Small bug in json-search changes

Small bug in json-search changes

[Mark Walters]

Hi

I have found a small bug in the recent changes to notmuch search to use
the JSON output. If you refresh the search buffer "point" does not stay
on the same thread. 

I think the problem is that notmuch-search-refresh-view calls notmuch
search with target-thread set to notmuch-search-find-thread-id which
returns the thread-id with "thread:" prefixed, whereas
notmuch-search-show-result checks for a match with (plist-get result
:thread) which is the thread id with no prefix.

The patch below fixes this but I am not sure it is the nicest fix.

Best wishes

Mark

diff --git a/emacs/notmuch.el b/emacs/notmuch.el
index d2d82a9..a53d5a0 100644
--- a/emacs/notmuch.el
+++ b/emacs/notmuch.el
@@ -798,7 +798,7 @@ non-authors is found, assume that all of the authors match."
 	(insert "\n")
 	(notmuch-search-color-line beg (point) (plist-get result :tags))
 	(put-text-property beg (point) 'notmuch-search-result result))
-      (when (string= (plist-get result :thread) notmuch-search-target-thread)
+      (when (string= (concat "thread:" (plist-get result :thread)) notmuch-search-target-thread)
 	(setq notmuch-search-target-thread "found")
 	(goto-char beg)))))
 

Re: Small bug in json-search changes

[Austin Clements]

Nice catch.  I think the solution is either exactly this, or we should
strip "thread:" off notmuch-search-target-thread when we set it.  I
would lean slightly toward the latter so we don't have to generate up
a new (identical) string on every notmuch-search-show-result call, but
I doubt it would make much difference in the grand scheme of things.

Quoth Mark Walters on Aug 07 at 12:10 am:
> 
> Hi
> 
> I have found a small bug in the recent changes to notmuch search to use
> the JSON output. If you refresh the search buffer "point" does not stay
> on the same thread. 
> 
> I think the problem is that notmuch-search-refresh-view calls notmuch
> search with target-thread set to notmuch-search-find-thread-id which
> returns the thread-id with "thread:" prefixed, whereas
> notmuch-search-show-result checks for a match with (plist-get result
> :thread) which is the thread id with no prefix.
> 
> The patch below fixes this but I am not sure it is the nicest fix.
> 
> Best wishes
> 
> Mark
> 
> diff --git a/emacs/notmuch.el b/emacs/notmuch.el
> index d2d82a9..a53d5a0 100644
> --- a/emacs/notmuch.el
> +++ b/emacs/notmuch.el
> @@ -798,7 +798,7 @@ non-authors is found, assume that all of the authors match."
>  	(insert "\n")
>  	(notmuch-search-color-line beg (point) (plist-get result :tags))
>  	(put-text-property beg (point) 'notmuch-search-result result))
> -      (when (string= (plist-get result :thread) notmuch-search-target-thread)
> +      (when (string= (concat "thread:" (plist-get result :thread)) notmuch-search-target-thread)
>  	(setq notmuch-search-target-thread "found")
>  	(goto-char beg)))))
>  
> 

[PATCH] emacs: notmuch search bugfix

[Mark Walters]

The recent change to use json for notmuch-search.el introduced a bug
in the code for keeping position on refresh. The problem is a
comparison between (plist-get result :thread) and a thread-id returned
by notmuch-search-find-thread-id: the latter is prefixed with
"thread:"

We fix this by adding an option to notmuch-search-find-thread-id to
return the bare thread-id. It appears that notmuch-search-refresh-view
is the only caller of notmuch-search that supplies a thread-id so this
change should be safe (but could theoretically break users .emacs
functions).
---

This is a different version of the bugfix suggested in the parent to
this thread. I think it is nicer but there is a slight risk of breaking
users .emacs lisp functions.

In general I think the thread: prefix should be present where the term
is "really" a search term, but absent when it is must be a thread-id,
and this fits with that idea.

