Re: [BUG] Custom headers in `notmuch-message-headers` are broken

2018-07-26 Thread Jan Malakhovski 
> 3) We should think carefully about whether we want to blindly send
>certain large headers like "Received". Some people use notmuch via
>ssh or equivalent, and it might (dunno) be a concern.

I'd prefer `notmuch show` to dump everything by default and have an
option like `--headers` to limit those. I.e. to get current behavior
you'd just dump comma-separated `notmuch-message-headers` into that
option in `notmuch.el` and be happy.

> 1) You'll want to avoid duplicating headers already emitted

Why? Wouldn't that prevent you from parsing "Received" headers in
`notmuch.el`?
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

[BUG] Custom headers in `notmuch-message-headers` are broken

2017-12-26 Thread Jan Malakhovski 
# What I did

I added "X-Github-Sender" to `notmuch-message-headers`.
Looked at a message sent via github with `notmuch-show`.

# What I expected

To see "X-Github-Sender" header displayed in `notmuch-show`.

# What I got

No such header was displayed.

# Why 

`(notmuch-show "query")` runs

```
notmuch show --format=sexp --format-version=4 query
```

internally. The latter produces a sexp with

```
:headers (:Subject "" :From "" :To "" :Reply-To "" :Date "")
```

even when the message has many more headers.

The end result is that `notmuch-message-headers` variable has no effect.
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: notmuch.el: controlling what does and doesn't get expanded in searches

2017-09-28 Thread Jan Malakhovski 
Hi.

I struggled with the same problem and wrote the following bit of code to
solve it. I've been using it for a couple of years now and I'm so happy
with it, that I forgot I wanted to get around to fix those TODOs. Feel
free do whatever with it. Hope it helps.

Cheers,
  Jan

;; A bit of notmuch-show-hook magic
;;
;; It makes notmuch-show
;; * show messages matching the specific query and hide the rest, or, if the 
query matches the whole thread,
;; * show unread or flagged messages of the thread and hide the rest, or, if 
there're no such messages,
;; * show all the messages
;; and jump to the very first one of those.
;;
;; This saves a lot of key presses.
;;
;; TODO: more generic one: by search query
;; TODO: rewrite using notmuch-show-mapc
(defun oxij/notmuch-show-only-tags (tags)
  "In notmuch-show-mode show messages matching tags, hide the rest"
  (interactive)
  (save-excursion
(goto-char (point-min))
(loop do
  (let ((props (notmuch-show-get-message-properties)))
   (if (remove-if-not (lambda (e) (member e tags))
  (plist-get props :tags))
   (notmuch-show-message-visible props t)
   (notmuch-show-message-visible props nil)))
until (not (notmuch-show-goto-message-next)

;; TODO: how do I break out of mapc?
(defun oxij/notmuch-show-interesting ()
  "Hide boring messages when all messages in the buffer match the specified 
query"
  (let ((all-matched t))
(notmuch-show-mapc (lambda ()
  (let ((props (notmuch-show-get-message-properties)))
(unless (plist-get props :match)
  (setq all-matched nil)
(when all-matched
  (oxij/notmuch-show-only-tags '("unread" "flagged")

;; notmuch-show-goto-first-wanted-message will be called just after
;; this hook and will undo the result of this when
;; oxij/notmuch-show-interesting hides everything. see the source of
;; notmuch-show-goto-first-wanted-message for details
(add-hook 'notmuch-show-hook 'oxij/notmuch-show-interesting)
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [BUG] gmime-3.0.1 (was: [PATCH] crypto: gracefully handle gmime errors)

2017-09-05 Thread Jan Malakhovski 
David Bremner  writes:

> There is code involved with configuring gmime-2.6, which nonetheless
> does not involve calling gpg itself.  As for notmuch-crypto.el, yes it's
> true that the emacs client sometimes does some stuff directly with gpg,
> but that has nothing to do with gmime.

That's what I wanted to hear. Thank you! (And another thanks for the
patience.)

