2) When doing attachments set the "name=" parameter of the
Content-Type field, in addition to the (already) set "filename="
parameter of the Content-Disposition field. Some utilities (e.g.,
metamail) use this parameter for the filename even though (I believe)
the "filename=" parameter of the Content-Disposition field is
preferred by the standard (I can't find this explicitly except for
"application/octet-stream" types - see RFC1521 sec 7.4.1 and RFC2046
sec 4.5.1). My impression is that the "name=" parameter of the
Content-Types field is really intended for use when retrieving a file
not in the message, e.g., "message/external-body" Content-Types, and
not for the filename.
(Thanks to wiz@ for noticing this in his spam logs.)
3) Be more careful when determining the Content-Type of 1-byte
attachments. libmagic(3) isn't helpful on such small files.
1) Add support for message selection based on the message body. The
pattern matching is done on the MIME decoded body as would be seen by
the print command.
2) Don't hook editline when doing headers only: that mode is never
interactive and it messes up piping if output is redirected to a
command that expects tty input, such as 'more'.
1) When detaching, don't try to close things we shouldn't (after
doing the header). Specifically, mip->mi_head_end was not
getting set correctly in mime_sendmessage().
2) Change paging to be closer to its old behavior: next, dt, and dp
should only page if the crt variable is set; and don't automatically
page most other commands - the user can always pipe them into more.
Partially restore the "crt" variable: if set, the [pP]rint and [tT]ype
commands invoke the PAGER making them identical to the [pP]age and
[mM]more commands. Its value is now ignored. If anyone really
objects, I will do my best to restore the old behavior, but it really
doesn't fit very well into the current paging architecture.
1) Statification of modules.
2) Implement the 'detach' and 'Detach' commands for extracting mime
parts from messages.
3) Teach mail to output "In-Reply-To" and "References" header fields
when replying so others can thread us.
4) Implement threading, sorting, and tagging, supported by the
following commands: 'flatten', 'reverse', 'sort', 'thread',
'unthread', 'down', 'tset', 'up', 'expose', 'hide', 'tag',
'untag', 'invtags', 'tagbelow', 'hidetags', 'showtags'.
See the manpage for details (when available - soon).
5) Implement a 'deldups' command to delete duplicate messages based on
their "Message-Id" field, e.g., in replies to a mailing list that
are also CCed to a subscriber. (This can also be accomplished with
the threading and tagging commands.)
6) Implement 'ifdef' and 'ifndef' commands, and make the conditionals
nestable (i.e., implement a conditional stack). The if/else/endif
commands existed before, but they were primitive and undocumented.
The 'if' command currently recognizes the "receiving", "sending",
and "headersonly" mode keywords.
7) Teach the message selecting routine to understand regular
expressions if "regex-search" is defined. Otherwise only case
insensitive substring matches are done (as in the past).
8) Teach the message selection routine to understand boolean
expressions. Improved "colon-modifier" support. See the manpage
for details (when available - soon).
9) Extend paging to all commands (where relevant).
10) Add shell like piping and redirection of (standard) output (if
"enable-piping" is defined). Extend completion to these contexts.
11) The manpage should follow soon!!!!
The problem is that mime_decode_close() closes all files registered
after mip->mi_pipe_end and this is getting set in mime_sendmessage()
after a fflush().
When you print a large number of messages fflush() blocks and prevents
mip->mi_pipe_end from getting set before the SIGPIPE comes in and
jumps to the close block that calls mime_decode_close(). As a result,
mime_decode_close() was closing all the registered files including
obuf (mi_pipe_end was NULL).
1) Removed the -B flag (it was stupid on my part) and added a short
description indicating how to accomplish the same thing under the
"Sending Mail" section of man mail(1).
2) Added a -H flag to dump the headers and exit. It takes optional
flags to restrict to old, new, read, unread, and deleted messages
(the later being kind of useless - it shares code with something
that already had it).
3) Restored the 'Save' command which somehow got mistakenly removed in
the last commit and add documentation for it! (My apologies to
its author.)
4) Added a 'mkread' command to mark messages as read (the inverse of
'unread'). Should we also have a 'mknew' command?
5) Added a 'smopts' command to keep a database of addresses and
sendmail options to be used when sending messages to those
addresses. See man mail(1) for a fuller description.
6) Added 'indentpreamble' and 'indentpostscript' variables whose
values are inserted before and after a quoted message (~m or ~M
escapes).
=20
7) Added string formatting abilities for the 'prompt', 'insertpreamble',
'insertpostscript', and header display strings. These strings
support all the strftime() format parameters as well as many more
specific to mail (see man mail(1)).
8) Fix the -a flag so that it only takes a single filename, unless
"mime-attach-list" is defined. This is more conventional and avoids
unexpected whitespace issues.
1) Use editline [optional]:
Most of this code was borrowed from src/usr.bin/ftp. It does the
appropriate editing, history, and completion for all mail commands
(from cmdtab[]) and also does editing on header strings ('~h' inside
the mail editor).
