Repository navigation
Tui round3 - #108
Merged
Merged
Tui round3#108
Conversation
m opens a card describing the run that just finished: what it allocated, what it left live, how many objects and collections that took, and how much the process holds from the OS. A modal rather than another status-line field, since that line already carries the pane hints, the follow badge and the search prompt. The run is fenced three times - before the scan, after it, and once the document has been rendered as JSON and YAML - so the card separates what codescan spent from what serializing its result spent on top. Two things the figures are not, stated on the card where they are read: the window is process-wide, because the redraw loop and the file watcher allocate on other goroutines meanwhile; and a rescan is measured while the document it replaces is still live, so its retained figure reads high by about one spec. Duration formatting moves to a humanize package alongside the new byte and count forms, so the header line and the card cannot drift into two spellings of the same measurement. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Signed-off-by: Frederic BIDON <fredbi@yahoo.com>
The run-cost card reports process-wide scalars, and a scalar cannot say who spent it: while a scan runs, the redraw loop and the file watcher allocate on other goroutines, and no arithmetic separates them out afterwards. Start with -profile and each scan is profiled as well, so the same noise arrives named — a row to discount rather than a confound. The card then reports, per phase, where the CPU went and what allocated the memory, by function; a split line at the top recaps the run as ratios (time, memory, and sampled CPU) so the phase to look at is the first thing read. It scrolls once it outgrows the terminal. Both phases are bracketed the same way: a CPU profile each, and an allocation profile differenced across the fences the cost figures already use. Rendering is often shorter than the sampler's 10 ms interval and catches nothing, which the card states — an absent section reads as an oversight, "nothing sampled" is a measurement. Flags rather than options-popup rows: runtime.MemProfileRate is meant to be set once and held for the life of the process, so a mid-session toggle would make consecutive readings incomparable. -profile-dir says where the five artifacts land, and the card carries the go tool pprof commands that open them, -base included so either phase reproduces in the real tool. What the mode costs itself is disclosed rather than hidden: it collects at each fence to keep the allocation profile current, so the summary figures are worse than the same run unprofiled. The profiler's own allocations — flushing one phase's profile through a compressor, allocating the next one's buffers — are dropped from the tables, where in the short phase they would otherwise crowd out its real sites. github.com/google/pprof/profile reads the profiles back. It is a self-contained package with no transitive requirements, and it joins this module rather than the library. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Signed-off-by: Frederic BIDON <fredbi@yahoo.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change type
Please select: 🆕 New feature or enhancement|🔧 Bug fix'|📃 Documentation update
Short description
Fixes
Full description
Checklist