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [BUG] gmime-3.0.1 (was: [PATCH] crypto: gracefully handle gmime errors)

2017-09-05 Thread Jan Malakhovski 
Servilio Afre Puentes <servi...@gmail.com> writes:

> On Tue, Sep 05 2017, Jan Malakhovski wrote:
>
> [...]
>> I added `gpgme` to `buildInputs` of `gmime` and now `notmuch` passes all
>> the tests. Yay! So that was the root problem.
>>
>> Is it correct to assume that when building with `gmime-3` `notmuch`
>> stops calling `gpg` binary and does all the things PGP using `gmime-3`?
>
> Yes, they use GPGME instead of a custom wrapper:
>
> https://mail.gnome.org/archives/gmime-devel-list/2017-April/msg2.html
>
> BTW, I created a merge-request to have GMime 3 in Nixpkgs:
>
> https://github.com/NixOS/nixpkgs/pull/28796

I continued this nix-specific discussion there.
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [BUG] gmime-3.0.1 (was: [PATCH] crypto: gracefully handle gmime errors)

2017-09-05 Thread Jan Malakhovski 
David Bremner  writes:

>> I added `gpgme` to `buildInputs` of `gmime` and now `notmuch` passes all
>> the tests. Yay! So that was the root problem.
>>
>> Is it correct to assume that when building with `gmime-3` `notmuch`
>> stops calling `gpg` binary and does all the things PGP using `gmime-3`?
>>
>
> notmuch does not call gpg directly in either case. What changes in the
> way gmime calls gpg. In 2.6 gmime called gpg, while in 3.0 it uses gpgme.
> It sounds like your gmime-3 package is missing a dependency on gpgme.

Why then at the very least

- crypto.c
- notmuch-config.c
- emacs/notmuch-crypto.el

files refer to "gpg" in the source with lines like

params.crypto.gpgpath = notmuch_config_get_crypto_gpg_path (config);
notmuch_config_set_crypto_gpg_path (config, "gpg");
gpgctx = g_mime_gpg_context_new (NULL, crypto->gpgpath ? crypto->gpgpath : 
"gpg")
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [BUG] gmime-3.0.1 (was: [PATCH] crypto: gracefully handle gmime errors)

2017-09-05 Thread Jan Malakhovski 
Hi.

David Bremner  writes:

> I'm fairly certain this something nix specific. 3.0.1 is the
> default version of gmime I develop against these days.
>
>> patching sources
>
> What patches, if any are applied here?

None.

>> T350-crypto: Testing PGP/MIME signature verification and decryption
>>  PASS   emacs delivery of signed message
>>  FAIL   signature verification
>>  --- T350-crypto.2.expected  2017-08-31 14:25:03.126885225 +
>>  +++ T350-crypto.2.output2017-08-31 14:25:03.126885225 +
>>  @@ -18,13 +18,7 @@
>>   ], 
>>   "content-type": "multipart/signed", 
>>   "id": 1, 
>>  -"sigstatus": [
>>  -{
>>  -"created": 946728000, 
>>  -"fingerprint": 
>> "5AEAB11F5E33DCE875DDB75B6D92612D94E46381", 
>>  -"status": "good"
>>  -}
>>  -]
>>  +"sigstatus": []
>>   }
>>   ], 
>>   "date_relative": "2000-01-01", 
>> Failed to verify signed part: Cannot verify multipart/signed part: 
>> unregistered signature protocol 'application/pgp-signature'.
>
> It seems like your gmime install doesn't understand PGP/MIME. That's
> pretty strange since afaik it enables SMIME and PGP/MIME with the same flag.
>
> Previously you wrote
> ,
> |I wonder why gnupg stops getting referenced with gmime-3.0.1. My guess
> |is that `./configure` does something very different when compiling with
> |gmime-3.
> `
>
> Although I don't think that configure is really the problem, the missing
> dependence on gnupg is suspicious.  Not having a gpg binary at all
> should cause more failures and/or messages about skipping. It's hard for
> me to test because on Debian there is a hard dependency of gmime-3.0 on
> gnupg.

