed H at, I nt el, IBM , E clips e, T w itt er, J etB r ains, et c. K ot lin: J etB r ains. T his is fine actu ally, but do not exp ect a mir acle. F or a s ingle comp any a damn gr eat langu age -
C onclusion T her e w ill be dr agons
K ot lin w ill not make y ou a bett er dev elop er. I t w ill giv e y ou t ools t hou gh R emember about J av a u nder neat h. S w ift can br eak t hings, K ot lin cannot T he langu age is t op‑not ch and w ell‑des igned. B e car efu l!