[PATCH 3/7] go: Allow notmuch objects to be garbage collected

Le 19/10/2012 06:55, Ethan Glasser-Camp a ?crit : > Adrien Bustany writes: > >> This makes notmuch appropriately free the underlying notmuch C objects >> when garbage collecting their Go wrappers. To make sure we don't break >> the underlying links be

[PATCH v2 0/2] Add flush/reopen methods to notmuch_database_t

Le 20/10/2012 18:49, Ethan Glasser-Camp a ?crit : > Jani Nikula writes: > >> On Wed, 17 Oct 2012, Adrien Bustany wrote: >>> The code of the patches in unchanged, but the formatting issues are now >>> hopefully fixed. >> >> Hi Adrien, please check

Re: [PATCH 3/7] go: Allow notmuch objects to be garbage collected

Le 19/10/2012 06:55, Ethan Glasser-Camp a écrit : Adrien Bustany writes: This makes notmuch appropriately free the underlying notmuch C objects when garbage collecting their Go wrappers. To make sure we don't break the underlying links between objects (for example, a notmuch_messages_t

Re: [PATCH v2 0/2] Add flush/reopen methods to notmuch_database_t

Le 20/10/2012 18:49, Ethan Glasser-Camp a écrit : Jani Nikula writes: On Wed, 17 Oct 2012, Adrien Bustany wrote: The code of the patches in unchanged, but the formatting issues are now hopefully fixed. Hi Adrien, please check at what version flush and reopen have been introduced to xapian

[PATCH 2/2] Add notmuch_database_reopen method

Calling notmuch_database_reopen is needed to refresh the database contents when the database on disk was modified by another notmuch_database_t instance, for example in a different thread. --- lib/database.cc | 17 + lib/notmuch.h | 8 2 files changed, 25 insertions(+)

[PATCH 1/2] Add notmuch_database_flush method

This method explicitly flushes the pending modifications to disk. It is useful if your program has various threads, each with a read only DB and one writer thread with a read/write DB. In that case, you most likely want the writer to sync the changes to disk so that the readers can see them, withou

[PATCH v2 0/2] Add flush/reopen methods to notmuch_database_t

The code of the patches in unchanged, but the formatting issues are now hopefully fixed.

[PATCH 1/2] Add notmuch_database_flush method

Le 17/10/2012 18:53, Ethan Glasser-Camp a ?crit : > Adrien Bustany writes: > >> This method explicitly flushes the pending modifications to disk. It is >> useful if your program has various threads, each with a read only DB and >> one writer thread with a read/write DB

[PATCH 2/2] Add notmuch_database_reopen method

Calling notmuch_database_reopen is needed to refresh the database contents when the database on disk was modified by another notmuch_database_t instance, for example in a different thread. --- lib/database.cc | 17 + lib/notmuch.h | 8 2 files changed, 25 insertions(+)

[PATCH 1/2] Add notmuch_database_flush method

This method explicitly flushes the pending modifications to disk. It is useful if your program has various threads, each with a read only DB and one writer thread with a read/write DB. In that case, you most likely want the writer to sync the changes to disk so that the readers can see them, withou

[PATCH v2 0/2] Add flush/reopen methods to notmuch_database_t

The code of the patches in unchanged, but the formatting issues are now hopefully fixed. ___ notmuch mailing list notmuch@notmuchmail.org http://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH 1/2] Add notmuch_database_flush method

Le 17/10/2012 18:53, Ethan Glasser-Camp a écrit : Adrien Bustany writes: This method explicitly flushes the pending modifications to disk. It is useful if your program has various threads, each with a read only DB and one writer thread with a read/write DB. In that case, you most likely want

[PATCH 3/7] go: Allow notmuch objects to be garbage collected

Le 20/07/2012 06:23, Austin Clements a ?crit : > Quoth Adrien Bustany on Jul 19 at 9:25 pm: >> Le 18/07/2012 23:40, Austin Clements a ?crit : >>> This is subtle enough that I think it deserves a comment in the source >>> code explaining that tracking the talloc owner r

Re: [PATCH 3/7] go: Allow notmuch objects to be garbage collected

Le 20/07/2012 06:23, Austin Clements a écrit : Quoth Adrien Bustany on Jul 19 at 9:25 pm: Le 18/07/2012 23:40, Austin Clements a écrit : This is subtle enough that I think it deserves a comment in the source code explaining that tracking the talloc owner reference, combined with the fact that

[PATCH 2/2] Add notmuch_database_reopen method

Calling notmuch_database_reopen is needed to refresh the database contents when the database on disk was modified by another notmuch_database_t instance, for example in a different thread. --- lib/database.cc | 17 + lib/notmuch.h |8 2 files changed, 25 insertions

[PATCH 1/2] Add notmuch_database_flush method

This method explicitly flushes the pending modifications to disk. It is useful if your program has various threads, each with a read only DB and one writer thread with a read/write DB. In that case, you most likely want the writer to sync the changes to disk so that the readers can see them, withou

[PATCH 0/2] Add flush/reopen methods to notmuch_database_t

ase_t for this purpose (the signaling of the writer to the readers that they need to reopen is left to the application). Adrien Bustany (2): Add notmuch_database_flush method Add notmuch_database_reopen method lib/database.cc | 35 +++ lib/notmuch.h |

[PATCH 0/7] Various fixes for the Go bindings

oks good to me other than the few things I commented on. > It's nice to see the Go bindings get a bit of love! > > Quoth Adrien Bustany on Jul 18 at 9:34 pm: >> The following patches fix some serious memory management issues with >> the Go bindings, and add some missi