I added `gpgme` to `buildInputs` of `gmime` and now `notmuch` passes all
the tests. Yay! So that was the root problem.

Is it correct to assume that when building with `gmime-3` `notmuch`
stops calling `gpg` binary and does all the things PGP using `gmime-3`?

It's the only explanation I have for why `notmuch` package stops
directly referencing `gpg` even when the sources get patched with
's/gpg/${pkgs.gpg}/bin/gpg/g' (not exactly, but close enough).

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH] cli: propagate NULL from _mime_node_create, handle it in callers.

2017-09-01 Thread Jan Malakhovski 
Hi.

David Bremner  writes:

> Can you check if this fixes the segfault for you? It's based on your
> patch, and your followup comment. It tries to handle all of the places
> mime_node_child is called.

It fixes segfaults the same way my patch does, but it also suffers from
the same problems.

It fixes the SIGSEGV, but, as with my patch, without some more code
around those NULLs `notmuch search --format=sexp --decrypt` simply
produces empty bodies (`:body ()`) for suspect messages. That breaks
rendering of those messages in `notmuch.el` (it fails with some
exception and stops rendering the rest of the thread; which is better
than rendering no messages in the thread, but still very far from
ideal). This solution also doesn't print any content when it could (in
case of an invalid signature).

Compiling with gmime-3.0.1, on the other hand, works perfectly. It just
shows that the signature is invalid and that's it, the rest of the
message and the rest of the thread renders perfectly. The problem with
gmime-3.0.1, as my other message still stuck in moderation shows, is
that a lot of other tests fail with some silly differences in the
produced JSONs.

Maybe its a bug in the tests. Maybe its a problem of `gmime` from
`nixpkgs` (apparently, it can't do pkcs7 and I didn't investigate why
yet). I'm not sure.

TBH, I would just `exit(1)` with an error like "your gmime is broken"
instead of producing NULLs whet gmime produces a NULL (to be
future-proof), assert non-NULL in all the other places, drop support for
gmime-2.6.*, require gmime-3.*, fix the tests (or `nixpkgs`) and get
happy.

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH] crypto: gracefully handle gmime errors

2017-08-31 Thread Jan Malakhovski 
David Bremner <da...@tethera.net> writes:

> Jan Malakhovski <o...@oxij.org> writes:
>
>>
>> The test suite has such a message already. "signature verification
>> (notmuch CLI)" test fails with a SIGSEGV when built with gmime-2.6.23.
>>
>> T355-smime: Testing S/MIME signature verification and decryption
>> gpgsm: keybox 
>> '/tmp/nix-build-notmuch-0.25.drv-0/notmuch-test-8789.qe2sRf/gnupg/pubring.kbx'
>>  created
>>  PASS   emacs delivery of S/MIME signed message
>>  PASS   emacs delivery of S/MIME encrypted + signed message
>>  PASS   Signature verification (openssl)
>>  FAIL   signature verification (notmuch CLI)
>
> Actually now that I try, I can't duplicate this failure on Debian (also
> with gmime-2.6.23). Matt, can you confirm or deny that the failures you
> were having were with SMIME signed messages?

Versions of all dependencies as per diff of `nix-store -q --requisites`
between two builds, if it helps:

--- /proc/self/fd/112017-08-31 23:53:53.387411873 +
+++ /proc/self/fd/122017-08-31 23:53:53.405412185 +
@@ -1,10 +1,9 @@
 /nix/store/y4ihw837rm0s0ps5f31zi18w7ri1a0y8-glibc-2.25
 /nix/store/bcg4yk2kinfd9703m3qlh7gkivkg1kqj-zlib-1.2.11
 /nix/store/2hd1rag058ija0ivn3ip4lfjdsnk1l9r-glib-2.52.1
-/nix/store/3ia3cz1hikmfhwgmbyf0j005myh9y7rb-gmime-2.6.23
+/nix/store/8ms9r6qw41rds9f59swiaaypmx07p376-gmime-3.0.1
 /nix/store/axldd4qjamb2c1lgi7d6k3s51qyy4610-bash-4.4-p12
 /nix/store/ds1qs36gzkzr4z6fvpdchwkhgkwq94p5-gcc-6.4.0-lib
 /nix/store/9636ksx6g2yifzx4d3wcfir87s3dzlja-xapian-1.4.4
-/nix/store/g3px2vjcfhaq7f6hdm1xnmbk5ry5qjrv-gnupg-2.1.23
 /nix/store/wfq4aip4ljnv0irx0r8b54pihfhz6im9-talloc-2.1.8
-/nix/store/h0lzwh8pmpcdzg7q8xl9bj5lzd9rvf71-notmuch-0.25
+/nix/store/kfwjlk0mqzisc77zxiflbamnrc52mfhq-notmuch-0.25

