On stable Rust, Clippy's wildcard_enum_match_arm flags new variants of a #[non_exhaustive] enum. Once the wildcard shares an arm, as in its own `Variant | _` fix, it stops checking that match. https://importstatic.com/rust/rust-non-exhaustive-enum-new-variants

How to notice new variants of a #[non_exhaustive] Rust enum

A #[non_exhaustive] enum sends new variants into your _ arm without a warning. A nightly rustc lint flags them; Clippy's stable lint skips `Variant | _` arms.

importstatic.com

Go 1.26's go/parser parses type parameters on a method only to report "method must have no type parameters", then drops them. A go/types tool built with Go 1.26 then flags every use of R in a Go 1.27 generic method as "undefined: R".

In the OpenAI CLI, --input @alice is sent as the literal "@alice" only because no file named alice exists. Create one and the CLI sends its contents instead. A name with a dot or slash and no file is an error. We checked v1.40.2.

One generic method, three errors: Go 1.27 with go 1.26 in go.mod says "requires go1.27", Go 1.26 reports a syntax error, and a go/types tool built with Go 1.25 says "undefined: R". https://importstatic.com/go/go-method-type-parameters-error

Go "method must have no type parameters": causes and fixes

Go's "method must have no type parameters" error means an old toolchain or go line, unless it's an interface. Real output from Go 1.25, 1.26 and 1.27.

importstatic.com

The OpenAI CLI reads a file for any flag value starting with @: --input "@.env" sent our .env, even with OPENAI_UNTRUSTED_STDIN=1. Piping the text to --input @- kept it literal. https://importstatic.com/ai/openai-cli-at-file-input

The OpenAI CLI reads your .env if the prompt starts with @

The OpenAI CLI reads a local file whenever a flag value starts with @, even with OPENAI_UNTRUSTED_STDIN=1. How to pass prompts and model output safely.

importstatic.com

PostgreSQL's recovery_prefetch has sighup context, so a lagging standby can try it off with ALTER SYSTEM SET recovery_prefetch = off and SELECT pg_reload_conf(), no restart. wal_decode_buffer_size, its lookahead limit, does need a restart.

To see which WAL records carry full-page images, PostgreSQL 18's pg_walinspect has pg_get_wal_block_info(start_lsn, end_lsn) gives one row per block reference; block_fpi_length > 0 means the page was imaged. After a checkpoint, the first change to every page was.

Postgres 18's WAL prefetcher skips a page whose record carries a full-page image, then prefetches that same page one record later anyway. Is anyone running standbys with recovery_prefetch off? https://importstatic.com/databases/postgres-recovery-prefetch-full-page-writes

PostgreSQL's recovery_prefetch prefetches pages the WAL already holds

recovery_prefetch skips a page with a full-page image in the WAL, then prefetches it one record later. Spot it in pg_stat_recovery_prefetch on PostgreSQL 18.

importstatic.com

Python 3.15 deprecates `import` lines in .pth files, the hook the bad LiteLLM 1.107.0 wheels used to run on every interpreter start. They still execute until 3.18, which ignores them in favour of new .start files. Worth a look at your site-packages.

Same Rust mistake, two error codes. With an error type lacking From<ParseIntError>, rustc 1.99.1 reports raw.parse::<u16>()? as E0277 ("? couldn't convert the error") but let n: u16 = raw.parse()?; as E0271, a type mismatch on FromStr::Err.

The bad LiteLLM 1.107.0 wheels drop a 12-byte cook.pth that says 'import cook'. Python runs that at every start, so even pip uninstall executes the stealer. Clean up the files before you touch pip. https://importstatic.com/ai/litellm-1-107-0-compromise-check

How to check a machine for the compromised LiteLLM 1.107.0

LiteLLM 1.107.0 on PyPI and Docker shipped a credential stealer that runs at every Python start. The affected windows, the files to find and the keys to rotate.

importstatic.com

We switched off tsrs's lazy member tables (ports of two TypeScript PRs) with --noLazyMembers --checkerAssignment go. It got 9% slower and was still 2.7x faster than tsgo 7.1 beta at 4 checkers. Most of the gap isn't those PRs.

On a 12-core Mac, tsrs's default (one checker per core) checked TypeScript's 456-file source slower than --checkers 4: 0.53 s vs 0.48 s median, and 37 MiB more memory. On smaller projects, try fewer checkers first.

Rust's E0277 also appears as "doesn't implement Debug", "? couldn't convert the error" and "cannot be sent between threads". Fixes for each, including two versions of one crate. https://importstatic.com/rust/rust-e0277-trait-bound

Rust error E0277 and "the trait bound is not satisfied", case by case

Rust's E0277 shows up as "the trait bound is not satisfied" and six other messages. The rustc 1.99.1 output for each, and the fix that made it build.

importstatic.com

openai-go v3.83.0 still documents "30m" as the only prompt cache ttl, but sends "12h" without complaint. For a per-breakpoint ttl, which its struct lacks, call SetExtraFields(map[string]any{"ttl": "12h"}) on the breakpoint param.