With 0.14 freeze pretty close it might be best to apply this to master
and the simpler no breakage option (in id:"87boinpqfg.fsf@qmul.ac.uk")
to the branch.

Best wishes 

Mark

 emacs/notmuch.el |   10 ++++++----
 1 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/emacs/notmuch.el b/emacs/notmuch.el
index d2d82a9..0fff58a 100644
--- a/emacs/notmuch.el
+++ b/emacs/notmuch.el
@@ -475,10 +475,12 @@ BEG."
 	(push (plist-get (notmuch-search-get-result pos) property) output)))
     output))
 
-(defun notmuch-search-find-thread-id ()
-  "Return the thread for the current thread"
+(defun notmuch-search-find-thread-id (&optional bare)
+  "Return the thread for the current thread
+
+If BARE is set then do not prefix with \"thread:\""
   (let ((thread (plist-get (notmuch-search-get-result) :thread)))
-    (when thread (concat "thread:" thread))))
+    (when thread (concat (when (not bare) "thread:") thread))))
 
 (defun notmuch-search-find-thread-id-region (beg end)
   "Return a list of threads for the current region"
@@ -993,7 +995,7 @@ same relative position within the new buffer."
   (interactive)
   (let ((target-line (line-number-at-pos))
 	(oldest-first notmuch-search-oldest-first)
-	(target-thread (notmuch-search-find-thread-id))
+	(target-thread (notmuch-search-find-thread-id t))
 	(query notmuch-search-query-string)
 	(continuation notmuch-search-continuation))
     (notmuch-kill-this-buffer)
-- 
1.7.9.1

Re: [PATCH] emacs: notmuch search bugfix

[Austin Clements]

Quoth Mark Walters on Aug 07 at  4:40 pm:
> 
> The recent change to use json for notmuch-search.el introduced a bug
> in the code for keeping position on refresh. The problem is a
> comparison between (plist-get result :thread) and a thread-id returned
> by notmuch-search-find-thread-id: the latter is prefixed with
> "thread:"
> 
> We fix this by adding an option to notmuch-search-find-thread-id to
> return the bare thread-id. It appears that notmuch-search-refresh-view
> is the only caller of notmuch-search that supplies a thread-id so this
> change should be safe (but could theoretically break users .emacs
> functions).
> ---
> 
> This is a different version of the bugfix suggested in the parent to
> this thread. I think it is nicer but there is a slight risk of breaking
> users .emacs lisp functions.
> 
> In general I think the thread: prefix should be present where the term
> is "really" a search term, but absent when it is must be a thread-id,
> and this fits with that idea.

Agreed.

> With 0.14 freeze pretty close it might be best to apply this to master
> and the simpler no breakage option (in id:"87boinpqfg.fsf@qmul.ac.uk")
> to the branch.
> 
> Best wishes 
> 
> Mark
> 
>  emacs/notmuch.el |   10 ++++++----
>  1 files changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/emacs/notmuch.el b/emacs/notmuch.el
> index d2d82a9..0fff58a 100644
> --- a/emacs/notmuch.el
> +++ b/emacs/notmuch.el
> @@ -475,10 +475,12 @@ BEG."
>  	(push (plist-get (notmuch-search-get-result pos) property) output)))
>      output))
>  
> -(defun notmuch-search-find-thread-id ()
> -  "Return the thread for the current thread"
> +(defun notmuch-search-find-thread-id (&optional bare)
> +  "Return the thread for the current thread
> +
> +If BARE is set then do not prefix with \"thread:\""
>    (let ((thread (plist-get (notmuch-search-get-result) :thread)))
> -    (when thread (concat "thread:" thread))))
> +    (when thread (concat (when (not bare) "thread:") thread))))

(unless bare "thread:")?

