David Bremner [Tue, 16 Jun 2020 14:17:55 +0000 (11:17 -0300)]
build: produce signatures for release candidate tarballs.
Today Defalos on #notmuch asked for a signed tarball for
0.30~rc2. This is a minimal change to support this in the future. The
question of automagically uploading will need more thought; currently
I like the fact that tags from pre-releases are only pushed manually.
configure: can gpgme can verify signatures when decrypting with a session key?
If https://dev.gnupg.org/T3464 is unresolved in the version of gpgme
we are testing against, then we should know about it, because it
affects the behavior of notmuch.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
David Bremner [Wed, 24 Jun 2020 14:32:34 +0000 (11:32 -0300)]
test: mark two tests broken on machines with 32 bit time_t
I haven't traced the code path as exhaustively for the SMIME test, but
the expiry date in question is larger then representable in a signed
32 bit integer.
David Bremner [Wed, 24 Jun 2020 14:32:33 +0000 (11:32 -0300)]
configure: detect 64 bit time_t
Certain tests involving timestamps > 32 bits cannot pass with the
current libnotmuch API. We will avoid this issue for now by disabling
those tests on "old" architectures with 32 bit time_t.
Tomi Ollila [Wed, 24 Jun 2020 20:11:20 +0000 (23:11 +0300)]
Makefile.global: drop -std=gnu99. C11 (or later) compiler required
Since October 2018 building notmuch has actually required compiler
that knows C11.
Also this -std=gnu99 was not used in code compiled by configure,
so in theory this could have caused problems...
...but no related reports have been sent, perhaps ever.
Both gcc and clang has been shipping compilers supporting C11
(or later) by default for more than four years now.
Therefore, just dropping -std=gnu99 (and not checking C11
compatibility for now, for simplicity) is easiest to do,
and removes inconsistency between configure and build time
compilations.
Since it is possible to use an atomic context to abort a number of
changes support this usage. Because the only way to actually abort
the transaction is to close the database this must also do so.
Amended by db: Note the limitation requiring close is a limitation of
the underlying notmuch API, which should be fixed in a future notmuch
release.
This reverses the logic of StandaloneMessage to instead create a
OwnedMessage. Only the Thread class allows retrieving messages more
then once so it can explicitly create such messages.
The added test fails with SIGABRT without the fix for the message
re-use in threads being present.
Anton Khirnov [Mon, 15 Jun 2020 20:58:49 +0000 (22:58 +0200)]
python/notmuch2: do not destroy messages owned by a query
Any messages retrieved from a query - either directly via
search_messages() or indirectly via thread objects - are owned by that
query. Retrieving the same message (i.e. corresponding to the same
message ID / database object) several times will always yield the same
C object.
The caller is allowed to destroy message objects owned by a query before
the query itself - which can save memory for long-lived queries.
However, that message must then never be retrieved again from that
query.
The python-notmuch2 bindings will currently destroy every message object
in Message._destroy(), which will lead to an invalid free if the same
message is then retrieved again. E.g. the following python program leads
to libtalloc abort()ing:
import notmuch2
db = notmuch2.Database(mode = notmuch2.Database.MODE.READ_ONLY)
t = next(db.threads('*'))
msgs = list(zip(t.toplevel(), t.toplevel()))
msgs = list(zip(t.toplevel(), t.toplevel()))
Fix this issue by creating a subclass of Message, which is used for
"standalone" message which have to be freed by the caller. Message class
is then used only for messages descended from a query, which do not need
to be freed by the caller.
Even though we use collections.abc.Set which implements all these
methods under their operator names, the actual named variations of
these methods are shockingly missing. So let's add them manually.
Tomi Ollila [Tue, 9 Jun 2020 12:32:27 +0000 (15:32 +0300)]
configure: use cffi.FFI().verify() to test buildability of CFFI bindings
Checking existence of pyconfig.h to determine whether CFFI-based
notmuch bindings are buildable is not enough; for example Fedora 32
ships pyconfig.h in python3-libs package, but python3-devel is required
to be installed for the bindings to build.
Executing cffi.FFI().verify() is pretty close to what is done in
bindings/python-cffi/notmuch2/_build.py to get the c code part of the
bindings built.