I wonder why gnupg stops getting referenced with gmime-3.0.1. My guess
is that `./configure` does something very different when compiling with
gmime-3.

Also, I think another message of mine got stuck in premoderation because
it had a biggish attachment with a build log.

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH] crypto: gracefully handle gmime errors

2017-08-31 Thread Jan Malakhovski 
David Bremner  writes:

> Oh, by the way, I think this was reported already in
> id:qf5d17wzq90@google.com, but we don't have a test case.

Yep. Seems to be the same bug. Added to References.

> So if you have (or can make) a message that triggers these segfaults,
> and that you can share, that would be helpful.

The test suite has such a message already. "signature verification
(notmuch CLI)" test fails with a SIGSEGV when built with gmime-2.6.23.

T355-smime: Testing S/MIME signature verification and decryption
gpgsm: keybox 
'/tmp/nix-build-notmuch-0.25.drv-0/notmuch-test-8789.qe2sRf/gnupg/pubring.kbx' 
created
 PASS   emacs delivery of S/MIME signed message
 PASS   emacs delivery of S/MIME encrypted + signed message
 PASS   Signature verification (openssl)
 FAIL   signature verification (notmuch CLI)
--- T355-smime.4.expected   2017-08-31 14:57:24.145052329 +
+++ T355-smime.4.output 2017-08-31 14:57:24.145052329 +
@@ -1,56 +1 @@
-[
-[
-[
-{
-"body": [
-{
-"content": [
-{
-"content": "This is a test signed 
message.\n",
-"content-type": "text/plain",
-"id": 2
-},
-{
-"content-disposition": "attachment",
-"content-length": "NONZERO",
-"content-transfer-encoding": "base64",
-"content-type": 
"application/x-pkcs7-signature",
-"filename": "smime.p7s",
-"id": 3
-}
-],
-"content-type": "multipart/signed",
-"id": 1,
-"sigstatus": [
-{
-"created": 946728000,
-"expires": 424242424,
-"fingerprint": 
"616F46CD73834C63847756AF0DFB64A6E0972A47",
-"status": "good"
-}
-]
-}
-],
-"date_relative": "2000-01-01",
-"excluded": false,
-"filename": [
-"Y"
-],
-"headers": {
-"Date": "Sat, 01 Jan 2000 12:00:00 +",
-"From": "Notmuch Test Suite 
",
-"Subject": "test signed message 001",
-"To": "test_su...@notmuchmail.org"
-},
-"id": "X",
-"match": true,
-"tags": [
-"inbox",
-"signed"
-],
-"timestamp": 946728000
-},
-[]
-]
-]
-]
+
Failed to construct pkcs7 context.
No JSON object could be decoded
 PASS   Decryption and signature verification (openssl)

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH] crypto: gracefully handle gmime errors

2017-08-31 Thread Jan Malakhovski 
Hi.

On second though

@@ -354,6 +357,8 @@ _mime_node_seek_dfs_walk (mime_node_t *node, int *n)
 *n -= 1;
 for (i = 0; i < node->nchildren; i++) {
mime_node_t *child = mime_node_child (node, i);
+   if (!child)
+   return NULL;
mime_node_t *ret = _mime_node_seek_dfs_walk (child, n);
if (ret)
return ret;

is probably an incorrect way to do this since the next child node can
succeed. Maybe it should be `continue` instead of `return NULL`. Not
sure.

