Raka'a
1. Gabatarwa ga AI a Ci gaban Wayar hannu: Matsayi, Iyakoki, Tabbatarwa da Tsaro 2. Ƙirƙirar Lambar Wayar hannu tare da Hankali na Artificial: Kotlin, Swift da Ci gaban Platform 3. Zane-zanen Interface da Ƙwararrun Lambobin UI tare da Hankali na Artificial 4. Kan-Na'urar AI: Core ML, TensorFlow Lite da ML Kit 5. Haɗin Cloud AI da LLM API: Taɗi, Yawo da Tsaro 6. Ƙirƙirar Gwaji tare da Hankali na Artificial: Unit, Interface and Automation Tests 7. Ƙwarewa da Binciken Crash tare da Hankali na Artificial 8. Aiki da Inganta Baturi: Aikace-aikace masu sauri da inganci tare da Hankali na Artificial 9. Keɓantawa, Izini da Amintaccen Amfani 10. Sakin Store: App Store, Google Play da AI Compatibility 11. Ƙarshe-zuwa-Ƙarshe Project, Amfani da Haƙƙin Amfani da Hankali na Artificial da Taswirar Hanya a cikin Sana'a
Naúrar 7 / 11

Ƙwarewa da Binciken Crash tare da Hankali na Artificial

Riba:

  • Ability don ƙunsar abubuwan da za a iya haifar da sauri ta hanyar ba da bayanan faɗuwa (hanyoyi tara) zuwa hankali na wucin gadi tare da lambar da ta dace da mahallin yanayi.
  • Ikon warware tushen tushen har abada maimakon tabbatar da cutar AI a matsayin hasashe a cikin lamba da gwaji da rufe alamar.
  • Kare keɓantawa yayin da ake yin kuskure ta hanyar rufe bayanan sirri a cikin bayanan faɗuwa da rajistan ayyukan

Kowane aikace-aikacen yana ba da kurakurai; Abin da ke bambanta mai haɓaka mai kyau shine yadda sauri suke samun da gyara kwari. Gyara wayar hannu - ganowa da gyara tushen matsala - yana da wahala musamman saboda kuskuren yana faruwa akan na'urar mai amfani, a cikin yanayin da ba za ku iya gani ba. Yawancin lokaci, duk abin da kuke da shi shine rikodin faɗuwar faɗuwa (faɗuwar log / stack trace — fasarar fasaha ta inda aikace-aikacen ya tafi lokacin da ya fado). AI yana da ƙarfi sosai a karanta waɗannan bayanan sirri, jera dalilai masu yiwuwa, da ba da shawarar mafita. A cikin wannan rukunin za mu koyi yadda ake amfani da AI a matsayin "mai binciken kwaro" amma bar ku da alhakin tabbatar da ganewar asali da gyara.

Karatun log log: Inda AI ke haskakawa

Rubutun hatsari dogon rubutu ne mai ban tsoro; ƙwararren mai haɓakawa ba zai san inda zai duba ba. AI parses this text in seconds: at which line it crashed, which exception was thrown, what is the possible reason. Common mobile errors are obvious and the AI ​​recognizes them quickly: NullPointerException (trying to access a null value), IndexOutOfBoundsException (accessing a non-existent list element) on Android, EXC_BAD_ACCESS (accessing freed memory) on iOS, unexpectedly found nil (forcing a nil optional).

Mafi yawan ire-iren hadarurruka na wayar hannu da abubuwan da suka saba faruwa sune kamar haka:

Kuskure (banda)

Dandalin

na al'ada dalilin

NullPointerException

Android

Samun damar ƙima mara kyau

IndexOutOfBoundsException

Android

Samun dama ga ɓangaren lissafin da babu shi

ba zato ba tsammani

iOS

Tilasta buɗe wani zaɓi (!)