>  
>  (defun notmuch-search-find-thread-id-region (beg end)
>    "Return a list of threads for the current region"
> @@ -993,7 +995,7 @@ same relative position within the new buffer."
>    (interactive)
>    (let ((target-line (line-number-at-pos))
>  	(oldest-first notmuch-search-oldest-first)
> -	(target-thread (notmuch-search-find-thread-id))
> +	(target-thread (notmuch-search-find-thread-id t))

Perhaps pass 'bare instead of t to make this self-documenting?

>  	(query notmuch-search-query-string)
>  	(continuation notmuch-search-continuation))
>      (notmuch-kill-this-buffer)

Re: [PATCH] emacs: notmuch search bugfix

[Tomi Ollila]

On Tue, Aug 07 2012, Mark Walters wrote:

> The recent change to use json for notmuch-search.el introduced a bug
> in the code for keeping position on refresh. The problem is a
> comparison between (plist-get result :thread) and a thread-id returned
> by notmuch-search-find-thread-id: the latter is prefixed with
> "thread:"
>
> We fix this by adding an option to notmuch-search-find-thread-id to
> return the bare thread-id. It appears that notmuch-search-refresh-view
> is the only caller of notmuch-search that supplies a thread-id so this
> change should be safe (but could theoretically break users .emacs
> functions).
> ---
>
> This is a different version of the bugfix suggested in the parent to
> this thread. I think it is nicer but there is a slight risk of breaking
> users .emacs lisp functions.
>
> In general I think the thread: prefix should be present where the term
> is "really" a search term, but absent when it is must be a thread-id,
> and this fits with that idea.
>
> With 0.14 freeze pretty close it might be best to apply this to master
> and the simpler no breakage option (in id:"87boinpqfg.fsf@qmul.ac.uk")
> to the branch.

LGTM.

IMHO you're too shy with this patch to be included in 0.14 :D
-- at least I cannot imagine a case this would fail...

My change would have been:
  (when thread (if bare thread (concat "thread:" thread)))

or even (which changes irrelevant part):
  (if thread (if bare thread (concat "thread:" thread)))

Anyway, this is good as it is "in the grand scheme of things".

Tomi

>
> Best wishes 
>
> Mark
>
>  emacs/notmuch.el |   10 ++++++----
>  1 files changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/emacs/notmuch.el b/emacs/notmuch.el
> index d2d82a9..0fff58a 100644
> --- a/emacs/notmuch.el
> +++ b/emacs/notmuch.el
> @@ -475,10 +475,12 @@ BEG."
>  	(push (plist-get (notmuch-search-get-result pos) property) output)))
>      output))
>  
> -(defun notmuch-search-find-thread-id ()
> -  "Return the thread for the current thread"
> +(defun notmuch-search-find-thread-id (&optional bare)
> +  "Return the thread for the current thread
> +
> +If BARE is set then do not prefix with \"thread:\""
>    (let ((thread (plist-get (notmuch-search-get-result) :thread)))
> -    (when thread (concat "thread:" thread))))
> +    (when thread (concat (when (not bare) "thread:") thread))))
>  
>  (defun notmuch-search-find-thread-id-region (beg end)
>    "Return a list of threads for the current region"
> @@ -993,7 +995,7 @@ same relative position within the new buffer."
>    (interactive)
>    (let ((target-line (line-number-at-pos))
>  	(oldest-first notmuch-search-oldest-first)
> -	(target-thread (notmuch-search-find-thread-id))
> +	(target-thread (notmuch-search-find-thread-id t))
>  	(query notmuch-search-query-string)
>  	(continuation notmuch-search-continuation))
>      (notmuch-kill-this-buffer)
> -- 
> 1.7.9.1
>
> _______________________________________________
> notmuch mailing list
> notmuch@notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch

[PATCH v2] emacs: notmuch search bugfix

[Mark Walters]

The recent change to use json for notmuch-search.el introduced a bug
in the code for keeping position on refresh. The problem is a
comparison between (plist-get result :thread) and a thread-id returned
by notmuch-search-find-thread-id: the latter is prefixed with
"thread:"

We fix this by adding an option to notmuch-search-find-thread-id to
return the bare thread-id. It appears that notmuch-search-refresh-view
is the only caller of notmuch-search that supplies a thread-id so this
change should be safe (but could theoretically break users .emacs
functions).
---

I think this implements all of Austin's suggestions, and it adds a short
remark to NEWS (since in unlikely cases it could break users' functions)
and updates the docstring for notmuch-search. 

