SShortSingh.
Back to feed

Bol's dev team shares hard lessons from migrating Backstage platform to pnpm

0
·54 views

Developers at Dutch retailer bol migrated their internal Backstage developer platform from Yarn 4 to pnpm after five years, seeking better support for parallel Git worktree workflows used with coding agents. While the initial switch — pinning pnpm, replacing commands, and committing a new lockfile — was straightforward, CI pipelines repeatedly downloaded all 4,226 packages from scratch despite a cache being in place. The root cause was a mismatch between where pnpm wrote its store and where GitLab expected to find it, resolved by explicitly setting the store directory within the CI configuration. A further issue emerged when the cached archive ballooned to 1.42 GB across roughly 601,000 files because node_modules trees were included unnecessarily; dropping those reduced cache size by about 72%. The team concluded that package manager migrations surface hidden architectural assumptions, particularly around cache paths and what is actually worth caching.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Guide Shows How to Scrape LinkedIn Job Postings Without Logging In

A developer tutorial explains how to collect LinkedIn job listings using public pages that are visible to logged-out visitors. It describes two sources: a paginated search endpoint that returns job cards with IDs, titles, companies, locations and posting dates, and individual job pages that include descriptions, criteria, applicant counts and sometimes salary. The author notes that some fields, such as external apply links, workplace type, job function and the poster, are only available to signed-in users or must be inferred from the text. The piece includes Python code using requests and BeautifulSoup, and the author discloses that they maintain a commercial LinkedIn jobs scraper on Apify. The approach was tested in November 2026.

0
ProgrammingDEV Community ·

dol8: A Python CI/CD gate that blocks pull requests with energy regressions

dol8 is an open-source CI/CD gate that stops pull requests from increasing a Python program's CPU energy use. It reads real RAPL energy measurements on bare-metal Linux and falls back to perf counters on virtual machines or CPU-time estimates on Windows. Its measurement core is written in Rust via PyO3, keeping the measurement separate from Python's own overhead, while Python handles statistics and orchestration. Every result is labeled with its measurement source and confidence, the gate avoids failing builds on noisy data, and comparisons across measurement types are blocked. The project is at v0.4-alpha under the MIT license, validated on AMD Ryzen, with Intel validation still in progress.

0
ProgrammingDEV Community ·

Kaze: A Go Terminal Client That Searches Multiple Sources and Drives mpv Over IPC

After repeated gaps in ani-cli, the developer built Kaze v4.3.0, a single static Go binary for Windows, macOS, and Linux that searches several sources at once. Queries run concurrently under a 3-second timeout, with results streamed into a Bubble Tea UI, and AUTO mode switches to another mirror when a CDN fails. The app controls mpv through IPC, using Unix sockets on Linux and macOS and Windows Named Pipes on Windows, which avoids shell injection and enables live shader swaps and skip-chapter controls. A security audit also removed an unsafe mpv TLS-bypass flag and added IP checks at dial time to block SSRF and DNS rebinding attacks.

0
ProgrammingDEV Community ·

A Single Line Made a Python Queue 2,500 Times Slower: Why list.pop(0) Is Costly

A developer found that processing one million tasks from a Python list took about 58 seconds because each pop(0) call shifted every remaining element. Replacing the list with collections.deque and using popleft() cut the runtime to roughly 0.023 seconds on Python 3.15.2 on an Apple M4. The article explains that list removal and insertion at the front cost O(n), so total work grows quadratically as data increases. It also covers related slowdowns, such as repeatedly building lists with out = out + [x] and string concatenation with += when the string has extra references. The author recommends deque for queues, append or += for lists, and str.join for building strings.