Skip to content

Pmk 2059 - #164

Open
ewood-ac wants to merge 6 commits into
mainfrom
pmk-2059
Open

Pmk 2059#164
ewood-ac wants to merge 6 commits into
mainfrom
pmk-2059

Conversation

@ewood-ac

Copy link
Copy Markdown
Contributor

No description provided.

ewood-ac and others added 6 commits January 22, 2026 09:12
The composer constraint moves from `~8.1 || …` to `~8.2 || … || ~8.5`, and three
PostmarkAttachment factories gain `string` parameter types while two PostmarkOpen
getters become nullable. All of those are breaking for consumers and none were
recorded, so a release cut from this branch would have shipped as a minor.

Adds the CHANGELOG entry marking this as v8.0.0 with each break called out.

Also replaces the empty catch in the suppressions teardown with a STDERR warning
— a cleanup failure left the list populated and broke the following run somewhere
unrelated, with nothing on record pointing back here.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
ewood-ac added a commit that referenced this pull request Aug 5, 2026
Three changes to the client-identification headers:

The version was a hand-maintained `public static $SDK_VERSION = '7.0.0'`. That
has to be bumped by hand on every release and will silently go stale — at which
point the telemetry these headers exist to provide is actively wrong, which is
worse than absent. It now reads Composer's runtime metadata, falling back to a
constant only when that is unavailable (source checkout, no installed package).
The property was also public and mutable; the replacement is a method plus a
const.

The new User-Agent dropped the OS field that the old format carried. Anything
parsing that string for platform breakdown would have gone blank without
warning, so OS is back: `Postmark-SDK/<v> (PHP/<ver>; OS/<os>)`.

Also drops this branch's composer.json and CI edits. They overlap with the ones
in #164, which sets a wider range (adding 8.5 as well as dropping 8.1) and so
supersedes them — leaving both would just conflict. The PHP-version decision
belongs to that PR; this one is only about the headers.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant