This seems to always happens with MFH. My new theory is that I need to populate the "Differential Revision" for both the commit (which I did) and with MFH commit (which I did not).

This seems to always happens with MFH. My new theory is that I need to populate the "Differential Revision" for both the commit (which I did) and with MFH commit (which I did not).

No reason not to format the description using the preferred style.

Here's an update, please let me know if the <li> guys are formatted ok (I found examples that formatted the way I did but also some that indented multi line<li>'s).

Address mat@'s comments
Populate "reports" and reformat the description.

If you add a blank line before PATCH_SITES the Makefile will pass portlint. Otherwise looks good to me.

I tested my change in a ports tree that had not been updated in a few hours then copied to the ports tree I use for committer work; I lost the race with r472884. I'll try to be more careful moving forward.

Update diff to not step on r472884 PY_FLAVOR change.

I have more than one poudriere jail on my systems, I just stopped using the 10.3-RELEASE ones a few weeks ago.

In D15271#322012, @matthew wrote:

Test builds I ran overnight all failed on 10.3-RELEASE due to 'bad C++ code' in many of the dependencies. Of course, since yesterday I should be testing on 10.4-RELEASE...

(It still says I have 4 unsubmitted comment.)

Remove gcc and python from BUILD_DEPENDS, fully understand how BINARY_ALIAS works and use PATH from MAKE_ENV when building.

Restore comment/header to patch-ssh.c and Bump PORTREVISION (oops).

Accidentally quit from arc without clearing all the fields...

"oops"