1) Use editline [optional]:
Most of this code was borrowed from src/usr.bin/ftp. It does the
appropriate editing, history, and completion for all mail commands
(from cmdtab[]) and also does editing on header strings ('~h' inside
the mail editor).
2) '-B' flag:
This will suppress the "To:" line passed to sendmail. In most
configurations it will lead to sendmail adding "To: undisclosed
recipients;". Currently, AFAIK mail requires at least one exposed
recipient address.
3) Comments in rcfile:
Currently, comments in .mailrc are only supported if the first
(non-white) character on a line is '#' followed by white space,
i.e., '#' is a 'nop' command. This (trivial) patch allows the more
normal/expected use of '#' as a comment character. It does not
respect quoting, so that might be an objection which I should fix.
4) Sendmail option editing:
This adds the sendmail option string to the strings editable by the
'~h' command within the mail editor. Currently, you can only set
this string from the command-line, which is particularly annoying
when replying to mail.
5) Reply from:
When replying to a message, grab the "To:" address from the message
and, if there is only one such address and it does not match a list of
allowed addresses (set in the "ReplyFrom" variable), pass it to
sendmail as the "From:" address for the reply (with the '-f' option).
I often make aliases for myself so that my primary address is not
given out; if the alias gets out, I know who to blame. Unfortunately,
a reply to such a message would normally use the primary address
without this patch. A warning is displayed when this is going to
happen so that it can be modified with '~h'.
6) CC and BCC lists:
Allow '-c' and '-b' to accept white-space or ',' delimited lists.
Currently, a white-space delimited list of addresses work, but a
list of aliases will not get expanded. For example, currently:
mail -c "foo bar" christos
will fail to send mail to 'foo' and 'bar' if these are mail aliases
(in ~/.mailrc); sendmail aliases (in /etc/aliases) do work.
7) pipe command:
This pipes the current message into a shell command. I use this for
quick decoding of uuencoded mail, but I can imagine it might be
useful for decrypting encrypted mail, too.
8) show command:
This command takes a list of variables and shows their values. It
is probably stupid as the 'set' command without any argument
displays all variable values. Of course, if there are a lot of
variables you have to sift through the list for the one(s) you want.
target) instead of using home-grown 'distribution' targets or using
FILES with the 'install' target.
Add some etc/ subdir Makefiles where appropriate.
XXX: some of etc/Makefile install-etc-files could be converted to CONFIGFILES.
program/tool from "FOO" to "TOOL_FOO". The new variables are:
TOOL_ASN1_COMPILE TOOL_CAP_MKDB TOOL_CAT TOOL_CKSUM TOOL_COMPILE_ET
TOOL_CONFIG TOOL_CRUNCHGEN TOOL_CTAGS TOOL_DB TOOL_EQN TOOL_FGEN
TOOL_GENCAT TOOL_GROFF TOOL_HEXDUMP TOOL_INDXBIB TOOL_INSTALLBOOT
TOOL_INSTALL_INFO TOOL_M4 TOOL_MAKEFS TOOL_MAKEINFO TOOL_MAKEWHATIS
TOOL_MDSETIMAGE TOOL_MENUC TOOL_MKCSMAPPER TOOL_MKESDB
TOOL_MKLOCALE TOOL_MKMAGIC TOOL_MKTEMP TOOL_MSGC TOOL_MTREE
TOOL_PAX TOOL_PIC TOOL_PREPMKBOOTIMAGE TOOL_PWD_MKDB TOOL_REFER
TOOL_ROFF_ASCII TOOL_ROFF_DVI TOOL_ROFF_HTML TOOL_ROFF_PS
TOOL_ROFF_RAW TOOL_RPCGEN TOOL_SOELIM TOOL_SUNLABEL TOOL_TBL
TOOL_UUDECODE TOOL_VGRIND TOOL_ZIC
For each, provide default in <bsd.sys.mk> of the form:
TOOL_FOO?= foo
and for the ${USETOOLS}=="yes" case in <bsd.own.mk>, provide override:
TOOL_FOO= ${TOOLDIR}/bin/${_TOOL_PREFIX}foo
Document all of these in bsd.README.
This cleans up a chunk of potential (and actual) namespace collision
within our build infrastructure, as well as improves consistency in
the share/mk documentation and provision of appropriate defaults for
each of these variables.
He accidently moved the call to block the SIGCHLD to the wrong side of
the call to findchild().
Caused a coredump in one in every N thousand invocations of mail,
which I have been hunting for literally years.
Fixes PR 19696 from Mason Loring Bliss, too.
By the way, this is a prime example of why declarations of the form
struct foo *p = funcall();
are pure liquid evil.
when deciding whether to run $PAGER, otherwise it may start up the
pager for a two line message if all 55 header lines are the subject
of a .mailrc ignore command.
(And no, I don't find this program directly useful for reading
today's mail volumes, but it's great as a component run from wrapper
scripts, pretty good for scanning archived mail, and more than
adequate for sending mail.)