EXC_BAD_ACCESS

iOS

Samun dama ga ƙwaƙwalwar ajiya

ANR/daskare

Android

Dogon aiki/nauyi mai nauyi akan babban zaren

Gudun gyara kuskure mataki-mataki:

  1. Tattara rikodin. Haɗa log ɗin faɗuwar rana, saƙon kuskure, da matakai don sake buga shi idan zai yiwu.
  2. Ba da yanayin AI. Faɗa mani ba kawai kuskuren ba, amma yanki mai dacewa da abin da ya fado yana yi.
  3. Tambayi dalilai masu yiwuwa. "Ku gaya mani dalilai guda 3 da suka fi dacewa da kuma yadda zan tabbatar da kowanne."
  4. Tabbatar. Tabbatar da dalilin da aka tsara a cikin lamba da gwaji; Kar a gyara ta da zato.
  5. Gyara shi kuma gwada sake. Bincika cewa a zahiri kuskuren ya tafi kuma ba a sami sabbin kurakurai ba.
Tukwici: Lokacin ba da rajistar hadarin ga AI, haɗa da snippet ɗin da ya dace kuma. Sai kawai tare da alamar tari AI yana yin hasashen gabaɗaya; Lokacin da kuka ga lambar, yuwuwar gano ainihin layin da ainihin dalilin yana ƙaruwa sosai. Yanayi yana ƙayyade ingancin ganewar asali.

Tarkon bayanan sirri

Rikodin hadarurruka da rajistan ayyukan galibi suna ƙunshe da bayanan mai amfani: imel, ID na mai amfani, wuri, har ma da samar da abun ciki. Manna wannan rikodin a cikin AI kamar yadda ake watsa bayanan sirri ga ɓangare na uku kuma cin zarafi ne na KVKK / GDPR. Share wuraren sirri (mask) kafin ƙaddamar da rikodin. Har ila yau, a kula kada ku rubuta bayanan sirri zuwa majigin aikace-aikacenku tun daga farko; Kyakkyawan log yana bayyana matsalar amma baya bayyana ainihin.

Tsanaki: Gyaran da AI ya ba da shawara na iya "shiru da kwaro" amma maiyuwa bazai warware tushen dalilin ba. Misali, rufe NullPointerException tare da duba maras kyau zai dakatar da hadarin, amma idan ba ku gano dalilin da yasa darajar ta lalace ba ainihin kuskuren dabaru zai ci gaba. Yi maganin cutar, ba alamar ba.

Tushen bincike

Manufar ƙwararrun gyara kuskuren ba shine don rufe kuskuren ba amma don nemo tushen tushen. Na tambayi AI "me yasa wannan zai zama banza, a ina zai iya ɓacewa a cikin kwararar bayanai?" tambayar, "Yaya zan yi shiru?" Yana da daraja fiye da tambaya. Da zarar an sami tushen tushen, yawancin bambance-bambancen kuskure iri ɗaya ana warware su lokaci ɗaya. AI yana da kyau a wannan tunanin sarkar: bi bayanan daga shigarwa zuwa fitarwa kuma tambaye shi don tunani game da inda ya rushe.

uku mini lokuta

Case 1 - 2 hours na aiki a cikin minti 10. Wani mai haɓakawa ya kwashe awanni 2 yana neman kwaro wanda kawai ya fado akan takamaiman samfurin Samsung. Ya ba da log log (share wuraren sirri) ga AI; YZ ya ce kuskuren yana nuni zuwa ambaton ƙwaƙwalwar ajiya da ke faruwa tare da ƙudurin kyamara daban na waccan na'urar. Tare da alamar, an gano dalilin a cikin minti 10. AI ya hanzarta binciken, ɗan adam ya tabbatar da mafita.

Hali na 2 - Kwaron da aka yi shiru ya dawo. Ƙungiya ɗaya ta dakatar da wani karo mai maimaitawa ta hanyar amfani da shawarar AI don gwada kama shi. Hadarin ya tsaya, amma masu amfani sun fara korafin cewa "ba a adana bayanai"; saboda ainihin matsalar (Database connection) tana nan har yanzu, ta zama marar gani. Da zarar an gano tushen tushen, an warware matsalar da asarar bayanai. Darasi: shiru ba warwarewa ba ne.

Case 3 - Bayanai sun leka cikin log. Wani bincike ya gano cewa an rubuta cikakkun sunayen masu amfani da lambobin waya a cikin rajistan ayyukan hadarin. Masu haɓakawa akai-akai liƙa waɗannan rajistan ayyukan cikin AI da ƙayyadaddun kwari; Don haka bayanan sirri ke fita tsawon watanni. An rufe gundumomi kuma an gyara tsarin. Darasi: Sirri yana aiki ko da lokacin yin kuskure.

Rauni mai ƙarfi / Ƙarfi mai ƙarfi

Ba daidai ba: "Me yasa wannan kuskuren ke faruwa?