Tomi Ollila [Tue, 2 Jun 2020 20:21:59 +0000 (23:21 +0300)]
doc: field processor support now always included, adjust manual pages
The features that require field processor support, are now just
documented w/o mentioning **Xapian Field Processors**' is needed
for those.
Replaced "compact" and "field_processor" with "retry_lock" in
build_with config option, as it is currently the only one that
is optionally excluded. The former 2 are now documented as
features always included.
Dropped one 'we' "passive" in notmuch-search-terms.rst. It was the
only one, and inconsistent with rest of the documentation in that
file.
Dropped message about conditional open-ended ranges support, as
those are now always supported.
Tomi Ollila [Fri, 5 Jun 2020 21:57:55 +0000 (00:57 +0300)]
configure: require python 3.5 for CFFI-based notmuch bindings
Also tell users what the consequences of a "No" answer is when
python version is less than 3.5, cffi or setuptools is missing,
or no pytest >= 3.0 is available.
David Bremner [Wed, 3 Jun 2020 16:05:12 +0000 (13:05 -0300)]
devel: script to calculate a list of authors.
As an initial heuristic, report anyone with at least 15 lines of code
in the current source tree. Test corpora are excluded, although
probabably this doesn't change much about the list of authors
produced.
David Bremner [Sun, 31 May 2020 16:15:03 +0000 (13:15 -0300)]
doc: fix for out-of-tree builds of notmuch-emacs docs
The sphinx-doc include directive does not have the ability to include
files from the build tree, so we replace the include with reading the
files in conf.py. The non-trivial downside of this is that the emacs
docstrings are now defined for every rst source file. They are
namespaced with docstring::, so hopefully there will not be any
surprises. One thing that is noticable is a small (absolute) time
penalty in running sphinx-doc.
This is a simple hack to enable out-of-tree builds, a concern raised
by Tomi in id:m24kzjib9a.fsf@guru.guru-group.fi
This change at least enables "make check" to complete without error,
but I'm sure it could be improved. I am not expert enough in
setuptools to know how.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Amended by db per id:87d06usa31.fsf@powell.devork.be
Tomi Ollila [Thu, 23 Apr 2020 21:26:43 +0000 (00:26 +0300)]
test/test-lib.sh: fix two out of tree test issues
json_check_nodes.py exists in source tree, not in out of tree
build tree. Added -B to the execution so source tree is not
"polluted" by a .pyc file when json_check_nodes.py is executed.
When creating run_emacs.sh make it load .elc files from out of
tree build tree, not from source tree if such files existed.
If existed, those may be outdated, or even created by some other
emacs than the one that was used to build .elc files in out of
tree build dir.
smime: Pass PKCS#7 envelopedData to node_decrypt_and_verify
This change means we can support "notmuch show --decrypt=true" for
S/MIME encrypted messages, resolving several outstanding broken tests,
including all the remaining S/MIME protected header examples.
We do not yet handle indexing the cleartext of S/MIME encrypted
messages, though.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
crypto: handle PKCS#7 envelopedData in _notmuch_crypto_decrypt
In the two places where _notmuch_crypto_decrypt handles
multipart/encrypted messages (PGP/MIME), we should also handle PKCS#7
envelopedData (S/MIME).
This is insufficient for fully handling S/MIME encrypted data because
_notmuch_crypto_decrypt isn't yet actually invoked for envelopedData
parts, but that will happen in the following changes.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
crypto: Make _notmuch_crypto_decrypt take a GMimeObject
As we prepare to handle S/MIME-encrypted PKCS#7 EnvelopedData (which
is not multipart), we don't want to be limited to passing only
GMimeMultipartEncrypted MIME parts to _notmuch_crypto_decrypt.
There is no functional change here, just a matter of adjusting how we
pass arguments internally.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
cli/show: If a leaf part has children, show them instead of omitting
Until we did PKCS#7 unwrapping, no leaf MIME part could have a child.
Now, we treat the unwrapped MIME part as the child of the PKCS#7
SignedData object. So in that case, we want to show it instead of
deliberately omitting the content.