Best wishes

Mark

 NEWS             |    3 +++
 emacs/notmuch.el |   12 +++++++-----
 2 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/NEWS b/NEWS
index 761b2c1..9916c91 100644
--- a/NEWS
+++ b/NEWS
@@ -55,6 +55,9 @@ user-specified formatting
   you've customized this variable, you may have to change your date
   format from `"%s "` to `"%12s "`.
 
+The thread-id for the `target-thread` argument for `notmuch-search` should
+now be supplied without the "thread:" prefix.
+
 Notmuch 0.13.2 (2012-06-02)
 ===========================
 
diff --git a/emacs/notmuch.el b/emacs/notmuch.el
index d2d82a9..7b61e9b 100644
--- a/emacs/notmuch.el
+++ b/emacs/notmuch.el
@@ -475,10 +475,12 @@ BEG."
 	(push (plist-get (notmuch-search-get-result pos) property) output)))
     output))
 
-(defun notmuch-search-find-thread-id ()
-  "Return the thread for the current thread"
+(defun notmuch-search-find-thread-id (&optional bare)
+  "Return the thread for the current thread
+
+If BARE is set then do not prefix with \"thread:\""
   (let ((thread (plist-get (notmuch-search-get-result) :thread)))
-    (when thread (concat "thread:" thread))))
+    (when thread (concat (unless bare "thread:") thread))))
 
 (defun notmuch-search-find-thread-id-region (beg end)
   "Return a list of threads for the current region"
@@ -936,7 +938,7 @@ If `query' is nil, it is read interactively from the minibuffer.
 Other optional parameters are used as follows:
 
   oldest-first: A Boolean controlling the sort order of returned threads
-  target-thread: A thread ID (with the thread: prefix) that will be made
+  target-thread: A thread ID (without the thread: prefix) that will be made
                  current if it appears in the search results.
   target-line: The line number to move to if the target thread does not
                appear in the search results."
@@ -993,7 +995,7 @@ same relative position within the new buffer."
   (interactive)
   (let ((target-line (line-number-at-pos))
 	(oldest-first notmuch-search-oldest-first)
-	(target-thread (notmuch-search-find-thread-id))
+	(target-thread (notmuch-search-find-thread-id 'bare))
 	(query notmuch-search-query-string)
 	(continuation notmuch-search-continuation))
     (notmuch-kill-this-buffer)
-- 
1.7.9.1

Re: [PATCH v2] emacs: notmuch search bugfix

[Austin Clements]

On Tue, 07 Aug 2012, Mark Walters <markwalters1009@gmail.com> wrote:
> The recent change to use json for notmuch-search.el introduced a bug
> in the code for keeping position on refresh. The problem is a
> comparison between (plist-get result :thread) and a thread-id returned
> by notmuch-search-find-thread-id: the latter is prefixed with
> "thread:"
>
> We fix this by adding an option to notmuch-search-find-thread-id to
> return the bare thread-id. It appears that notmuch-search-refresh-view
> is the only caller of notmuch-search that supplies a thread-id so this
> change should be safe (but could theoretically break users .emacs
> functions).
> ---
>
> I think this implements all of Austin's suggestions, and it adds a short
> remark to NEWS (since in unlikely cases it could break users' functions)
> and updates the docstring for notmuch-search. 

LGTM.

