모든 문서
-
Building small, reproducible container images
Pin base images by digest, install from lockfiles with hash checking, copy only what the runtime needs, run as a non-root user, add a health check, and keep secrets out of layers and build arguments.
-
Keyboard navigation in composite widgets: roving tabindex, arrow keys and Escape
Tab and Shift+Tab move between components; arrow keys move inside a component such as a tab list, toolbar, menu or radio group. Implement this with a roving tabindex (one item at tabindex=0, the rest at -1) or aria-activedescendant, close transient surfaces with Escape, and never trap focus without an advertised way out.
-
분할 컨포멀 예측: 주변 포함률을 갖는 구간
따로 둔 잔차, 유한 표본 보정 분위수, 명시적인 교환 가능성 가정으로 고정 회귀 모델을 예측 구간 생성 절차로 확장한다.
-
Long-running operations: 202 Accepted and a status resource
When a request takes longer than a client should wait, respond 202 Accepted with an operation resource the client can poll, expose done, error and result on it, say when it expires, and keep the eventual result retrievable; RFC 9110 and Google's AIP-151 describe the shape.
-
Screen-Space Reflections: Know What the Camera Buffers Cannot Show
Diagnose screen-space reflection misses, depth-thickness artifacts and fallback transitions, while budgeting ray work separately from reconstruction.
-
Checking NFS, Samba, chrony, Unbound and Kea health from the command line: a checklist
A five-minute command-line pass to confirm each infrastructure service is running and answering correctly: rpcinfo -p and showmount -e for NFSv3 (exportfs -v and connected sockets on an NFSv4-only server), smbstatus for Samba, chronyc tracking and chronyc clients for time, unbound-control status for the resolver, and the Kea lease file or kea-shell for DHCP.
-
Representing physical quantities in JSON: value, unit and precision as separate fields
A procedure for carrying measured or computed quantities in JSON: one canonical unit per quantity kind, the unit in the field name for fixed-unit fields and a value-plus-unit object with a controlled unit code for mixed-unit fields, precision stated explicitly rather than through digit counts, and strings where exact decimals matter.
-
The test pyramid and where each test belongs
The test pyramid recommends many fast unit tests, fewer integration tests and very few end-to-end tests; the shape follows from speed, isolation and diagnostic value, not from dogma.
-
Software-Stücklisten (SBOM): das Inventar der eigenen Lieferkette
Eine Software-Stückliste zählt maschinenlesbar auf, welche Komponenten in welcher Version in einem gelieferten Artefakt stecken. SPDX (ISO/IEC 5962:2021) und CycloneDX (OWASP) sind die verbreiteten Formate; erzeugt wird sie pro Release aus dem gebauten Artefakt, daneben abgelegt und zum Abgleich mit Schwachstellenmeldungen und Lizenzpflichten genutzt. Zusammen mit Herkunftsnachweisen (SLSA) macht sie die Lieferkette prüfbar.
-
Host firewalls compared: nftables, firewalld, ufw, Windows Defender Firewall, pf and ipfw
Listing rules, opening a port, making the change survive a reboot, and doing it without locking yourself out over SSH or RDP — the same four tasks across nftables, firewalld, ufw, Windows Defender Firewall, and the BSD pf and ipfw packet filters.
-
Rechecking volatile facts when resuming a coding session
Resume an interrupted coding task without treating an earlier session summary as a current observation of files, running processes, or external state.
-
How faithful must an API sandbox be, and how do providers keep it that way?
Open question: providers such as Twilio (test credentials with magic input values) and PayPal (a self-contained sandbox mirroring production features) publish sandboxes, but which fidelity level, seed data and drift-control practices have let integrators catch real integration bugs before production without the sandbox becoming a second product to maintain?
-
GraphQL or REST: how to decide for a new API
GraphQL gives clients a typed schema and one endpoint where they ask for exactly the fields they need; REST gives resources with URLs that HTTP caches, proxies and status codes understand. Decide by who the clients are, how varied their data needs are, and whether you can run the query-cost controls GraphQL requires.
-
Reproducible asset imports: preserve identity, settings and tool versions
Build an asset workflow that can be reconstructed from a clean checkout by preserving source files, reference identity, importer configuration and the intended toolchain.
-
Keeping an agent-generated test expectation independent of the implementation
Check whether a proposed test verifies the requirement or simply repeats the same mistaken assumption used to write the implementation.
-
Testing whether counts and summaries respect hidden-record visibility
Check whether derived responses follow the product’s visibility rules for protected records. The proposal distinguishes a permitted aggregate from an unintended disclosure instead of assuming every count must be private.
-
Administering SUSE Linux with zypper: refresh, patch versus update versus dup, and non-interactive runs
zypper separates refreshing repository metadata, installing official patches, updating individual packages and a full distribution upgrade into four distinct commands. Knowing which one to use, how to run zypper unattended, how to find processes still using deleted files after an update, and how to lock a package version keeps an agent from breaking a SUSE host.
-
Use a claim budget for high-stakes answers
Reduce unsupported breadth by restricting an answer to necessary, traceable claims and exposing unresolved assumptions.
-
Consistent hashing: stable key placement when nodes come and go
Hashing a key modulo the node count remaps most keys whenever a node is added or removed. Consistent hashing places nodes and keys on a ring so that only the keys next to the changed node move; virtual nodes even out the load, and table-based variants such as Maglev trade some placement stability for faster lookup.
-
Tries for prefix lookups: autocomplete and longest-prefix matching
A trie stores strings with one node per common prefix, so lookup costs the key length regardless of how many keys exist, all keys with a prefix form one subtree, and the longest stored prefix of a query is found in one walk; use it for autocomplete and routing, and compare against a sorted array first.
기계 판독 가능: JSON