FujiyNotepad v4.11.0
A fast, read-only viewer for very large text and log files on Windows. Self-contained Native AOT build — no install, no .NET runtime required. Download the zip for your architecture below, unzip, and run FujiyNotepad.WinUI.exe.
Highlights
-
Much faster Filter on big logs. Filtering a huge file down to the lines matching a plain (non-regex) term now uses the same vectorized engine that Find uses, instead of decoding every line. On the engine benchmark a rare-literal filter went from ~213 ms to ~6 ms (~33x faster) and from ~196 MB to under 1 KB allocated — so narrowing a multi-million-line log to a handful of lines is near-instant. Regex filters are unchanged.
-
Selection highlighting now works during a search. Selecting a word while the Find bar is open now also highlights that word's other occurrences, alongside the search matches (previously the selection highlight switched off whenever Find was active).
-
Scrollbar stays usable with many matches. Find-match and bookmark ticks are now drawn in a slim overview strip on the inner edge of the scrollbar instead of across the whole track, so they no longer cover the scrollbar thumb when a search matches thousands of lines.
-
New Spark / YARN highlight preset. The Highlight Rules dialog's Insert preset menu has a preset tuned for Apache Spark driver/executor logs: it emphasises scheduler and lifecycle events and failures while dimming the repetitive token/SAS boilerplate.
Under the hood
- Less memory churn and GC pressure across the hot paths: the ~1 MiB search buffer is pooled, on-screen line layout takes an allocation-free fast path for plain lines, full-file Filter/export no longer evicts the viewport's cached lines, and random navigation allocates roughly half as much per checkpoint block.
- Internal cleanup: the Find/Filter option toggles share one style, and the file open/close paths share their reset helpers.
A behaviour note on the faster Filter: like Find, it has no per-line 64 KB cap, so a plain-term filter can now match a term that occurs beyond a very long line's first 64 KB (the old per-line filter could miss it) — making Filter and Find consistent.
See CHANGELOG.md for the full list.