Ƙarfi mai ƙarfi: "Wannan haɗarin yana faruwa a cikin app ɗina na Android. Yanayi: - Yayin da ake yin: mai amfani yana ƙarawa zuwa cart daga cikakkun bayanai - A kan wasu na'urori kawai, ƙananan ƙirar RAM- Lambobi masu dangantaka: [ViewModel da Repository part] - Crash log (sharar da bayanan sirri): Lissafin 3 mafi kusantar tushen tushen. Ga kowane: 1) Ta yaya zan tabbatar, 2 ba tare da tabbatarwa ba.

Samfura masu kwafi

Samfurin bincike na Crash:"Bincika haɗarin da ke biyo baya. Ma'anar: [abin da kuke yi, wace na'ura / nau'in]. Lambar da ta dace: [lambar]. Rarraba log (sharar da bayanan sirri): [trace]. Ba 3 mafi kusantar tushen tushen da tabbatarwa + gyarawa na dindindin ga kowane. Har ila yau, yi alama wuraren aiki da ke rufe alamar."

Tushen samfurin samfurin: "Wannan ƙimar ta zo [rasa/ƙarya] ba zato ba tsammani. Bi bayanan da ke gudana daga shigarwar zuwa wannan batu: ina zai iya ɓacewa ko lalata? Faɗa mini inda zan duba a kowane mataki. [code]"

Samfurin karatun shiga: "Fassara wannan fitowar log: menene abubuwan da suka faru a cikin tsari, ina rashin daidaituwa, menene matakin lafiya na ƙarshe kafin kuskuren? [log - share bayanan sirri]"

Samfuran Sakewa: "Waɗanne matakai, jihohin na'ura, da bayanai zan yi ƙoƙarin sake haifar da wannan kuskure cikin dogaro?

Kuskuren gama gari

  • Bayar da alamar tari mara mahallin. Ba tare da lambar da ta dace da yanayin ba, AI yana yin hasashen gabaɗaya.
  • Manna bayanan sirri a cikin AI tare da rajistan ayyukan. Watsewar sirri; abin rufe fuska da farko.
  • Shiru alamar. Boye hadarin tare da gwada kamawa yana barin tushen matsalar kuma yana haifar da sabbin matsaloli.
  • Aiwatar da shawarar farko ba tare da tabbatar da ita ba. Sakamakon ganewar AI shine hasashe; Tabbatar da lambar.
  • Ƙoƙarin sake haifar da shi a cikin abin koyi. Wasu kurakurai suna bayyana akan ainihin na'urar/yanayin kawai.
  • Ba sake gwadawa ba bayan gyara. Gyaran zai iya karya wani abu dabam; Duba koma baya.

a takaice

One of the areas where AI excels is reading crash logs and sorting out possible causes; Ana inganta ingancin ganewar asali sosai lokacin da aka ba da mahallin. Amma ganewar asali da gyara na ƙarshe na ɗan adam ne: shawarar AI shine hasashe, an tabbatar da shi a cikin lamba da gwaji. Manufar ba shine a rufe alamar ba amma don magance tushen dalilin; Kuskuren shiru yakan dawo ta wata siga. Rukunin hadarurruka na iya ƙunsar bayanan sirri; Matsar da shi kafin ba da shi ga AI kuma kada ku rubuta bayanan sirri a cikin rajistan ayyukanku daga farkon.

Aikin aikace-aikace

Ɗauki log log ɗin da kuke da shi (ko samfurin da kuke samarwa daga AI), rufe duk wani bayanan sirri / keɓantacce a ciki, kuma ku ba AI tare da "samfurin bincike na Crash". Distinguish which of the root causes AI lists are actual fixes and which are just silencing. Aiwatar da gyara na dindindin da kuka zaɓa kuma tabbatar da cewa kuskuren ya ɓace kuma babu wasu sabbin matsaloli da suka taso.

jerin abubuwan dubawa

  • [ ] Na ba da log log tare da dacewa lamba da mahallin yanayi
  • [ ] Na rufe bayanan sirri / keɓaɓɓen bayanai a cikin rajistan ayyukan
  • [ ] Na tambayi AI don tushen tushen da gyarawa na dindindin, ba shiru ba
  • [ ] Na tabbatar da ganewar asali a cikin lamba da gwaji, ban yi amfani da shi a makance ba
  • [ ] Bayan gyara, na gwada cewa kuskuren ya tafi kuma babu koma baya
  • [ ] Na duba cewa aikace-aikacena baya rubuta bayanan sirri a cikin rajistan ayyukansa
A baya Na gaba