| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
We only want to decode the display part of the addresses only.
Mails with a From: like
From: "Non-ASCII Lastname, Firstname" <mail@example.org>
resulted in To: lines like
To: Non-ASCII Lastname, Firstname <mail@example.org>
which would send to two addresses. Use -A to ensure proper decoding
and quoting, even if its just a single address in From.
|
|
|
|
|
|
| |
RFC 2047.6.2:
NOTE: Decoding and display of encoded-words occurs *after* a
structured field body is parsed into tokens.
|
|
|
|
|
|
| |
RFC 2047.6.2:
NOTE: Decoding and display of encoded-words occurs *after* a
structured field body is parsed into tokens.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is maybe a foul compromise between correctness and complexity of
implementation, but it should do the right thing in most cases, and
does not require fully parsing and reconstructing all headers that can
contain phrases.
An 'encoded-word' MUST NOT appear within a 'quoted-string'. We thus
completely encode the quoted-string (if necessary) as a single
encoded-word, and strip off the quotes.
This should fix encoding of addresses that have both non-ASCII and
special chars such as , and ;.
|
|
|
|
|
|
|
|
| |
Characters such as , or ; mustn't appear in qp-encoded strings,
as they have a meaning in phrases. To be safe, encode all special
characters except for the safe ones in RFC 2047 5.(3).
_ is dealt with already.
|
|
|
|
| |
An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
|
| |
|
| |
|
| |
|
|
|
|
| |
Found by skarnet.
|
| |
|
|
|
|
| |
This is better recognizable as "no data".
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
https://www.loc.gov/preservation/digital/formats/fdd/fdd000383.shtml
> Each message is immediately prefaced by a separation line and
> terminated by an empty line.
Bug discovered by skarnet.
Fixes #207.
|
|
|
|
|
|
|
|
| |
https://www.loc.gov/preservation/digital/formats/fdd/fdd000383.shtml
> Each message is immediately prefaced by a separation line and
> terminated by an empty line.
Bug discovered by skarnet.
|
|
|
|
|
|
|
|
|
| |
Do not add additional Content-Type and Content-Transfer-Encoding headers
when using mmime on input already containing them.
Do not reencode the message if Content-Transfer-Encoding is set.
Based on a patch by Felix Van der Jeugt and duncaen.
|
|
|
|
| |
Closes #206.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
This is in conformance with RFC 1341.
|
|
|
|
| |
Closes #198.
|
|
|
|
| |
Closes: #199 [via git-merge-pr]
|
| |
|
|
|
|
|
|
| |
Avoids matching lines like '#!/usr/bin/env sh'.
Closes: #196 [via git-merge-pr]
|
| |
|
|
|
|
|
| |
Either the mail got truncated or was wrongly generated, try to go on
by taking the remains of the buffer as part.
|
|
|
|
| |
Closes: #193 [via git-merge-pr]
|
| |
|
|
|
|
| |
This triggered an invalid read in strchr on 'Content-Type: text/html; '
|
| |
|
|
|
|
| |
Closes #191.
|
| |
|
|
|
|
| |
Closes: #190 [via git-merge-pr]
|
|
|
|
|
|
|
|
| |
Closes #189.
This does not work with terminal emulators that don't use -e or
properly support arbitrary arguments after it. Here's a nickel,
get yourself a proper terminal emulator.
|
|
|
|
|
|
|
| |
On the first run or if you manually delete your sequence file, mseq
errors will break your prompt when you hit tab.
Closes: #188 [via git-merge-pr]
|
| |
|
|
|
|
| |
Closes: #187 [via git-merge-pr]
|
|
|
|
|
|
|
|
| |
Though POSIX requires this conversion to work correctly (for dlsym),
it is not valid in ISO C, so it is better to just uses the appropriate
function pointer type.
Closes: #185 [via git-merge-pr]
|
|
|
|
|
|
| |
These POSIX functions are declared in strings.h, so include this
header explicitly instead of relying on the libc's default feature-test
macros to include it through string.h.
|
| |
|
| |
|
|
|
|
| |
Found by lhynes.
|
|
|
|
|
| |
This allows mpick to create files and run popen.
In the future, we can only pledge this if these features are actually used.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
All programs except mshow have a very tight set of promises. mshow
has a broad set of promises and might be a good future candidate
to further restrict using unveil(2).
This patch is based on commit 0300a112 by Alex Holst (dated
2017-12-07), which was proposed in GH PR #79.
* pledged mpick, mflow and mdate so that now all programs are pledged
* removed some unneeded promises and added some missing promises
* move err.h include and OpenBSD ifdef into a new xpledge.h
* cleaned up code aligning and whitespace
Closes: #179 [via git-merge-pr]
|