Wrapper architecture smoke test
Side-by-side compare of two parser contexts. Recipient: {email}
🌐 In TOP-LEVEL content_html (API parser context)
Predicted: apisecret is empty here, so the broken hash matches what /blast bakes into production newsletters today.
| Probe | Got | Predicted |
|---|---|---|
{strlen(apisecret)} |
{strlen(apisecret)} | 0 |
{md5(email)} |
{md5(email)} | 591ac677c2562831cb79d75e4c3f1403 |
{md5(email+apisecret)} |
{md5(email+apisecret)} | 591ac677… (same as md5(email) — confirms bug) |
{u(email)} |
{u(email)} | triet%2Bdummy%40workweek.com |
Multi-pass var probes
The script passes probe_zephyr with the literal value
"{md5(email+apisecret)}" — braces and all.
| Substitution | Got | Means |
|---|---|---|
{probe_text} |
{probe_text} | Sanity — should be HELLO_FROM_VARS |
{probe_zephyr} |
{probe_zephyr} |
Hash → multi-pass / vars eval-at-parse-time. Literal text → single-pass. |
{probe_zephyr|raw} |
{probe_zephyr|raw} |
Control · {optout_confirm_url}
sent via /blast with content_html embedding
{include "wm-newsletter-wrapper-test"}