{"id":11382,"date":"2026-10-04T00:34:55","date_gmt":"2026-10-04T00:34:55","guid":{"rendered":"https:\/\/bitep.net\/blog\/2026\/10\/04\/githubcom-3\/"},"modified":"2026-10-04T00:34:55","modified_gmt":"2026-10-04T00:34:55","slug":"githubcom-3","status":"publish","type":"post","link":"https:\/\/bitep.net\/blog\/2026\/10\/04\/githubcom-3\/","title":{"rendered":"github.com"},"content":{"rendered":"

S\u00fcni \u0130ntellekt D\u0259st\u0259kli Proqramla\u015fd\u0131rmada Kontekst \u0130nqilab\u0131: Niy\u0259 H\u0259r \u015eey D\u00fczg\u00fcn Giri\u015fd\u0259n As\u0131l\u0131d\u0131r?<\/h2>\n

M\u00fcasir proqram t\u0259minat\u0131 m\u00fch\u0259ndisliyind\u0259 s\u00fcni intellekt (AI) al\u0259tl\u0259ri, x\u00fcsusil\u0259 GitHub Copilot v\u0259 ox\u015far k\u00f6m\u0259k\u00e7il\u0259r g\u00fcnd\u0259lik inki\u015faf prosesinin ayr\u0131lmaz hiss\u0259sin\u0259 \u00e7evrilib. Lakin bir \u00e7ox t\u0259rtibat\u00e7\u0131 bu al\u0259tl\u0259rd\u0259n istifad\u0259 ed\u0259rk\u0259n g\u00f6zl\u0259nilm\u0259z, qeyri-d\u0259qiq v\u0259 ya m\u00f6vcud layih\u0259 arxitekturas\u0131na uy\u011fun g\u0259lm\u0259y\u0259n kod t\u0259klifl\u0259ri il\u0259 qar\u015f\u0131la\u015f\u0131r. GitHub m\u00fch\u0259ndisl\u0259rinin son t\u00f6vsiy\u0259l\u0259ri v\u0259 t\u0259cr\u00fcb\u0259l\u0259ri g\u00f6st\u0259rir ki, AI-\u0131n kod keyfiyy\u0259ti birba\u015fa ona \u00f6t\u00fcr\u00fcl\u0259n kontekstin<\/strong> d\u0259qiqliyind\u0259n as\u0131l\u0131d\u0131r. \u0130r\u0259li s\u00fcr\u00fcl\u0259n \u0259sas prinsip sad\u0259dir: i\u015f sah\u0259nizi (workspace) laz\u0131ms\u0131z fayllardan t\u0259miz saxlay\u0131n v\u0259 s\u00fcni intellekt\u0259 n\u0259y\u0259 diqq\u0259t yetirm\u0259li oldu\u011funu d\u0259qiq g\u00f6st\u0259rin.<\/p>\n

Texniki Arxa Plan: Kontekst P\u0259nc\u0259r\u0259si v\u0259 “Diqq\u0259t Mexanizmi”<\/h2>\n

B\u00f6y\u00fck Dil Modell\u0259ri (LLM) t\u0259rtibat\u00e7\u0131n\u0131n redaktorunda ba\u015f ver\u0259n h\u0259r \u015feyi bir anda qavram\u0131r. Modell\u0259r m\u0259hdud sayda token q\u0259bul ed\u0259 bil\u0259n “kontekst p\u0259nc\u0259r\u0259si” (context window) \u00e7\u0259r\u00e7iv\u0259sind\u0259 i\u015fl\u0259yir. GitHub Copilot arxa planda i\u015fl\u0259y\u0259rk\u0259n yaln\u0131z kursorun oldu\u011fu fayl\u0131 deyil, h\u0259m d\u0259 redaktorda a\u00e7\u0131q qalm\u0131\u015f dig\u0259r tablar\u0131 (open tabs), son d\u0259yi\u015fiklikl\u0259ri v\u0259 qon\u015fu s\u0259n\u0259dl\u0259ri skan ed\u0259r\u0259k x\u00fcsusi bir kontekst matrisi formala\u015fd\u0131r\u0131r.<\/p>\n