Feel free to ignore this patch if there's a better way to fix those
SIGSEGVs. I get them by building against `gmime-2.6.23` and running
`tests/T355-smime.sh` and by trying to open some mails in my inbox
(which breaks display in emacs as `notmuch` just crashed in the middle
of an s-expression).

Cheers,
  Jan
___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

[PATCH] crypto: gracefully handle gmime errors

2017-08-31 Thread Jan Malakhovski 
Before: dies by SIGSEGV when gmime produces an error.
After: doesn't print NULL mime nodes.
---
 mime-node.c|  5 +
 notmuch-show.c | 18 +-
 2 files changed, 18 insertions(+), 5 deletions(-)

diff --git a/mime-node.c b/mime-node.c
index 24d73afa..f27aeb9a 100644
--- a/mime-node.c
+++ b/mime-node.c
@@ -317,7 +317,10 @@ mime_node_child (mime_node_t *parent, int child)
INTERNAL_ERROR ("Unexpected GMimeObject type: %s",
g_type_name (G_OBJECT_TYPE (parent->part)));
 }
+
 node = _mime_node_create (parent, sub);
+if (!node)
+   return NULL;
 
 if (child == parent->next_child && parent->next_part_num != -1) {
/* We're traversing in depth-first order.  Record the child's
@@ -354,6 +357,8 @@ _mime_node_seek_dfs_walk (mime_node_t *node, int *n)
 *n -= 1;
 for (i = 0; i < node->nchildren; i++) {
mime_node_t *child = mime_node_child (node, i);
+   if (!child)
+   return NULL;
mime_node_t *ret = _mime_node_seek_dfs_walk (child, n);
if (ret)
return ret;
diff --git a/notmuch-show.c b/notmuch-show.c
index cdcc2a98..2c591e37 100644
--- a/notmuch-show.c
+++ b/notmuch-show.c
@@ -567,8 +567,11 @@ format_part_text (const void *ctx, sprinter_t *sp, 
mime_node_t *node,
}
 }
 
-for (i = 0; i < node->nchildren; i++)
-   format_part_text (ctx, sp, mime_node_child (node, i), indent, params);
+for (i = 0; i < node->nchildren; i++) {
+   mime_node_t * child = mime_node_child (node, i);
+   if (child)
+   format_part_text (ctx, sp, child, indent, params);
+}
 
 if (GMIME_IS_MESSAGE (node->part))
g_mime_stream_printf (stream, "\fbody}\n");
@@ -619,7 +622,9 @@ format_part_sprinter (const void *ctx, sprinter_t *sp, 
mime_node_t *node,
if (output_body) {
sp->map_key (sp, "body");
sp->begin_list (sp);
-   format_part_sprinter (ctx, sp, mime_node_child (node, 0), TRUE, 
include_html);
+   mime_node_t * child = mime_node_child (node, 0);
+   if (child)
+   format_part_sprinter (ctx, sp, child, TRUE, include_html);
sp->end (sp);
}
sp->end (sp);
@@ -720,8 +725,11 @@ format_part_sprinter (const void *ctx, sprinter_t *sp, 
mime_node_t *node,
nclose = 3;
 }
 
-for (i = 0; i < node->nchildren; i++)
-   format_part_sprinter (ctx, sp, mime_node_child (node, i), TRUE, 
include_html);
+for (i = 0; i < node->nchildren; i++) {
+   mime_node_t * child = mime_node_child (node, i);
+   if (child)
+   format_part_sprinter (ctx, sp, child, TRUE, include_html);
+}
 
 /* Close content structures */
 for (i = 0; i < nclose; i++)
-- 
2.14.1

___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

[PATCH] build: fix unbound variable in configure

2017-08-31 Thread Jan Malakhovski 
---
 configure | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configure b/configure
index 364854f3..c0a550c7 100755
--- a/configure
+++ b/configure
@@ -651,6 +651,7 @@ if command -v cppcheck > /dev/null; then
 have_cppcheck=1
 printf "Yes.\n"
 else
+have_cppcheck=0
 printf "No.\n"
 fi
 
-- 
2.14.1

___
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

[PATCH] python: fix get_filenames() and make it actually usable

2015-07-31 Thread Jan Malakhovski 
The problem with the previous implementation is that different
versions of python exhaust __iter__() differently and the
implementation that can be exhausted is not only absolutely unusable
in the user code, but it also can not be consistently used with both
python 2.* and 3.*.

