I was getting an "abort: No such file or directory" message whenever I issued an hg command that hit the server (e.g. clone, update, incoming, etc). It was only after I'd spent two days that I finally noticed the paths to my cert and private key in my .hgrc config file were wrong. If the message had said something like "cert file specified in /home/scott/hgrc auth.myserver.cert = /etc/ssl/certs/yadda.pem not found", the problem would have been immediately obvious. Thanks
Perhaps we'll need to catch IOError somewhere in sslutil.py, probably load_cert_chain().
I can see how this error could be very annoying. I'll try to get a fix for this in Mercurial 4.3.
Fixed by the following changeset: changeset: 46072:3bdbbadddecc user: Gregory Szorc <gregory.szorc@gmail.com> date: Mon Jul 10 21:09:46 2017 -0700 summary: sslutil: check for missing certificate and key files (issue5598) (Where is the bot that is supposed to be updating bugs?)