Again (as earlier last week) it happened that some of the postings to R-help were using a version of unicode that seemed invalid to the unicode-handler (for the digest) inside Mailman (the mailing list software). As a consequence, yesterday's and today's daily digests were not sent out. This time, it ended up better and worse: Better is that, thanks to google, I found a patch to the python script inside mailman which made it not to choke any more on the invalid code *) ==> The problem should not happen anymore in the future. Worse I did this time by fouling up the manual digest resending, such that 3 digests were sent out today, and in wrong order, too. With apologies for any confusion to the R-help digest readers, Martin Maechler, ETH Zurich
*) interestingly these problems only happened for the first time last week, even though we've been using mailman 2.1.6 for a long time. Interestingly too, that the fix to the problem I found with google is only 5 weeks old, too. My guess is that only quite recently some e-mail software has been changed to the worse such that these illegal characters are now sent around in e-mails.