This is mainly to lay out the structure of the final code. The problem
isn't really solved yet, although some very simple cases are
better (hence the fixed test). We need two passes through the messages
because we need to be careful not to re-parent too many messages and
end up without any toplevel messages.
There is no public notmuch_message_list_t public interface, so to this
is added to the private API. We use it immediately in thread.cc;
future commits will use it further.
For non-root messages, this should not should anything currently, as
the messages are already added in date order. In the future we will
add some non-root messages in a second pass out of order and the
sorting will be useful. It does fix the order of multiple
root-messages (although it is overkill for that).
This documents the bug discussed at
id:87efgmmysi.fsf@len.workgroup
The underlying issue is that the reply to a ghost (missing) message is
falsely classified as a root message in _resolve_thread_relationships.
There are two pairs of tests; in each case the the first test is
simpler / more robust, but also easier to fool.
There are 3 threads here, two synthetic, and one anonymized one using
data from Gregor. They test various aspects of thread
ordering/construction in the presence of replies to ghost messages.
This clarifies that the breakage seen with Xapian 1.4.6 does not have
to do with "funny" tags.
This test is "known broken", but only with xapian 1.4.6, and there's
curently no convenient way to mark that.
Messages that contain Windows-1252 are frequently mislabeled as ISO
8859-1, which may result in non-printable characters when displaying
the message. The test asserts that such characters (in this case
curved quotes) are displayed correctly.
On some platforms (e.g. macOS), it is necessary to add a real sentinel
process for the error buffer used by `notmuch-start-notmuch' rather
than a no-op sentinel.
Many of the manpages didn't treat literal text as literal text. I've
tried to normalize some of the restructured text to make it a bit more
regular.
several of the synopsis lines are still untouched by this cleanup, but
i'm not sure what the right way to represent those is in .rst,
actually.
In particular find that if i rebuild the manpages, sometimes i end up
with some of the synopsis lines showing – (U+2013 EN DASH) where they
should have -- (2 × U+002D HYPHEN-MINUS) in the generated nroff
output, though i have not tracked down the source of this error yet.
If e.g. /lib is a link to /usr/lib, then the latter may not show up in
the way we expect in the output of ldconfig. 'test foo -ef bar' checks
if foo and bar have the same device and inode numbers. Since (at least
in bash, dash, ksh, and zsh) the shell dereferences symlinks before
applying the test, this includes both the case where file1 is equal to
file2 and the case where one is a symlink to the other.
This is mainly to improve discoverability. It seems that doing
variable cross-references is not easy without using some sphinx
extension/customization.
This is a simple way to make the info pages available, since the emacs
package system deals with the annoying bits. If there is many people
who want the info pages outside of emacs, we could consider a separate
binary package, but currently it doesn't seem worthwhile.
