From: Mark Walters Date: Fri, 25 Apr 2014 06:42:00 +0000 (+0100) Subject: Re: [PATCH 08/11] emacs: Support caching in notmuch-get-bodypart-{binary, text} X-Git-Url: http://git.tremily.us/?a=commitdiff_plain;h=faa1d4877b6ae6c34ca5f0771a01259ae97ede66;p=notmuch-archives.git Re: [PATCH 08/11] emacs: Support caching in notmuch-get-bodypart-{binary, text} --- diff --git a/cb/cd2a9d109c03f970c4227d9068c51ca891dc36 b/cb/cd2a9d109c03f970c4227d9068c51ca891dc36 new file mode 100644 index 000000000..0480a67af --- /dev/null +++ b/cb/cd2a9d109c03f970c4227d9068c51ca891dc36 @@ -0,0 +1,213 @@ +Return-Path: +X-Original-To: notmuch@notmuchmail.org +Delivered-To: notmuch@notmuchmail.org +Received: from localhost (localhost [127.0.0.1]) + by olra.theworths.org (Postfix) with ESMTP id 6789A431FBD + for ; Thu, 24 Apr 2014 23:42:10 -0700 (PDT) +X-Virus-Scanned: Debian amavisd-new at olra.theworths.org +X-Spam-Flag: NO +X-Spam-Score: 0.502 +X-Spam-Level: +X-Spam-Status: No, score=0.502 tagged_above=-999 required=5 + tests=[DKIM_ADSP_CUSTOM_MED=0.001, FREEMAIL_FROM=0.001, + NML_ADSP_CUSTOM_MED=1.2, RCVD_IN_DNSWL_LOW=-0.7] autolearn=disabled +Received: from olra.theworths.org ([127.0.0.1]) + by localhost (olra.theworths.org [127.0.0.1]) (amavisd-new, port 10024) + with ESMTP id 3rljN1cILgEi for ; + Thu, 24 Apr 2014 23:42:04 -0700 (PDT) +Received: from mail2.qmul.ac.uk (mail2.qmul.ac.uk [138.37.6.6]) + (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) + (No client certificate requested) + by olra.theworths.org (Postfix) with ESMTPS id 093AB431FAE + for ; Thu, 24 Apr 2014 23:42:04 -0700 (PDT) +Received: from smtp.qmul.ac.uk ([138.37.6.40]) + by mail2.qmul.ac.uk with esmtp (Exim 4.71) + (envelope-from ) + id 1WdZph-0003ze-Il; Fri, 25 Apr 2014 07:42:01 +0100 +Received: from 5751dfa2.skybroadband.com ([87.81.223.162] helo=localhost) + by smtp.qmul.ac.uk with esmtpsa (TLSv1:AES128-SHA:128) (Exim 4.71) + (envelope-from ) + id 1WdZph-0007Wc-1X; Fri, 25 Apr 2014 07:42:01 +0100 +From: Mark Walters +To: Austin Clements +Subject: Re: [PATCH 08/11] emacs: Support caching in + notmuch-get-bodypart-{binary, text} +In-Reply-To: <20140424181216.GO25817@mit.edu> +References: <1398105468-14317-1-git-send-email-amdragon@mit.edu> + <1398105468-14317-9-git-send-email-amdragon@mit.edu> + <87sip3t37f.fsf@qmul.ac.uk> <20140424181216.GO25817@mit.edu> +User-Agent: Notmuch/0.15.2+615~g78e3a93 (http://notmuchmail.org) Emacs/23.4.1 + (x86_64-pc-linux-gnu) +Date: Fri, 25 Apr 2014 07:42:00 +0100 +Message-ID: <87d2g5uczr.fsf@qmul.ac.uk> +MIME-Version: 1.0 +Content-Type: text/plain; charset=us-ascii +X-Sender-Host-Address: 87.81.223.162 +X-QM-Geographic: According to ripencc, + this message was delivered by a machine in Britain (UK) (GB). +X-QM-SPAM-Info: Sender has good ham record. :) +X-QM-Body-MD5: 14239cbd11f4e8c3a5792e1fe7b93f1a (of first 20000 bytes) +X-SpamAssassin-Score: -0.1 +X-SpamAssassin-SpamBar: / +X-SpamAssassin-Report: The QM spam filters have analysed this message to + determine if it is + spam. We require at least 5.0 points to mark a message as spam. + This message scored -0.1 points. + Summary of the scoring: + * 0.0 FREEMAIL_FROM Sender email is commonly abused enduser mail + provider * (markwalters1009[at]gmail.com) + * -0.1 AWL AWL: From: address is in the auto white-list +X-QM-Scan-Virus: ClamAV says the message is clean +Cc: notmuch@notmuchmail.org +X-BeenThere: notmuch@notmuchmail.org +X-Mailman-Version: 2.1.13 +Precedence: list +List-Id: "Use and development of the notmuch mail system." + +List-Unsubscribe: , + +List-Archive: +List-Post: +List-Help: +List-Subscribe: , + +X-List-Received-Date: Fri, 25 Apr 2014 06:42:10 -0000 + + +On Thu, 24 Apr 2014, Austin Clements wrote: +> Quoth Mark Walters on Apr 24 at 11:46 am: +>> +>> On Mon, 21 Apr 2014, Austin Clements wrote: +>> > (The actual code change here is small, but requires re-indenting +>> > existing code.) +>> > --- +>> > emacs/notmuch-lib.el | 52 ++++++++++++++++++++++++++++++---------------------- +>> > 1 file changed, 30 insertions(+), 22 deletions(-) +>> > +>> > diff --git a/emacs/notmuch-lib.el b/emacs/notmuch-lib.el +>> > index fc67b14..fee8512 100644 +>> > --- a/emacs/notmuch-lib.el +>> > +++ b/emacs/notmuch-lib.el +>> > @@ -503,33 +503,39 @@ (defun notmuch-parts-filter-by-type (parts type) +>> > (lambda (part) (notmuch-match-content-type (plist-get part :content-type) type)) +>> > parts)) +>> > +>> > -(defun notmuch-get-bodypart-binary (msg part process-crypto) +>> > +(defun notmuch-get-bodypart-binary (msg part process-crypto &optional cache) +>> > "Return the unprocessed content of PART in MSG as a unibyte string. +>> > +>> > This returns the \"raw\" content of the given part after content +>> > transfer decoding, but with no further processing (see the +>> > discussion of --format=raw in man notmuch-show). In particular, +>> > this does no charset conversion." +>> > - (let ((args `("show" "--format=raw" +>> > - ,(format "--part=%d" (plist-get part :id)) +>> > - ,@(when process-crypto '("--decrypt")) +>> > - ,(notmuch-id-to-query (plist-get msg :id))))) +>> > - (with-temp-buffer +>> > - ;; Emacs internally uses a UTF-8-like multibyte string +>> > - ;; representation by default (regardless of the coding system, +>> > - ;; which only affects how it goes from outside data to this +>> > - ;; internal representation). This *almost* never matters. +>> > - ;; Annoyingly, it does matter if we use this data in an image +>> > - ;; descriptor, since Emacs will use its internal data buffer +>> > - ;; directly and this multibyte representation corrupts binary +>> > - ;; image formats. Since the caller is asking for binary data, a +>> > - ;; unibyte string is a more appropriate representation anyway. +>> > - (set-buffer-multibyte nil) +>> > - (let ((coding-system-for-read 'no-conversion)) +>> > - (apply #'call-process notmuch-command nil '(t nil) nil args) +>> > - (buffer-string))))) +>> > - +>> > -(defun notmuch-get-bodypart-text (msg part process-crypto) +>> > + (let ((data (plist-get part :binary-content))) +>> > + (when (not data) +>> > + (let ((args `("show" "--format=raw" +>> > + ,(format "--part=%d" (plist-get part :id)) +>> > + ,@(when process-crypto '("--decrypt")) +>> > + ,(notmuch-id-to-query (plist-get msg :id))))) +>> > + (with-temp-buffer +>> > + ;; Emacs internally uses a UTF-8-like multibyte string +>> > + ;; representation by default (regardless of the coding +>> > + ;; system, which only affects how it goes from outside data +>> > + ;; to this internal representation). This *almost* never +>> > + ;; matters. Annoyingly, it does matter if we use this data +>> > + ;; in an image descriptor, since Emacs will use its internal +>> > + ;; data buffer directly and this multibyte representation +>> > + ;; corrupts binary image formats. Since the caller is +>> > + ;; asking for binary data, a unibyte string is a more +>> > + ;; appropriate representation anyway. +>> > + (set-buffer-multibyte nil) +>> > + (let ((coding-system-for-read 'no-conversion)) +>> > + (apply #'call-process notmuch-command nil '(t nil) nil args) +>> > + (setq data (buffer-string))))) +>> > + (when cache +>> > + (plist-put part :binary-content data))) +>> > + data)) +>> +>> I am a little puzzled by this but that could be lack of familiarity with +>> elisp. As far as I can see plist-put will sometimes modify the original +>> plist and sometimes return a new plist. If the latter happens then I +>> think it works out as if we hadn't cached anything as the part passed to +>> the function is unmodified. That might not matter in this case (though I +>> find the lack of determinism disturbing). +>> +>> Or is something else going on? +> +> No, your familiarity with Elisp serves you well. I'm completely +> cheating here. According to the specification of plist-put, it's +> allowed to return a new list but in reality this only happens when the +> original plist is nil. We lean on this already all over the +> notmuch-emacs code, but maybe that doesn't excuse me adding one more +> cheat. +> +> I could add a comment here explaining what's going on, I could +> manually do the list insertion in a way that's guaranteed to mutate it +> in place, or I could add a nil :binary-content property when parts are +> created (since plist-put is guaranteed to mutate existing keys in +> place). + +I think a comment is fine. + +(Incidentally what is the best way of telling if emacs has changed an +object or returned a new one for other commands? Something like (setq +oldobject object) (setq object (operation-on object)) (if (eq object +oldobject) ... )) + +Also, I think the function should have a comment about the lifetime of +the caching. I think in some cases the addition of :binary-content could +occur on load and thus the plist with binary content added would get +saved in the buffer when the msg plist was saved as a +text-property. However, maybe in other cases this gets called after the +initial insertion and thus the cached value is just used during this +operation on msg? Sorry that is a little incoherent as I haven't checked +all callers. + +Best wishes + +Mark + + + +>> Best wishes +>> +>> Mark +>> +>> +>> +>> > + +>> > +(defun notmuch-get-bodypart-text (msg part process-crypto &optional cache) +>> > "Return the text content of PART in MSG. +>> > +>> > This returns the content of the given part as a multibyte Lisp +>> > @@ -546,7 +552,9 @@ (defun notmuch-get-bodypart-text (msg part process-crypto) +>> > (npart (apply #'notmuch-call-notmuch-sexp args))) +>> > (setq content (plist-get npart :content)) +>> > (when (not content) +>> > - (error "Internal error: No :content from %S" args)))) +>> > + (error "Internal error: No :content from %S" args))) +>> > + (when cache +>> > + (plist-put part :content content))) +>> > content)) +>> > +>> > ;; Workaround: The call to `mm-display-part' below triggers a bug in