\u018fg\u0259r redaktorda cari tap\u015f\u0131r\u0131qla \u0259laq\u0259si olmayan 15\u201320 f\u0259rqli fayl (m\u0259s\u0259l\u0259n, konfiqurasiya fayllar\u0131, k\u00f6hn\u0259 miqrasiyalar, as\u0131l\u0131 olmayan modullar) a\u00e7\u0131qd\u0131rsa, AI-\u0131n diqq\u0259t mexanizmi (attention mechanism) “\u00e7irkl\u0259nir”. N\u0259tic\u0259d\u0259, model cari arxitekturaya fokuslana bilmir v\u0259 s\u0259hv sinif strukturlar\u0131, k\u00f6hn\u0259lmi\u015f API \u00e7a\u011f\u0131r\u0131\u015flar\u0131 v\u0259 ya \u00fcmumiyy\u0259tl\u0259 hall\u00fcsinasiyalar istehsal edir.<\/p>\n

Effektiv Fayl \u0130dar\u0259etm\u0259si v\u0259 @mentions G\u00fcc\u00fc<\/h2>\n

GitHub Copilot Chat kimi m\u00fcasir interfeysl\u0259rd\u0259 konteksti idar\u0259 etm\u0259yin \u0259n effektiv yolu @mentions<\/code> sistemind\u0259n v\u0259 tab gigiyenas\u0131ndan ke\u00e7ir:<\/p>\n

    \n
  • Tab T\u0259mizliyi:<\/strong> Yaln\u0131z \u00fcz\u0259rind\u0259 i\u015fl\u0259diyiniz arxitekturaya aid fayllar\u0131 a\u00e7\u0131q saxlay\u0131n. \u018flaq\u0259siz fayllar\u0131n ba\u011flanmas\u0131 modelin diqq\u0259tini yaln\u0131z vacib olan komponentl\u0259r\u0259 c\u0259ml\u0259yir.<\/li>\n
  • D\u0259qiq Fayl H\u0259d\u0259fl\u0259m\u0259si:<\/strong> Sual ver\u0259rk\u0259n \u00fcmumi m\u00fchit \u0259v\u0259zin\u0259 konkret fayllara istinad etm\u0259k \u00fc\u00e7\u00fcn #file<\/code> v\u0259 ya @file<\/code> strukturundan istifad\u0259 etm\u0259k kontekst s\u0259m\u0259r\u0259liliyini art\u0131r\u0131r.<\/li>\n
  • Sistem S\u0259viyy\u0259li \u0130stiqam\u0259tl\u0259ndirm\u0259:<\/strong> @workspace<\/code> \u0259mri model\u0259 b\u00fct\u00fcn repozitoriyan\u0131n strukturunu v\u0259 indeksini t\u0259hlil etm\u0259k imkan\u0131 verir, lakin bu zaman da sor\u011funu spesifik saxlamaq \u015f\u0259rtdir.<\/li>\n<\/ul>\n

    Add\u0131m-Add\u0131m B\u0259l\u0259d\u00e7i: Kontekstd\u0259n Maksimum \u0130stifad\u0259 Qaydalar\u0131<\/h2>\n

    GitHub Copilot v\u0259 VS Code m\u00fchitind\u0259 \u0259n y\u00fcks\u0259k d\u0259qiqlikd\u0259 kod generasiyas\u0131 \u0259ld\u0259 etm\u0259k \u00fc\u00e7\u00fcn a\u015fa\u011f\u0131dak\u0131 ard\u0131c\u0131ll\u0131\u011f\u0131 t\u0259tbiq edin:<\/p>\n

    Add\u0131m 1: A\u00e7\u0131q Tablar\u0131 Filtrl\u0259yin<\/h3>\n

    \u00dcz\u0259rind\u0259 i\u015fl\u0259diyiniz modul il\u0259 birba\u015fa \u0259laq\u0259si olmayan b\u00fct\u00fcn fayllar\u0131 redaktorda ba\u011flay\u0131n. M\u0259s\u0259l\u0259n, yaln\u0131z \u00f6d\u0259ni\u015f servisinin arxitekturas\u0131n\u0131 qurursunuzsa, k\u00f6hn\u0259 autentifikasiya testl\u0259rini v\u0259 ya frontend konfiqurasiyalar\u0131n\u0131 a\u00e7\u0131q qoymay\u0131n.<\/p>\n

    Add\u0131m 2: Repozitoriya Kontekstini Qura\u015fd\u0131r\u0131n<\/h3>\n

    GitHub Copilot-un layih\u0259 standartlar\u0131n\u0131 anlamas\u0131 \u00fc\u00e7\u00fcn repozitoriyan\u0131n k\u00f6k qovlu\u011funda x\u00fcsusi t\u0259limat fayl\u0131 yarad\u0131n:<\/p>\n

