[PATCH] cli: add --output=address-{from,to,all} to notmuch search

2014-09-19 Thread David Bremner
Jani Nikula writes: > + if (name && *name) > + full_address = talloc_asprintf (NULL, "%s <%s>", name, addr); > + else > + full_address = talloc_asprintf (NULL, "<%s>", addr Is there some reason not to use sprinter as a talloc context here? > + > +

[PATCH 1/3] cli: add support for parsing multiple keyword arguments

2014-09-19 Thread David Bremner
Jani Nikula writes: > This allows having multiple --foo=bar --foo=baz options on the command > line, with the corresponding values OR'd together. > --- This looks OK. In hindsight I guess we could have had only keyword_flags from the beginning, but I guess that would be pretty intrusive to

[PATCH v4 3/3] Emacs: Add address completion based on company-mode

2014-09-19 Thread Michal Sojka
With this patch, address completion candidates are shown automatically after short typing delay in a nice popup box. This requires company-mode to be installed and it works only on Emacs >= 24. The completion is based entirely on the asynchronous address harvesting from notmuch-address.el so the

[PATCH v4 2/3] Emacs: Add address completion mechanism implemented in elisp

2014-09-19 Thread Michal Sojka
Currently, notmuch has an address completion mechanism that requires external command to provide completion candidates. This patch adds a completion mechanism inspired by https://github.com/tjim/nevermore, which is implemented in Emacs lisp only. The preexisting address completion mechanism,

[PATCH v4 1/3] Emacs: Display a message when generating address completion candidates

2014-09-19 Thread Michal Sojka
The TAB-initiated address completion generates completion candidates synchronously, blocking the UI. Since this can take long time, it is better to let the use know what's happening. --- emacs/notmuch-address.el | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git

[PATCH v4 0/3] Emacs: Address completion implemented in elisp

2014-09-19 Thread Michal Sojka
Hi, this is another version of my address completion patchset. It obsoletes id:1407771091-12651-1-git-send-email-sojkam1 at fel.cvut.cz. This version works reasonably well on systems with rotating disk, which was my main problem with the previous version. It also incorporates suggestions pointed

[PATCH v4 0/3] Emacs: Address completion implemented in elisp

2014-09-19 Thread Michal Sojka
Hi, this is another version of my address completion patchset. It obsoletes id:1407771091-12651-1-git-send-email-sojk...@fel.cvut.cz. This version works reasonably well on systems with rotating disk, which was my main problem with the previous version. It also incorporates suggestions pointed

[PATCH v4 1/3] Emacs: Display a message when generating address completion candidates

2014-09-19 Thread Michal Sojka
The TAB-initiated address completion generates completion candidates synchronously, blocking the UI. Since this can take long time, it is better to let the use know what's happening. --- emacs/notmuch-address.el | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git

[PATCH v4 3/3] Emacs: Add address completion based on company-mode

2014-09-19 Thread Michal Sojka
With this patch, address completion candidates are shown automatically after short typing delay in a nice popup box. This requires company-mode to be installed and it works only on Emacs = 24. The completion is based entirely on the asynchronous address harvesting from notmuch-address.el so the

[PATCH v4 2/3] Emacs: Add address completion mechanism implemented in elisp

2014-09-19 Thread Michal Sojka
Currently, notmuch has an address completion mechanism that requires external command to provide completion candidates. This patch adds a completion mechanism inspired by https://github.com/tjim/nevermore, which is implemented in Emacs lisp only. The preexisting address completion mechanism,

Re: [PATCH 1/3] cli: add support for parsing multiple keyword arguments

2014-09-19 Thread David Bremner
Jani Nikula j...@nikula.org writes: This allows having multiple --foo=bar --foo=baz options on the command line, with the corresponding values OR'd together. --- This looks OK. In hindsight I guess we could have had only keyword_flags from the beginning, but I guess that would be pretty

Re: [PATCH] cli: add --output=address-{from,to,all} to notmuch search

2014-09-19 Thread David Bremner
Jani Nikula j...@nikula.org writes: + if (name *name) + full_address = talloc_asprintf (NULL, %s %s, name, addr); + else + full_address = talloc_asprintf (NULL, %s, addr Is there some reason not to use sprinter as a talloc context here? + +

Re: [PATCH] python: Add binding for notmuch_query_add_tag_exclude

2014-09-19 Thread David Bremner
Thibaut Horel thibaut.ho...@gmail.com writes: Implemented as the method `exclude_tag` of the class `Query`. This method takes one argument, a string containing the name of the tag to exclude This tag looks pretty trivial. I'll merge it unless I hear some objections over the weekend. d

[PATCH] test/emacs: globally force the html renderer to html2text

2014-09-19 Thread David Bremner
Previously we did this for a single test, but some other proposed tests ( id:1398105468-14317-3-git-send-email-amdra...@mit.edu ) show similar breakage when switching renderers. --- test/T310-emacs.sh | 2 +- test/test-lib.el | 5 + 2 files changed, 6 insertions(+), 1 deletion(-) diff