This fixes the test of the protected subject in
id:smime-onepart-signed@protected-headers.example.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
cli: include wrapped part of PKCS#7 SignedData in the MIME tree
Unwrap a PKCS#7 SignedData part unconditionally when the cli is
traversing the MIME tree, and return it as a "child" of what would
otherwise be a leaf in the tree.
Unfortunately, this also breaks the JSON output. We will fix that
next.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
When we are indexing, we should treat SignedData parts the same way
that we treat a multipart object, indexing the wrapped part as a
distinct MIME object.
Unfortunately, this means doing some sort of cryptographic
verification whose results we throw away, because GMime doesn't offer
us any way to unwrap without doing signature verification.
I've opened https://github.com/jstedfast/gmime/issues/67 to request
the capability from GMime but for now, we'll just accept the
additional performance hit.
As we do this indexing, we also apply the "signed" tag, by analogy
with how we handle multipart/signed messages. These days, that kind
of change should probably be done with a property instead, but that's
a different set of changes. This one is just for consistency.
Note that we are currently *only* handling signedData parts, which are
basically clearsigned messages. PKCS#7 parts can also be
envelopedData and authEnvelopedData (which are effectively encryption
layers), and compressedData (which afaict isn't implemented anywhere,
i've never encountered it). We're laying the groundwork for indexing
these other S/MIME types here, but we're only dealing with signedData
for now.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
smime: tests of X.509 certificate validity are known-broken on GMime < 3.2.7
When checking cryptographic signatures, Notmuch relies on GMime to
tell it whether the certificate that signs a message has a valid User
ID or not.
If the User ID is not valid, then notmuch does not report the signer's
User ID to the user. This means that the consumer of notmuch's
cryptographic summary of a message (or of its protected headers) can
be confident in relaying the reported identity to the user.
However, some versions of GMime before 3.2.7 cannot report Certificate
validity for X.509 certificates. This is resolved upstream in GMime
at https://github.com/jstedfast/gmime/pull/90.
We adapt to this by marking tests of reported User IDs for
S/MIME-signed messages as known-broken if GMime is older than 3.2.7
and has not been patched.
If GMime >= 3.2.7 and certificate validity still doesn't work for
X.509 certs, then there has likely been a regression in GMime and we
should fail early, during ./configure.
To break out these specific User ID checks from other checks, i had to
split some tests into two parts, and reuse $output across the two
subtests.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Several functions in test/test-lib.sh used variable names that are
also used outside of those functions (e.g. $output and $expected are
used in many of the test scripts), but they are not expected to
communicate via those variables.
We mark those variables "local" within test-lib.sh so that they do not
get clobbered when used outside test-lib.
We also move the local variable declarations to beginning of each
function, to avoid weird gotchas with local variable declarations as
described in https://tldp.org/LDP/abs/html/localvar.html.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
notmuch(1): clarify documentation about --option/value separators
id:CA+Tk8fzRiqxWpd=r8=DRvEewNZXUZgD7MKyRLB1A=R-LxxGEZw@mail.gmail.com
started a thread of discussion that showed that the cli's current
idiosyncrasies around dealing with boolean options were not
understandable.
This attempts to improve the documentation at least (actual changes to
the API might be better, but have not reached consensus).
Note that no one in the discussion thread identified any other
(non-boolean) command-line options that cannot use space as a
separator. If such an option is identified (or introduced in the
future), it should be added explicitly to this part of the manual.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
David Bremner [Mon, 4 May 2020 13:55:43 +0000 (10:55 -0300)]
lib: replace STRNCMP_LITERAL in __message_remove_indexed_terms
strncmp looks for a prefix that matches, which is very much not what
we want here. This fixes the bug reported by Franz Fellner in
id:1588595993-ner-8.651@TPL520
David Bremner [Mon, 4 May 2020 13:49:43 +0000 (10:49 -0300)]
test: known broken test for reindex tag preservation
In id:1588595993-ner-8.651@TPL520 Franz Fellner reported that tags
starting with 'attachment' are removed by 'notmuch reindex'. This is
probably related to the use of STRNCMP_LITERAL in
_notmuch_message_remove_indexed_terms.
GPGME has a strange failure mode when it is in offline mode, and/or
when certificates don't have any CRLs: in particular, it refuses to
accept the validity of any certificate other than a "root" cert.
This can be worked around by setting the `disable-crl-checks`
configuration variable for gpgsm.
I've reported this to the GPGME upstream at
https://dev.gnupg.org/T4883, but I have no idea how it will be
resolved. In the meantime, we'll just work around it.
Note that this fixes the test for verification of
id:smime-multipart-signed@protected-headers.example, because
multipart/signed messages are already handled correctly (one-part
PKCS#7 messages will get fixed later).
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Add a simple S/MIME SignedData message, taken from an upcoming draft
of
https://datatracker.ietf.org/doc/draft-autocrypt-lamps-protected-headers/
RFC 8551 describes a SignedData, a one-part clearsigned object that is
more resistant to common patterns of MTA message munging than
multipart/signed (but has the downside that it is only readable by
clients that implement S/MIME).
To make sure sure notmuch can handle this kind of object, we want to
know a few things:
Already working:
- Is the content of the SignedData object indexed? It actually is
right now because of dumb luck -- i think we're indexing the raw
CMS object and it happens to contain the cleartext of the message
in a way that we can consume it before passing it on to Xapian.
- Are we accidentally indexing the embedded PKCS#7 certificates? We
don't want to, and for some reason I don't understand, our indexing
is actually skipping the embedded certificates already. That's
good!
Still need fixing:
- do we know the MIME type of the embedded part?
- do we know that the message is signed?
- can notmuch-show read its content?
- can notmuch-show indicate the signature validity?
- can notmuch-reply properly quote and attribute content?
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
tests/smime: Test indexing cleartext of envelopedData
These tests describe some simple behavior we would expect to work if
we were to correctly index the cleartext of encrypted S/MIME messages
(PKCS#7 envelopedData).
Of course, they don't currently pass, so we mark them known-broken.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
When consuming a signed+encrypted S/MIME message generated by emacs,
we expect to see the same cryptographic properties for the message as
a whole. This is not done correctly yet, so the test is marked as
known broken.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Note that this commit doesn't strictly pass the common git pre-commit
hook due to introducing some trailing whitespace. That's just the
nature of the corpus, though. We should have that trailing
whitespace, so I've made this commit with --no-verify.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
This is taken from the same Internet Draft that test/smime/ca.crt
comes from. See that draft for more details.
https://www.ietf.org/id/draft-dkg-lamps-samples-02.html#name-pkcs12-object-for-bob
We don't use it yet, but it will be used to decrypt other messages in
the test suite.
Note that we include it here with an empty passphrase, rather than
with the passphrase "bob" that it is supplied with in the I-D. The
underlying cryptographic material is the same, but this way we can
import cleanly into gpgsm without having a passphrase set on it (gpgsm
converts an empty-string passphrase into no passphrase at all on
import).
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
tests/smime: avoid copying the key+cert.pem around
No functional change.
We no longer need to identify the key and cert to mml-mode when
sending an S/MIME message, so making a copy of key+cert.pem to
test_suite.pem is superfluous. Get rid of the extra file.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
To ensure that this mode works, we just need to import the secret key
in question into gpgsm in addition to the public key. gpgsm should be
able pick the right keys+certificates to use based on To/From headers,
so we don't have to specify anything manually in the #secure mml tag.
The import process from the OpenSSL-preferred form (cert+secretkey) is
rather ugly, because gpgsm wants to see a PKCS#12 object when
importing secret keys.
Note that EasyPG generates the more modern Content-Type:
application/pkcs7-signature instead of application/x-pkcs7-signature
for the detached signature.
We are also obliged to manually set gpgsm's include-certs setting to 1
because gpgsm defaults to send "everything but the root cert". In our
weird test case, the certificate we're using is self-signed, so it
*is* the root cert, which means that gpgsm doesn't include it by
default. Setting it to 1 forces inclusion of the signer's cert, which
satisfies openssl's smime subcommand. See https://dev.gnupg.org/T4878
for more details.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
David Bremner [Mon, 27 Apr 2020 12:24:22 +0000 (09:24 -0300)]
util/zlib-extra: de-inline gzerror_str
It turns out that putting inline functions in C header files is not a
good idea, and can cause linking problems if the compiler decides not
to inline them. In principle this is solvable by using a "static
inline" declaration, but this potentially makes a copy in every
compilation unit. Since we don't actually care about the performance
of this function, just use a non-inline function.
Jonas Bernoulli [Sat, 25 Apr 2020 20:18:07 +0000 (22:18 +0200)]
emacs: Use `cl-lib' instead of deprecated `cl'
Starting with Emacs 27 the old `cl' implementation is finally
considered obsolete. Previously its use was strongly discouraged
at run-time but one was still allowed to use it at compile-time.
For the most part the transition is very simple and boils down to
adding the "cl-" prefix to some symbols. A few replacements do not
follow that simple pattern; e.g. `first' is replaced with `car',
even though the alias `cl-first' exists, because the latter is not
idiomatic emacs-lisp.
In a few cases we start using `pcase-let' or `pcase-lambda' instead
of renaming e.g. `first' to `car'. That way we can remind the reader
of the meaning of the various parts of the data that is being
deconstructed.
An obsolete `lexical-let' and a `lexical-let*' are replaced with their
regular variants `let' and `let*' even though we do not at the same
time enable `lexical-binding' for that file. That is the right thing
to do because it does not actually make a difference in those cases
whether lexical bindings are used or not, and because this should be
enabled in a separate commit.
We need to explicitly depend on the `cl-lib' package because Emacs
24.1 and 24.2 lack that library. When using these releases we end
up using the backport from GNU Elpa.
We need to explicitly require the `pcase' library because
`pcase-dolist' was not autoloaded until Emacs 25.1.
test: sort the output of the "prefix" test in T610-message-property
This test extracts values from a (key,value) map where multiple entries
can have the same key, and the entries are sorted by key, but not by
value. The test incorrectly assumes that the values will be sorted as
well, so sort the output.
Tomi Ollila [Tue, 21 Apr 2020 21:07:29 +0000 (00:07 +0300)]
build: drop support for xapian versions less than 1.4
Xapian 1.4 is over 3 years old now (1.4.0 released 2016-06-24),
and 1.2 has been deprecated in Notmuch version 0.27 (2018-06-13).
Xapian 1.4 supports compaction, field processors and retry locking;
conditionals checking compaction and field processors were removed
but user may want to disable retry locking at configure time so it
is kept.
util: after gzgets(), Z_STREAM_END means EOF, not error
Context: I am compiling notmuch on OpenBSD which has a rather old zlib
1.2.3. It seems that the behaviour of gzgets() changed slightly between
this version and more recent versions, but the manual does not reflect
that change. Note that zlib's manual:
- does not specify which error code (Z_OK or Z_STREAM_END) is set when
EOF is reached,
- does not indicate the meaning of Z_STREAM_END after gzgets(), but
based on its meaning as a possible return value of inflate(), I would
guess that it means EOF.
emacs: use def instead of initial-input for notmuch-show-browse-urls
This is the non-deprecated way to use completing-read. Additionally
the old use was broken when using ivy for completing-read. For user's
using completing-read-default they won't see the default URL now, but
if they hit enter it will be visited. Alternatively they can select
it with M-n.
From the completing-read documentation for initial-input:
This feature is deprecated--it is best to pass nil for INITIAL-INPUT
and supply the default value DEF instead. The user can yank the
default value into the minibuffer easily using M-n.
Additionally collection is now all urls, rather than all but the
first. I'm not sure why "(cdr urls)" was previously done.
This is like notmuch-search-filter-by-tag, but creates a new search
rather than filtering the current search. We add this to
notmuch-common-keymap since this can be used by many contexts. We bind
to the key "t", which is the same key used by
notmuch-search-filter-by-tag in notmuch-search-mode-map. This is done
intentionally since the keybinding for notmuch-search-mode-map can be
seen as a specialization of creating a new search.
This change was motivated for use in "notmuch-hello". It is a more
convenient way to search a tag than expanding the list of all tags. I
also noticed many saved searches people use are simply tags.
David Bremner [Sun, 12 Apr 2020 17:30:12 +0000 (14:30 -0300)]
test: add known_broken test for dumping large stored queries
'qsx' reported a bug on #notmuch with notmuch-dump and large stored
queries. This test will pass (on my machine) if the value of `repeat'
is made smaller.
projects / notmuch / log