This doesn't change the interface.
---
 bindings/python/notmuch/filenames.py | 84 +++-
 1 file changed, 26 insertions(+), 58 deletions(-)

diff --git a/bindings/python/notmuch/filenames.py 
b/bindings/python/notmuch/filenames.py
index 229f414..e2b8886 100644
--- a/bindings/python/notmuch/filenames.py
+++ b/bindings/python/notmuch/filenames.py
@@ -65,6 +65,18 @@ class Filenames(Python3StringMixIn):
 _get.argtypes = [NotmuchFilenamesP]
 _get.restype = c_char_p

+_valid = nmlib.notmuch_filenames_valid
+_valid.argtypes = [NotmuchFilenamesP]
+_valid.restype = bool
+
+_move_to_next = nmlib.notmuch_filenames_move_to_next
+_move_to_next.argtypes = [NotmuchFilenamesP]
+_move_to_next.restype = None
+
+_destroy = nmlib.notmuch_filenames_destroy
+_destroy.argtypes = [NotmuchFilenamesP]
+_destroy.restype = None
+
 def __init__(self, files_p, parent):
 """
 :param files_p: A pointer to an underlying *notmuch_tags_t*
@@ -83,68 +95,24 @@ class Filenames(Python3StringMixIn):
 if not files_p:
 raise NullPointerError()

-self._files_p = files_p
+self._list = []
+while self._valid(files_p):
+file_ = self._get(files_p)
+self._move_to_next(files_p)
+self._list.append(file_.decode('utf-8', 'ignore'))
+self._destroy(files_p)
+
 #save reference to parent object so we keep it alive
 self._parent = parent

 def __iter__(self):
-""" Make Filenames an iterator """
-return self
-
-_valid = nmlib.notmuch_filenames_valid
-_valid.argtypes = [NotmuchFilenamesP]
-_valid.restype = bool
-
-_move_to_next = nmlib.notmuch_filenames_move_to_next
-_move_to_next.argtypes = [NotmuchFilenamesP]
-_move_to_next.restype = None
-
-def __next__(self):
-if not self._files_p:
-raise NotInitializedError()
-
-if not self._valid(self._files_p):
-self._files_p = None
-raise StopIteration
-
-file_ = Filenames._get(self._files_p)
-self._move_to_next(self._files_p)
-return file_.decode('utf-8', 'ignore')
-next = __next__ # python2.x iterator protocol compatibility
-
-def __unicode__(self):
-"""Represent Filenames() as newline-separated list of full paths
-
-.. note::
-
-This method exhausts the iterator object, so you will not be able 
to
-iterate over them again.
-"""
-return "\n".join(self)
-
-_destroy = nmlib.notmuch_filenames_destroy
-_destroy.argtypes = [NotmuchMessageP]
-_destroy.restype = None
-
-def __del__(self):
-"""Close and free the notmuch filenames"""
-if self._files_p:
-self._destroy(self._files_p)
+"""Make Filenames an iterator"""
+return self._list.__iter__()

 def __len__(self):
-"""len(:class:`Filenames`) returns the number of contained files
+"""len(:class:`Filenames`) returns the number of contained files"""
+return self._list.__len__()

-.. note::
-
-This method exhausts the iterator object, so you will not be able 
to
-iterate over them again.
-"""
-if not self._files_p:
-raise NotInitializedError()
-
-i = 0
-while self._valid(self._files_p):
-self._move_to_next(self._files_p)
-i += 1
-self._files_p = None
-return i
+def __unicode__(self):
+"""Represent Filenames() as newline-separated list of full paths"""
+return "\n".join(self)
-- 
2.4.1

[PATCH] python: fix get_filenames() and make it actually usable

2015-07-30 Thread Jan Malakhovski 
The problem with the previous implementation is that different
versions of python exhaust __iter__() differently and the
implementation that can be exhausted is not only absolutely unusable
in the user code, but it also can not be consistently used with both
python 2.* and 3.*.

