acf domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /home2/hyperm16/so3d.com.br/wp-includes/functions.php on line 6170wp-pagenavi domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /home2/hyperm16/so3d.com.br/wp-includes/functions.php on line 6170antispam-bee domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /home2/hyperm16/so3d.com.br/wp-includes/functions.php on line 6170woocommerce-gateway-paypal-express-checkout domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /home2/hyperm16/so3d.com.br/wp-includes/functions.php on line 6170Every few months someone posts the same complaint in a support forum: their mail account works fine in Apple Mail on a MacBook, then fails with a certificate error when the identical credentials are entered into a script or client on a Linux box. The usual suspects get blamed first, IMAP vs POP3 settings, the incoming mail server hostname, the port number, a password with a stray character. None of that is usually the problem. The difference is the platform certificate authority store, and it will break mail client configuration in ways that look exactly like a protocol misconfiguration.
Here is the case that pushed me to write this down. A small team ran an SEO automation tool on an Ubuntu server to file link-prospecting replies into project mailboxes. The tool fetched mail over POP3 from a catch-all domain hosted on a smaller provider, processed the messages, and archived them. It worked for months. Then the provider rotated its TLS certificate, and every fetch started returning a handshake error. The same POP3 account, same port, same password, still opened without complaint in Apple Mail on the team’s Macs.
The instinct was to assume the automation tool had switched something. It had not. The provider had deployed a new leaf certificate but forgot to include the intermediate certificate in the chain it served. macOS, which maintains its own trust store and can often complete an incomplete chain by fetching or caching the missing intermediate, accepted it. The Linux box ran OpenSSL 3, which by default refuses to build an incomplete chain unless the intermediate is either served by the host or present locally in the system CA directory. With the intermediate missing on both sides, the handshake died. Nothing about POP3 vs IMAP, ports, or authentication was wrong. The trust path was.
This is the part most guides skip. They treat certificate validation as a binary, valid or invalid, and assume every platform agrees. It does not. OpenSSL 3 tightened chain building compared with OpenSSL 1.1, so a chain that squeaked through on an older distribution can fail after an upgrade with no configuration change on your side. A mail client on Windows uses the Schannel store. macOS uses the Keychain. A Python script may use certifi’s bundled CA file, which is a fourth store entirely. Four platforms, four answers to the same handshake.
When a fetch fails, the error text tells you which layer broke. A message like “certificate verify failed: unable to get local issuer certificate” is a chain problem, not a credential problem. “Hostname mismatch” points at a certificate issued for a different name than the one in your client config, often the difference between mail.example.com and the bare domain. “Certificate has expired” is self-explanatory. If you see any of these, stop editing ports and protocols and inspect the certificate the server actually serves. On Linux, openssl s_client -connect host:995 -showcerts prints the chain the server sends. If the output shows only the leaf certificate and no intermediate, you have found the cause.
Compare that against what the working Mac sees. If the Mac client succeeds, the chain is probably still incomplete and macOS is compensating. That is the tell. A server that serves a full chain will pass on every platform. A server that serves a partial chain will pass on some and fail on others, which is exactly the pattern that sends people chasing POP3 for SEO automation settings that were never wrong. The fix belongs with the mail host or with your local trust store, not with the client’s protocol choice.
The protocol layer is simpler than the trust layer, and it is worth confirming the basics are correct so you can rule them out. As explained in an AllMail IMAP vs POP3 guide, IMAP keeps mail on the server and synchronizes folders and read states across devices, using encrypted port 993. POP3 downloads mail to a single client, uses encrypted port 995, and supports the inbox only. SMTP handles sending mail and is a separate configuration from either. If your account fetches on one machine and not another, and both use the same protocol and port, the protocol is not the variable. The certificate store is.
On Linux, the system trust anchors live in /etc/ssl/certs, usually symlinked from the ca-certificates package. Check whether the intermediate your provider uses is present. If it is missing, you have two legitimate options: ask the provider to serve the full chain, which is the correct fix because it helps every client, or add the intermediate to your local store and update it. The second option works but must be repeated on every machine and container, and it silently breaks when the intermediate rotates. For a single workstation that is tolerable. For a fleet of automation workers it is a maintenance trap. The SEO tool that defaulted to POP3 was never the issue; the missing intermediate in the served chain was.
On macOS, the Keychain Access utility shows the trust settings, and the system is more forgiving of incomplete chains because it can retrieve missing intermediates. That forgiveness is a feature, not a guarantee, and it hides server misconfiguration from the people best positioned to fix it. A team that only tests on Macs will ship automation that fails the first time it touches a Linux runner. If you maintain catch-all domains for projects through a service like AllMail, verify the chain on the mail host itself rather than trusting that it works because it opens on your laptop. The article covering this on the AllMail blog is worth reading alongside your own s_client output.
A few concrete checks shorten the diagnosis. Confirm the certificate the server presents matches the hostname in your client config, including any subdomain. Confirm the chain includes every intermediate up to a root the target platform trusts. Confirm the client is validating against the store you think it is, because a bundled certifi file and the system store can diverge. Confirm the TLS version and cipher negotiated, since an old client refusing TLS 1.3 or a server insisting on a deprecated cipher can produce failures that masquerade as trust errors. Run these on each platform you deploy to, not just the one on your desk.
One more trap deserves mention. Automation tools and libraries often ship their own CA bundle and ignore the system store entirely. A tool that worked when run by hand can fail under a service account with a different environment, or after the library updates its bundled certificates. When you debug, establish which store the failing process actually reads before you change anything. The same applies to IMAP4rev2, defined by RFC 9051, and to the older POP3 standard in RFC 1939: neither specification says anything about which CAs a client must trust, so protocol compliance tells you nothing about whether a given platform will accept a given chain. That gap is where cross-platform configuration failures live.
Treat the certificate chain as part of your mail configuration, not an afterthought. When an account works on one platform and fails on another, capture the served chain with openssl s_client, compare it against the trust store on the failing platform, and fix the chain at the source by having your provider serve the full intermediate set. Keep the protocol settings, IMAP on 993 with STARTTLS where offered, POP3 on 995 for single-device automation, SMTP for sending, correct and unchanged while you investigate. The next time a fetch fails after a certificate rotation, check the CA store first. Nine times out of ten the protocol was never the problem.
{“@context”:”https://schema.org”,”@type”:”Article”,”headline”:”CA Store Differences Break Mail Client Configuration”,”description”:”Platform CA stores differ: OpenSSL 3 on Linux rejects incomplete chains that macOS trusts. Fix cross-platform mail config by checking CA stores first.”,”inLanguage”:”en”,”datePublished”:”2026-11-16T20:00:22Z”,”articleBody”:”Every few months someone posts the same complaint in a support forum: their mail account works fine in Apple Mail on a MacBook, then fails with a certificate error when the identical credentials are entered into a script or client on a Linux box. The usual suspects get blamed first, IMAP vs POP3 settings, the incoming mail server hostname, the port number, a password with a stray character. None of that is usually the problem. The difference is the platform certificate authority store, and it wi”}