OpenAI bills each cached token at the longest lifetime of any breakpoint covering it. A 30m breakpoint at 2,000 tokens followed by a 12h one at 5,000 bills all 5,000 as 12-hour writes, at 2.5x input. Put the 12h breakpoint first.

tsrs, the Rust port of the TypeScript 7 compiler, checked TypeScript's own 456-file source 2.7x faster than tsgo 7.1 beta at 4 checkers, with 54% less memory. The published 8x is measured against 7.0. https://importstatic.com/rust/tsrs-vs-tsgo-typescript-7-1

tsrs vs tsgo: timing the Rust port against TypeScript 7.1 beta

tsrs vs tsgo on TypeScript's own 351,000-line source: at 4 checkers the Rust port was 2.7x faster than 7.1 beta and used 54% less memory. The 8x is against 7.0.

importstatic.com

Sending "SET lock_timeout = '2s'; CREATE INDEX CONCURRENTLY ..." as one query string fails on PostgreSQL 18 with SQLSTATE 25001: two statements in one string run as one implicit transaction. Send them as two calls on the same connection.

CREATE INDEX CONCURRENTLY waits for REPEATABLE READ transactions even if they never touched the table. On PostgreSQL 18.6 ours sat in 'waiting for old snapshots' until an idle transaction on another table committed. READ COMMITTED didn't hold it up.

OpenAI's 12-hour prompt cache write costs 2.5x input, so two requests 6 hours apart cost more than with the 30m default. It pays from the third reuse. Set request-wide, a chat cost 68% more. https://importstatic.com/ai/openai-12-hour-prompt-caching

Turn on OpenAI's 12-hour prompt cache for the shared prefix only

OpenAI prompt caching now has a 12-hour lifetime at 2.5x the input price. It pays from the third reuse, and set request-wide it raised a chat's input cost 68%.

importstatic.com

Go's crypto/x509 VerifyOptions has no CRL or OCSP setting. With its own root pool, Go 1.27.2 still verified eMudhra intermediates that CCADB lists as revoked on October 18 and 19. Go 1.27.3 rejects them by public key instead.

An instant ALTER TABLE ADD COLUMN froze every read and write on a Postgres table for 4.5 s. It waited behind one idle transaction, and every query queued behind it. https://importstatic.com/databases/postgres-lock-timeout-migrations

PostgreSQL lock_timeout for schema migrations, with retries

A waiting ALTER TABLE stalled every read and write on our table for 4.5 s. How lock_timeout, retries and CREATE INDEX CONCURRENTLY avoid it on PostgreSQL 18.

importstatic.com

With GOTOOLCHAIN=auto, a module whose go.mod says go 1.26.0 ran on go1.26.0 itself on our older machine, with none of the 1.26 security fixes. go get toolchain@go1.27.3 adds a toolchain line and keeps the go line as it is. https://go.dev/doc/toolchain

The sqlite3 in /usr/bin on our Mac reports sqlite_version() 3.54.0, but its source id is dated 2026-04-09 and it returns NULL for 3.54's 'end of month' modifier. Probe the feature, not the version string.

Go 1.27.3 rejects chains through eight eMudhra CA keys, used by 16 revoked intermediates, half of them cross-signed by Sectigo's AAA root. govulncheck still reports nothing for a 1.27.2 build. https://importstatic.com/go/go-1-27-3-x509-distrust

Go 1.27.3 blocks eight eMudhra CA keys, Sectigo cross-signs included

Go 1.27.3 and 1.26.10 reject any chain through 16 revoked eMudhra intermediates. How the crypto/x509 check works and which builds it doesn't reach.

importstatic.com

Pointing Claude Code at your own proxy with ANTHROPIC_BASE_URL doesn't route everything through it: the docs list the fast mode availability check and the WebFetch domain safety check as calling api.anthropic.com directly. https://code.claude.com/docs/en/llm-gateway-protocol

A Go httputil.ReverseProxy forwards the client's Accept-Encoding. If the client asked for gzip, the Transport hands you the body still gzipped: it only decodes gzip it requested itself. Reading responses in the proxy? Delete the header in Rewrite.

A proxy that logs Claude token usage from message_start records 2,679 input tokens for a web-search request the API bills at 10,682. The real numbers come at the end of the stream. https://importstatic.com/ai/llm-logging-proxy-go

Writing an LLM logging proxy in Go that counts tokens right

A small LLM logging proxy in Go for Claude Code and the Claude API: why the first usage numbers in a stream are wrong, and what breaks streaming.

importstatic.com

date('now','end of month') is new in SQLite 3.54. On 3.53 it returns NULL, no error, so a WHERE built on it matches zero rows. And 'weekday -0' on old builds moves forward a week. https://importstatic.com/databases/sqlite-end-of-month-modifier

SQLite's new 'end of month' modifier returns NULL on older versions

SQLite 3.54 added 'end of month' and 'weekday -N'. Older builds, including the Go and Rust drivers we tried, return NULL with no error. The portable forms.

importstatic.com

SQLite's .diskused is new in 3.54, the data under it isn't: SELECT name, pgsize, unused FROM dbstat WHERE aggregate = TRUE gives bytes and unused bytes per table and index. It ran from Go (modernc.org/sqlite) and rusqlite's bundled build. https://www.sqlite.org/dbstat.html