This doesn't change the interface.
---
 bindings/python/notmuch/filenames.py | 84 +++-
 1 file changed, 26 insertions(+), 58 deletions(-)

diff --git a/bindings/python/notmuch/filenames.py 
b/bindings/python/notmuch/filenames.py
index 229f414..e2b8886 100644
--- a/bindings/python/notmuch/filenames.py
+++ b/bindings/python/notmuch/filenames.py
@@ -65,6 +65,18 @@ class Filenames(Python3StringMixIn):
 _get.argtypes = [NotmuchFilenamesP]
 _get.restype = c_char_p
 
+_valid = nmlib.notmuch_filenames_valid
+_valid.argtypes = [NotmuchFilenamesP]
+_valid.restype = bool
+
+_move_to_next = nmlib.notmuch_filenames_move_to_next
+_move_to_next.argtypes = [NotmuchFilenamesP]
+_move_to_next.restype = None
+
+_destroy = nmlib.notmuch_filenames_destroy
+_destroy.argtypes = [NotmuchFilenamesP]
+_destroy.restype = None
+
 def __init__(self, files_p, parent):
 
 :param files_p: A pointer to an underlying *notmuch_tags_t*
@@ -83,68 +95,24 @@ class Filenames(Python3StringMixIn):
 if not files_p:
 raise NullPointerError()
 
-self._files_p = files_p
+self._list = []
+while self._valid(files_p):
+file_ = self._get(files_p)
+self._move_to_next(files_p)
+self._list.append(file_.decode('utf-8', 'ignore'))
+self._destroy(files_p)
+
 #save reference to parent object so we keep it alive
 self._parent = parent
 
 def __iter__(self):
- Make Filenames an iterator 
-return self
-
-_valid = nmlib.notmuch_filenames_valid
-_valid.argtypes = [NotmuchFilenamesP]
-_valid.restype = bool
-
-_move_to_next = nmlib.notmuch_filenames_move_to_next
-_move_to_next.argtypes = [NotmuchFilenamesP]
-_move_to_next.restype = None
-
-def __next__(self):
-if not self._files_p:
-raise NotInitializedError()
-
-if not self._valid(self._files_p):
-self._files_p = None
-raise StopIteration
-
-file_ = Filenames._get(self._files_p)
-self._move_to_next(self._files_p)
-return file_.decode('utf-8', 'ignore')
-next = __next__ # python2.x iterator protocol compatibility
-
-def __unicode__(self):
-Represent Filenames() as newline-separated list of full paths
-
-.. note::
-
-This method exhausts the iterator object, so you will not be able 
to
-iterate over them again.
-
-return \n.join(self)
-
-_destroy = nmlib.notmuch_filenames_destroy
-_destroy.argtypes = [NotmuchMessageP]
-_destroy.restype = None
-
-def __del__(self):
-Close and free the notmuch filenames
-if self._files_p:
-self._destroy(self._files_p)
+Make Filenames an iterator
+return self._list.__iter__()
 
 def __len__(self):
-len(:class:`Filenames`) returns the number of contained files
+len(:class:`Filenames`) returns the number of contained files
+return self._list.__len__()
 
-.. note::
-
-This method exhausts the iterator object, so you will not be able 
to
-iterate over them again.
-
-if not self._files_p:
-raise NotInitializedError()
-
-i = 0
-while self._valid(self._files_p):
-self._move_to_next(self._files_p)
-i += 1
-self._files_p = None
-return i
+def __unicode__(self):
+Represent Filenames() as newline-separated list of full paths
+return \n.join(self)
-- 
2.4.1

___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch