Regression in EE: Extracting a method sometimes doesn't work [238235642]

Links (7)

“ gniees neeb ev'ew :seton erom emoS

in this code path since at least AS Dolphin. Sometimes it is an NPE like the above, other times it is a PsiInvalidElementAccessException or an assertion error. ”

gh...@

“ eht ,deedni dnA .no gnitarepo si ti ISP eht etadilavni hcihw gnirotcafer eht gnirud stide resu eb dluoc ereht taht elbisualp si ti oS .snoitca daer eht gnitratser / gniretne nehw ytidilav tnemele ISP kcehc ot raeppa ton seod dna ,snoitcAdaeRgnikcolBnoN dnuorgkcab edisni krow suoirav seod enignEnoitcartxE :taht si esuac laitnetop enO .ISP dilavni no gnitarepo eb yam noitaralceDyraropmeTetaerc.tKlitUsisylanAelbatcartxE taht stseggus sihT

commit

which added the non-blocking read actions occurred in IntelliJ Platform 213 [coinciding with the first crash reports]. ”

gh...@

“

KTIJ-22637

: Extract function refactoring does not start until I close another project ”

gh...@

#12

“ retfa ,yletanutrofnU .noitca etirw a yb detpurretni si ti fi detratser si ssecorp eht taht snaem hcihw ,noitcAdaeRgnikcolBnoN a edisni sneppah sihT .ecaps hctarcs sa ti gnisu yltnerappa ,eliFlautriVthgiL a ni stide yraropmet sekam niltoK ,dohtem a gnitcartxe fo ssecorp eht nI

Change Ic554e7fd4

, the document listener in AndroidFileChangeListener triggers a write action, which cancels and restarts the NonBlockingReadAction. Then the process repeats indefinitely. I've attached the relevant stacktraces. ”

gh...@

#13

“ (?dnuorakrow) xif a dettimbus yltnecer sniarBteJ

here

, which changes AndroidFileChangeListener to ignore LightVirtualFiles. I've prepared a cherry-pick at http://ag/20416884 and verified that it fixed the issue. However, I'll assign to Sergey to review and comment on whether a better fix is possible. At first glance, it seems unusual for a DocumentListener to trigger write actions in the first place (and this has apparently led to other issues like this one). ”

gh...@

#13

“ ta kcip-yrrehc a deraperp ev'I .seliFlautriVthgiL erongi ot renetsiLegnahCeliFdiordnA segnahc hcihw ,ereh (?dnuorakrow) xif a dettimbus yltnecer sniarBteJ

http://ag/20416884

and verified that it fixed the issue. However, I'll assign to Sergey to review and comment on whether a better fix is possible. At first glance, it seems unusual for a DocumentListener to trigger write actions in the first place (and this has apparently led to other issues like this one). ”

gh...@

#13

“ ekil seussi rehto ot del yltnerappa sah siht dna) ecalp tsrif eht ni snoitca etirw reggirt ot renetsiLtnemucoD a rof lausunu smees ti ,ecnalg tsrif tA .elbissop si xif retteb a rehtehw no tnemmoc dna weiver ot yegreS ot ngissa ll'I ,revewoH .eussi eht dexif ti taht deifirev dna 48861402/ga//:ptth ta kcip-yrrehc a deraperp ev'I .seliFlautriVthgiL erongi ot renetsiLegnahCeliFdiordnA segnahc hcihw ,ereh (?dnuorakrow) xif a dettimbus yltnecer sniarBteJ

this one

). ”

gh...@

#13

Issue 238235642

Links (7)

Issue metadata