first stab at templating app_domain and app_name strings#627
Merged
Conversation
jpolitz
reviewed
May 23, 2026
|
|
||
| # Set absolute paths for tests (relative paths break when ChromeDriver spawns Chrome) | ||
| echo "GOOGLE_CHROME_BINARY=$GITHUB_WORKSPACE/chrome-linux64/chrome" >> $GITHUB_ENV | ||
| echo "CHROMEDRIVER_BINARY=$GITHUB_WORKSPACE/chromedriver-linux64/chromedriver" >> $GITHUB_ENV |
Member
There was a problem hiding this comment.
I'm pretty sure this is breaking the CI.
What was failing? Like what's the “Claude says”? This dynamic “fetch-chromedriver” has been working for years.
Contributor
Author
There was a problem hiding this comment.
Here's Claude's reasoning behind the fix (copy/pasted):
--
- Relative → absolute path: ./chrome-linux64/chrome → $GITHUB_WORKSPACE/chrome-linux64/chrome. ChromeDriver spawns Chrome as a subprocess and the working directory isn't guaranteed to be the repo root.
- ChromeDriver extracted and wired up directly: Now we unzip chromedriver-linux64.zip and set CHROMEDRIVER_BINARY so util.js uses chrome.setDefaultService() with the exact binary, bypassing the npm chromedriver package entirely (and those two npm install chromedriver steps that were adding noise).
- --disable-dev-shm-usage added: Chrome in Docker/CI containers shares /dev/shm by default (64MB limit) which is too small and causes crashes. This flag disables that.
jpolitz
added a commit
that referenced
this pull request
May 23, 2026
CC @schanzer Turns out this was "working" because package-lock.json was out of date, so the commands that were present were trying to update. This forces the rebuild (which was already kind of what was happening in #627 (comment)) *and* makes updates to util.js so that the environment variables work out. (I think) the problem with the PR was that util was running with the default headless chrome in the CI build, which is not the one that was just unzipped
This reverts commit 5771215.
Member
|
Thanks! I'm trying to fix the CI directly on horizon. I reverted the CI fix that's here. |
jpolitz
reviewed
May 23, 2026
| <title>{{APP_NAME}}</title> | ||
| <link crossorigin="anonymous" rel="preload" href="{{&PYRET}}" as="script"> | ||
| <script>window.PYRET = "{{&PYRET}}";</script> | ||
| <script>window.PYRET = "{{&PYRET}}"; window.APP_NAME = "{{APP_NAME}}"; window.APP_DOMAIN = "{{APP_DOMAIN}}";</script> |
Member
There was a problem hiding this comment.
Do we need the & here? Any escaping that we should include/avoid? Should probably justify the difference.
jpolitz
reviewed
May 23, 2026
| <head> | ||
| <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/> | ||
| <title>code.pyret.org</title> | ||
| <title>{{APP_NAME}}</title> |
Member
There was a problem hiding this comment.
Dumb question: does faq get templated by server.js? I forget which pages actually get all the template variables. Is it all of them?
Contributor
Author
There was a problem hiding this comment.
Looks like it doesn't:
app.get("/close.html", function(_, res) { res.render("close.html"); })
Is there any reason not to pass defaultOpts to all of them?
github-actions Bot
pushed a commit
that referenced
this pull request
Jun 25, 2026
Merge brownplt/code.pyret.org@horizon into the code.pyret.org/ subtree to pick up upstream work the monorepo import had missed: template-strings (#627), menu-item removal, sharedPrefetch non-2xx fallback, privacy.html removal, chart-save + sticky-scroll copy-icon fixes, chromedriver bump. Conflicts resolved: - package.json: kept the monorepo dependency-audit version -- pyret-lang and pyret-codemirror-mode git deps stay removed (resolved via ../lang and ../codemirror-mode siblings instead). - test-util/util.js: took upstream's dedup of the GOOGLE_CHROME_BINARY block; kept our CHROMEDRIVER_BINARY fix above it. - package-lock.json: regenerated from the merged package.json (npm install --package-lock-only), preserving the audit overrides. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.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
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.
Trying to template out strings that are user-facing vs. code-facing. This would make diffs with pyret.bootstrapworld.org a little easier, assuming I have the semantics right