With the shift to lexical-binding in org-clock.el (commit 35bc1c8853c5cfe37930e3540a944a328af5198b), the function org-load-clock fails to load the stored data in org-clock-persist-file.
These are the relevant setting in my .emacs: (setq org-clock-persist t) (setq org-clock-persist-file "~/data/org-clock-save.el") (org-clock-persistence-insinuate) Expected result: After calling the agenda for the first time, the variable org-clock-history is populated with data. Actual result: org-clock-history remains nil. As far as I can tell, the problem lies with the call to load-file (line 2986). Because of the lexical binding, the (setq stored-clock-history ....) statement stored in the org-clock-persist-file is evaluated but because of the lexical scoping it has no effect on the stored-clock-history in org-clock-load. Ironically, this also means that stored-clock-history is defined as an irrelevant global variable even while it remains undefined within org-load-clock as a local variable. (Before lexical binding, the opposite was the case: the call to load-file bound stored-clock-history locally, but not globally). Best, Matt