Tag analytics events with is_bot (54 pages) + fix a SyntaxError in the homepage snippet - #3
Merged
Merged
Conversation
…snippet Two changes, both in the inline PostHog snippet. 1. is_bot flag (54 pages) voicelogpro.com captured no bot signal, so traffic reports could not separate crawlers from customers. posthog-js only blocks a narrow default list (major search crawlers, headlesschrome, cypress); curl, python-requests, ClaudeBot, CCBot, Applebot, facebookexternalhit, Slackbot and generic bot/spider agents are all captured today and counted as human. Implemented as `before_send`, which is installed at init time and runs for EVERY event -- including the automatic $pageview fired inside init(). A posthog.register() call placed after init would miss exactly that event. The verdict is cached on window, so the UA list is scanned once per page rather than once per event. The inserted code is a single self-contained expression inside the options object, so it is safe even where init sits in a statement position. 2. Homepage snippet did not parse index.html shipped `e.__SV=!0}(document,...` -- missing the closing paren that matches `e.__SV||(`. That inline script threw `SyntaxError: Unexpected token '}'`, so its posthog.init never ran. Restoring the single `)` makes it parse; verified with `node --check`. The same defect exists on carshake and was found while sweeping. Every rewritten page's inline script was syntax-checked against its own pre-edit baseline: 0 regressions. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
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
2 participants
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.
Why
voicelogpro.com captured no bot signal at all, so traffic reports could not separate crawlers from customers.
posthog-js's built-in blocklist is narrower than it looks. It covers major search crawlers plus
headlesschromeandcypress— but notcurl,python-requests,ClaudeBot,CCBot,Applebot,Amazonbot,facebookexternalhit,Slackbot,Twitterbot, or genericbot/spider. Those are all captured today and counted as human. So the flag has real signal with no change to what gets captured.Why
before_sendand notposthog.register()This is a snippet-based static site:
posthog.initis inlined into every page, and the shared/ux.jsisdefer— it runs after the inline snippet already fired$pageview. There is no single shared file to hook.before_sendis installed at init time and runs for every event, so it also covers the automatic$pageviewfired insideinit(). Aregister()call placed after init would miss exactly that event — the most important one. The verdict is cached onwindow, so the UA list is scanned once per page, not once per event.Safety
The inserted code is a single self-contained expression inside the options object. An earlier version declared the detector as a statement before
posthog.init(...); that silently breaks pages where init sits in a statement position (if(cond)posthog.init(...)) — the declaration becomes the if-body and the call loses its guard. My syntax guard caught that before anything was written.Every rewritten page's inline script was extracted and checked with
node --checkagainst its own pre-edit baseline, so the gate is "did this sweep regress a page that previously parsed", not an absolute pass/fail: 0 regressions.Deliberately biased toward false negatives — mislabelling a real visitor as a bot hides a lead, which costs more than an over-counted pageview.
electronis excluded from the UA list for that reason.Property name and boolean type match sanctionsai.dev and hirenika.com, so
WHERE is_bot = falseworks portfolio-wide. No backfill is possible — queries spanning the cutover needis_bot != true.🤖 Generated with Claude Code
index.htmlshippede.__SV=!0}(document,...— missing the closing paren that matchese.__SV||(. That inline script threwSyntaxError: Unexpected token '}', so itsposthog.initnever ran.Restoring the single
)makes it parse (verified withnode --check). This was found while sweeping; carshake has the same defect on 121 pages and is being fixed separately.