Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: cakephp/log
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: 5.x
Choose a base ref
...
head repository: cakephp/log
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: 5.next
Choose a head ref
Checking mergeability… Don’t worry, you can still create the pull request.
  • 1 commit
  • 1 file changed
  • 1 contributor

Commits on Aug 7, 2026

  1. Stop a logger from reentering itself while it is dispatching (#19589) 

    A logger that writes through a subsystem which logs its own failures can be asked
to write again while it is still handling a message. Nothing noticed, so the two
called each other until the process ran out of memory.

The reachable route needs no exotic setup. A logger that stores rows in a table
performs an insert, Cake\Database\Log\QueryLogger picks that insert up and calls
Log::write('debug', ...), and the message comes straight back to the same logger,
which inserts another row. Scopes do not help: BaseLog defaults to an empty
scopes array, and Log::write() treats an empty array as matching everything,
scoped messages included. So a logger configured with nothing but a className is
exposed.

A second route runs through the cache. An engine that cannot reach its backend
logs the failure, and a table-backed logger asking for schema metadata lands back
in the cache pool that is still being built.

Track which streams are dispatching and skip a stream that is already inside a
write. The nested message goes to error_log() rather than being dropped
silently, and the other configured loggers still receive it, so one logger
looping does not silence the rest of the stack. The marker is cleared in a
finally, so a logger that throws does not disable itself for the remainder of the
process.

Keyed per stream rather than a single global flag on purpose. A global flag would
suppress every logger for the duration of any nested write, which is a much
broader change than the problem calls for. The tradeoff is that a cycle running
across two different streams is not caught; that needs a depth counter and is not
worth the complexity until someone hits it.
    @dereuromark
    dereuromark authored Aug 7, 2026
    Configuration menu
    Copy the full SHA
    364533a View commit details
    Browse the repository at this point in the history
Loading