    # Repozitoriyan\u0131n k\u00f6k\u00fcnd\u0259 Copilot \u00fc\u00e7\u00fcn x\u00fcsusi t\u0259limat fayl\u0131 yarad\u0131n\nmkdir -p .github\ntouch .github\/copilot-instructions.md\n<\/code><\/pre>\n
    Bu fayl\u0131n daxilind\u0259 layih\u0259nizin arxitektur prinsipl\u0259rini (m\u0259s\u0259l\u0259n: Clean Architecture, istifad\u0259 edil\u0259n kitabxanalar\u0131n versiyalar\u0131 v\u0259 qada\u011fan olunmu\u015f patternl\u0259r) qeyd edin. Copilot avtomatik olaraq bu fayl\u0131 \u0259sas kontekst kimi q\u0259bul ed\u0259c\u0259kdir.<\/p>\n
    Add\u0131m 3: Chat \u0130nterfeysind\u0259 `@` v\u0259 `#` Sintaksisind\u0259n \u0130stifad\u0259 Edin<\/h3>\n
    AI il\u0259 \u00fcnsiyy\u0259t zaman\u0131 konteksti dumanl\u0131 buraxmay\u0131n. T\u0259l\u0259binizi birba\u015fa fayllara ba\u011flay\u0131n:<\/p>\n
    # Copilot Chat daxilind\u0259 n\u00fcmun\u0259 sor\u011fu strukturu:\n@workspace \/explain #file:src\/services\/payment.ts daxilind\u0259ki tranzaksiya ax\u0131n\u0131n\u0131 nec\u0259 optimalla\u015fd\u0131ra bil\u0259r\u0259m?\n<\/code><\/pre>\n
    Add\u0131m 4: Terminal Kontekstini \u018flav\u0259 Edin<\/h3>\n
    H\u0259r hans\u0131 bir kompilasiya v\u0259 ya test x\u0259tas\u0131 il\u0259 qar\u015f\u0131la\u015fd\u0131qda, kodu yenid\u0259n yazmaq \u0259v\u0259zin\u0259 x\u0259tan\u0131 terminaldan birba\u015fa kontekst\u0259 \u00f6t\u00fcr\u00fcn:<\/p>\n
    # Terminal x\u0259talar\u0131n\u0131 analiz etm\u0259k \u00fc\u00e7\u00fcn:\n@terminal Son icra olunan test x\u0259tas\u0131n\u0131 analiz et v\u0259 #file:src\/tests\/payment.test.ts \u00fc\u00e7\u00fcn d\u00fcz\u0259li\u015f t\u0259klif et\n<\/code><\/pre>\n
    Sektora v\u0259 Proqram M\u00fch\u0259ndisliyin\u0259 T\u0259siri<\/h2>\n
    S\u00fcni intellektin kod yazma prosesin\u0259 c\u0259lb olunmas\u0131 proqram\u00e7\u0131lar\u0131n rolunu “kod yazandan” daha \u00e7ox “kontekst idar\u0259edicisin\u0259” (Context Engineer) \u00e7evirir. T\u0259rtibat\u00e7\u0131n\u0131n AI-a n\u0259 q\u0259d\u0259r informasiya verm\u0259si deyil, hans\u0131 informasiyan\u0131 verm\u0259m\u0259si h\u0259lledici amil\u0259 \u00e7evrilir.<\/p>\n
    \u015eirk\u0259tl\u0259r \u00fc\u00e7\u00fcn bu yana\u015fma bir ne\u00e7\u0259 m\u00fch\u00fcm \u00fcst\u00fcnl\u00fck t\u0259min edir:<\/p>\n
      \n
    • Kod \u0130cmallar\u0131nda (Code Review) Vaxta Q\u0259na\u0259t:<\/strong> D\u00fczg\u00fcn kontekstl\u0259 yaz\u0131lm\u0131\u015f kodlar \u015firk\u0259tin m\u00f6vcud dizayn patternl\u0259rin\u0259 d\u0259rhal uy\u011funla\u015fd\u0131\u011f\u0131 \u00fc\u00e7\u00fcn refaktorinq ehtiyac\u0131n\u0131 minimuma endirir.<\/li>\n
    • Token v\u0259 Resurs Q\u0259na\u0259ti:<\/strong> \u015ei\u015firdilmi\u015f kontekst h\u0259m gecikm\u0259ni (latency) art\u0131r\u0131r, h\u0259m d\u0259 hesablama x\u0259rcl\u0259rini y\u00fcks\u0259ldir. T\u0259miz i\u015f sah\u0259si daha s\u00fcr\u0259tli cavab m\u00fcdd\u0259ti dem\u0259kdir.<\/li>\n
