Yes, google is improperly flagging messages as spam (SPF Softfail) and citing the sender's client-IP-address as not an authorized sender. This is wrong on several levels.
The messages in question, were not sent to a google hosted mailbox.
Instead, the gmail user has configured gmail to fetch mail (POP3) from an account hosted on an other service and make those messages available in the gmail users inbox.
As such, these messages are NEVER actually received by a google mail-server. The messages are being fetched by Google, not received. Google should not be rewriting the mail headers as if they received them. If they wish to apply filters, they should only be reading the headers as written by the (non-google) mail server that is hosting the account.
SPF Softfail based on Client-IP: Google is reporting:
"spf=softfail (googlecom: domain of transitioning usr@example.com does not designate x.x.x.x (client-ip) as permitted sender)"
Client-IP is the internet address of the senders physical location (home, business, coffee-shop). It would be impossible to include all possible client-ips in an spf record. SPF check should only consider the IP of the mail-server that delivered the message to the recipients mail server, in this case, a non-google server.
In this example, Debi is the Gmail user. Debi has configured Gmail to fetch messages from an other mailbox 'editor' hosted at supercp.com.
Shawn sends a message to Editor which is received at supercp.com. Debi's gmail connects using POP3 to download the message. Google flags it as spam based on the IP address (residential ISP WAN address) where Shawn happened to be when he had Thunderbird mail client send the message.
Example email headers