Re: [PATCH v2] emacs: notmuch search bugfix

[Tomi Ollila]

On Tue, Aug 07 2012, Mark Walters <markwalters1009@gmail.com> wrote:

> The recent change to use json for notmuch-search.el introduced a bug
> in the code for keeping position on refresh. The problem is a
> comparison between (plist-get result :thread) and a thread-id returned
> by notmuch-search-find-thread-id: the latter is prefixed with
> "thread:"
>
> We fix this by adding an option to notmuch-search-find-thread-id to
> return the bare thread-id. It appears that notmuch-search-refresh-view
> is the only caller of notmuch-search that supplies a thread-id so this
> change should be safe (but could theoretically break users .emacs
> functions).
> ---

LGTM.

Tomi

>
> I think this implements all of Austin's suggestions, and it adds a short
> remark to NEWS (since in unlikely cases it could break users' functions)
> and updates the docstring for notmuch-search. 
>
> Best wishes
>
> Mark
>
>  NEWS             |    3 +++
>  emacs/notmuch.el |   12 +++++++-----
>  2 files changed, 10 insertions(+), 5 deletions(-)
>
> diff --git a/NEWS b/NEWS
> index 761b2c1..9916c91 100644
> --- a/NEWS
> +++ b/NEWS
> @@ -55,6 +55,9 @@ user-specified formatting
>    you've customized this variable, you may have to change your date
>    format from `"%s "` to `"%12s "`.
>  
> +The thread-id for the `target-thread` argument for `notmuch-search` should
> +now be supplied without the "thread:" prefix.
> +
>  Notmuch 0.13.2 (2012-06-02)
>  ===========================
>  
> diff --git a/emacs/notmuch.el b/emacs/notmuch.el
> index d2d82a9..7b61e9b 100644
> --- a/emacs/notmuch.el
> +++ b/emacs/notmuch.el
> @@ -475,10 +475,12 @@ BEG."
>  	(push (plist-get (notmuch-search-get-result pos) property) output)))
>      output))
>  
> -(defun notmuch-search-find-thread-id ()
> -  "Return the thread for the current thread"
> +(defun notmuch-search-find-thread-id (&optional bare)
> +  "Return the thread for the current thread
> +
> +If BARE is set then do not prefix with \"thread:\""
>    (let ((thread (plist-get (notmuch-search-get-result) :thread)))
> -    (when thread (concat "thread:" thread))))
> +    (when thread (concat (unless bare "thread:") thread))))
>  
>  (defun notmuch-search-find-thread-id-region (beg end)
>    "Return a list of threads for the current region"
> @@ -936,7 +938,7 @@ If `query' is nil, it is read interactively from the minibuffer.
>  Other optional parameters are used as follows:
>  
>    oldest-first: A Boolean controlling the sort order of returned threads
> -  target-thread: A thread ID (with the thread: prefix) that will be made
> +  target-thread: A thread ID (without the thread: prefix) that will be made
>                   current if it appears in the search results.
>    target-line: The line number to move to if the target thread does not
>                 appear in the search results."
> @@ -993,7 +995,7 @@ same relative position within the new buffer."
>    (interactive)
>    (let ((target-line (line-number-at-pos))
>  	(oldest-first notmuch-search-oldest-first)
> -	(target-thread (notmuch-search-find-thread-id))
> +	(target-thread (notmuch-search-find-thread-id 'bare))
>  	(query notmuch-search-query-string)
>  	(continuation notmuch-search-continuation))
>      (notmuch-kill-this-buffer)
> -- 
> 1.7.9.1
>
> _______________________________________________
> notmuch mailing list
> notmuch@notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH v2] emacs: notmuch search bugfix

[David Bremner]

Mark Walters <markwalters1009@gmail.com> writes:

> The recent change to use json for notmuch-search.el introduced a bug
> in the code for keeping position on refresh. The problem is a
> comparison between (plist-get result :thread) and a thread-id returned

pushed, 

d