    • T\u0259hl\u00fck\u0259sizlik:<\/strong> AI-\u0131n diqq\u0259tini yaln\u0131z t\u0259l\u0259b olunan komponentl\u0259r\u0259 y\u00f6nl\u0259ndirm\u0259k, t\u0259sad\u00fcf\u0259n konfidensial konfiqurasiya fayllar\u0131n\u0131n v\u0259 ya a\u00e7arlar\u0131n model\u0259 \u00f6t\u00fcr\u00fclm\u0259 riskini azald\u0131r.<\/li>\n<\/ul>\n
      BITEP X\u00fclas\u0259si<\/h2>\n
      GitHub-un diqq\u0259t \u00e7\u0259kdiyi bu qaydalar s\u00fcni intellektd\u0259n istifad\u0259 v\u0259rdi\u015fl\u0259rimizi yenid\u0259n n\u0259z\u0259rd\u0259n ke\u00e7irm\u0259y\u0259 m\u0259cbur edir. \u018fn g\u00fccl\u00fc model bel\u0259 nizams\u0131z i\u015f sah\u0259si v\u0259 h\u0259ddind\u0259n art\u0131q laz\u0131ms\u0131z fayl ax\u0131n\u0131 qar\u015f\u0131s\u0131nda faydas\u0131z qala bil\u0259r. Daha t\u0259miz i\u015f m\u00fchiti, m\u0259qs\u0259dy\u00f6nl\u00fc @mentions<\/code> istifad\u0259si v\u0259 layih\u0259y\u0259 uy\u011fun konfiqurasiya fayllar\u0131 GitHub Copilot-un arxitekturaya tam uy\u011fun, istehsala haz\u0131r kod yazmas\u0131n\u0131n \u0259sas a\u00e7ar\u0131d\u0131r.<\/p>\n","protected":false},"excerpt":{"rendered":"
      S\u00fcni \u0130ntellekt D\u0259st\u0259kli Proqramla\u015fd\u0131rmada Kontekst \u0130nqilab\u0131: Niy\u0259 H\u0259r \u015eey D\u00fczg\u00fcn Giri\u015fd\u0259n As\u0131l\u0131d\u0131r? M\u00fcasir proqram t\u0259minat\u0131 m\u00fch\u0259ndisliyind\u0259 s\u00fcni intellekt (AI) al\u0259tl\u0259ri, x\u00fcsusil\u0259 GitHub Copilot v\u0259 ox\u015far k\u00f6m\u0259k\u00e7il\u0259r g\u00fcnd\u0259lik inki\u015faf prosesinin ayr\u0131lmaz hiss\u0259sin\u0259 \u00e7evrilib. Lakin bir \u00e7ox t\u0259rtibat\u00e7\u0131 bu al\u0259tl\u0259rd\u0259n istifad\u0259 ed\u0259rk\u0259n g\u00f6zl\u0259nilm\u0259z, qeyri-d\u0259qiq v\u0259 ya m\u00f6vcud layih\u0259 arxitekturas\u0131na uy\u011fun g\u0259lm\u0259y\u0259n kod t\u0259klifl\u0259ri il\u0259 qar\u015f\u0131la\u015f\u0131r. GitHub m\u00fch\u0259ndisl\u0259rinin […]<\/p>\n","protected":false},"author":1,"featured_media":11381,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-11382","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"aioseo_notices":[],"jetpack_featured_media_url":"https:\/\/bitep.net\/blog\/wp-content\/uploads\/2026\/10\/bitep-ai-1791074092147-21.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/posts\/11382","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/comments?post=11382"}],"version-history":[{"count":0,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/posts\/11382\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/media\/11381"}],"wp:attachment":[{"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/media?parent=11382"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/categories?post=11382"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/bitep.net\/blog\/wp-json\/wp\/v2\/tags?post=11382"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}