* lisp/progmodes/dcl-mode.el (dcl-mode):
* lisp/progmodes/idlw-complete-structtag.el: Recommend
with-eval-after-load instead of load-hooks.
* lisp/calc/calc-ext.el (calc-ext-load-hook):
* lisp/emacs-lisp/bytecomp.el (bytecomp-load-hook):
* lisp/emacs-lisp/cl-extra.el (cl-extra-load-hook):
* lisp/emacs-lisp/cl-macs.el (cl-macs-load-hook):
* lisp/emacs-lisp/cl-seq.el (cl-seq-load-hook):
* lisp/gnus/message.el (message-load-hook):
* lisp/gnus/nnheader.el (nnheader-load-hook):
* lisp/gnus/nnmail.el (nnmail-load-hook):
* lisp/progmodes/dcl-mode.el (dcl-mode-load-hook):
* lisp/textmodes/tex-mode.el (tex-mode-load-hook):
* lisp/whitespace.el (whitespace-load-hook): Obsolete for
with-eval-after-load. Note that these variables are never declared,
but the byte-compiler will still warn about them if used.
* lisp/gnus/gnus-icalendar.el (gnus-icalendar-event-from-ical):
Sanitise the data before passing it on to the constructor. This
avoids backtraces on icals with extra, unknown slots (bug#43057).
* doc/misc/gnus.texi (Agent Variables): Document it.
* lisp/gnus/gnus-art.el (gnus-request-article-this-buffer): Ditto.
* lisp/gnus/gnus-async.el (gnus-async-article-callback): Use it.
* lisp/gnus/gnus.el (gnus-agent-eagerly-store-articles): New variable.
Includes work from Madhu <enometh@meer.net>.
* lisp/gnus/gnus-fun.el (gnus-convert-face-to-png): Do it.
* lisp/gnus/gnus-util.el (gnus-base64-repad): Allow not checking
anything, but just repadding.
* lisp/gnus/nnimap.el (nnimap-parse-flags):
Remove old hack that deletes MODSEQ entries in the buffer, as
Emacs now has bignums and so won't misparse MODSEQs (Bug#38938).
* lisp/gnus/gnus-icalendar.el (gnus-icalendar-event--find-attendee):
(gnus-icalendar-identities) `gnus-ignored-from-addresses' and
`message-alternative-emails' may be functions. This is not
supported by `gnus-icalendar-event--find-attendee' (bug#43908).
* lisp/gnus/mml-sec.el (mml-secure-epg-sign): Partially revert
"Make mml-secure-epg-sign bug out if we can't find an identity".
It causes signing to fail for people who have not set up
mml-secure-{smime,openpgp}-sign-with-sender, which is a regression
from Emacs-26 (Bug#40118). In such a situation gpg will use its
default key.
Do not merge to master. On master Emacs will query the user.
* lisp/gnus/mml-sec.el (mml-secure-sender-sign-query): New
function (bug#40118).
(mml-secure-epg-sign): Use it to determine the signer (bug#40118).
* lisp/gnus/mml-sec.el
(mml-secure-allow-signing-with-unknown-recipient): Remove.
* lisp/gnus/gnus-icalendar.el
(gnus-icalendar-event:recurring-interval): Fix parsing of weekly
recurring events (bug#43669).
Example: in the absence of explicit INTERVAL value in the calendar
event, a weekly event with occurrences scheduled for Mondays and
Wednesdays should receive the default recurring interval of "1" and the
org mode timestamp repeater should be "+1w".
Due to a bug in the current code we receive "WEEKLY" and "+WEEKLYw"
instead. The patch fixes the issue.
* lisp/gnus/gnus-util.el (gnus-write-active-file): In case of group
names with spaces in them (see Bug#42823). Names are later read with
`read', so this should be quite robust.
* lisp/gnus/nnselect.el (numbers-by-group,
nnselect-request-update-info, nnselect-push-info): Handle all three
mark types ('tuple, 'range, 'list) and general speedups.
* lisp/mail/rfc2047.el (rfc2047-fold-field): Return the end point.
* lisp/gnus/message.el (message--fold-long-headers): Use that to
reliably achieve progress.
* lisp/gnus/gnus-agent.el (gnus-agent-store-article): Made
obsolete.
* lisp/gnus/gnus-art.el (gnus-request-article-this-buffer): Don't
call it.
* lisp/gnus/gnus-async.el (gnus-async-article-callback): Ditto
(bug#43356).
This partially reverts f3b146e943, which
in was response to what appears to be an erroneous feature-request -
bug#8502.
Instead this should be done instead:
“If I read an article while plugged, do they get entered into the
Agent?”
*No*. If you want this behavior, add
‘gnus-agent-fetch-selected-article’ to
‘gnus-select-article-hook’.
This reverts commit de4f347901.
read-file-name already displays a long path when given a default
(like INITIAL in many other prompting functions), so using
format-prompt here is superfluous.
* lisp/gnus/gnus.el (gnus-group-get-parameter): Allow the group
argument to be either a group name or a group info structure. This is
then parallel to gnus-group-set-parameter.
This is part of removing code from nnir.el that isn't related to
searching backends and therefore belongs somewhere else.
* lisp/gnus/gnus-group.el (gnus-group-make-search-group)
(gnus-group-read-ephemeral-search-group): Put the logic for
determining the groups to search here, rather than in nnir. Improve
documentation.
* lisp/gnus/gnus-int.el (gnus-server-get-active): Renamed from
'nnir-get-active.
* lisp/gnus/nnir.el (nnir-run-imap, nnir-run-find-grep): Use it.
(nnir-get-active): Remove.
(nnir-make-specs): Make obsolete.
* lisp/gnus/nnselect.el (nnselect-group-server): Make obsolete in
favor of 'gnus-group-server.
* lisp/gnus/message.el (message-newline-and-reformat): Pick up any
longer white-space prefix before starting to fill (bug#43299).
This fixes the problem of hitting M-RET on a line that's just ">".
* lisp/gnus/gnus-score.el (gnus-score-date): The previous < and >
"date" scoring rules (added in the previous patch) had reversed
logic (bug#43270).
Copyright-paperwork-exempt: yes
* lisp/gnus/gnus-sum.el (gnus-summary-edit-article-done): Strip ^M
from the ends of lines after saving an edited article. Otherwise the
new header isn't always parsed properly, resulting in an incorrect
subject line in the summary buffer.