[PATCH 3/7] go: Allow notmuch objects to be garbage collected

messages_t alive? - If we do talloc_reference(parent, wrapped), then we reproduce the hierarchy again? Note that I have 0 experience with talloc, so I might as well be getting things wrong here. > > Quoth Adrien Bustany on Jul 18 at 9:34 pm: >> This makes notmuch appropriately free the

[PATCH 2/2] Add notmuch_database_reopen method

Calling notmuch_database_reopen is needed to refresh the database contents when the database on disk was modified by another notmuch_database_t instance, for example in a different thread. --- lib/database.cc | 17 + lib/notmuch.h |8 2 files changed, 25 insertions

[PATCH 0/2] Add flush/reopen methods to notmuch_database_t

ase_t for this purpose (the signaling of the writer to the readers that they need to reopen is left to the application). Adrien Bustany (2): Add notmuch_database_flush method Add notmuch_database_reopen method lib/database.cc | 35 +++ lib/notmuch.h |

[PATCH 1/2] Add notmuch_database_flush method

This method explicitly flushes the pending modifications to disk. It is useful if your program has various threads, each with a read only DB and one writer thread with a read/write DB. In that case, you most likely want the writer to sync the changes to disk so that the readers can see them, withou

Re: [PATCH 0/7] Various fixes for the Go bindings

good to me other than the few things I commented on. It's nice to see the Go bindings get a bit of love! Quoth Adrien Bustany on Jul 18 at 9:34 pm: The following patches fix some serious memory management issues with the Go bindings, and add some missing functions as well. Adrien Bustany

Re: [PATCH 3/7] go: Allow notmuch objects to be garbage collected

), then we reproduce the hierarchy again? Note that I have 0 experience with talloc, so I might as well be getting things wrong here. Quoth Adrien Bustany on Jul 18 at 9:34 pm: This makes notmuch appropriately free the underlying notmuch C objects when garbage collecting their Go wrappers. To

[PATCH 7/7] go: Bind notmuch_thread_t functions

--- bindings/go/src/notmuch/notmuch.go | 253 +++- 1 files changed, 252 insertions(+), 1 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index be4cb8c..f667dbb 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b

[PATCH 6/7] go: Bind notmuch_database_find_message_by_filename

--- bindings/go/src/notmuch/notmuch.go | 39 1 files changed, 39 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index 384d5a5..be4cb8c 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b/

[PATCH 5/7] go: Partially bind notmuch_database_upgrade

This binding does not handle the progress callback, but at least allows opening and upgrading a database if needed. --- bindings/go/src/notmuch/notmuch.go | 13 - 1 files changed, 12 insertions(+), 1 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/not

[PATCH 4/7] go: Make Destroy functions safe to call several times

Those methods were already checking if the underlying C object was NULL, but they were not setting the pointer to NULL after destroying it. --- bindings/go/src/notmuch/notmuch.go |6 ++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bin

[PATCH 3/7] go: Allow notmuch objects to be garbage collected

This makes notmuch appropriately free the underlying notmuch C objects when garbage collecting their Go wrappers. To make sure we don't break the underlying links between objects (for example, a notmuch_messages_t being GC'ed before a notmuch_message_t belonging to it), we add for each wraper struc

[PATCH 2/7] go: Add missing MESSAGE_FLAG_EXCLUDED in Flag enum

--- bindings/go/src/notmuch/notmuch.go |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index ecd7418..1d77fd2 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b/bindings/go/src/notmuch/notmuch.go @

[PATCH 1/7] go: Use iota in enum bindings

Using iota is the correct way to get the values in the enum increment automatically. The old code would just set all the enum values to 0. --- bindings/go/src/notmuch/notmuch.go |6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bind

[PATCH 0/7] Various fixes for the Go bindings

The following patches fix some serious memory management issues with the Go bindings, and add some missing functions as well. Adrien Bustany (7): go: Use iota in enum bindings go: Add missing MESSAGE_FLAG_EXCLUDED in Flag enum go: Allow notmuch objects to be garbage collected go: Make

[PATCH 3/7] go: Allow notmuch objects to be garbage collected

This makes notmuch appropriately free the underlying notmuch C objects when garbage collecting their Go wrappers. To make sure we don't break the underlying links between objects (for example, a notmuch_messages_t being GC'ed before a notmuch_message_t belonging to it), we add for each wraper struc

[PATCH 7/7] go: Bind notmuch_thread_t functions

--- bindings/go/src/notmuch/notmuch.go | 253 +++- 1 files changed, 252 insertions(+), 1 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index be4cb8c..f667dbb 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b

[PATCH 0/7] Various fixes for the Go bindings

The following patches fix some serious memory management issues with the Go bindings, and add some missing functions as well. Adrien Bustany (7): go: Use iota in enum bindings go: Add missing MESSAGE_FLAG_EXCLUDED in Flag enum go: Allow notmuch objects to be garbage collected go: Make

[PATCH 2/7] go: Add missing MESSAGE_FLAG_EXCLUDED in Flag enum

--- bindings/go/src/notmuch/notmuch.go |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index ecd7418..1d77fd2 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b/bindings/go/src/notmuch/notmuch.go @

[PATCH 6/7] go: Bind notmuch_database_find_message_by_filename

--- bindings/go/src/notmuch/notmuch.go | 39 1 files changed, 39 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/notmuch/notmuch.go index 384d5a5..be4cb8c 100644 --- a/bindings/go/src/notmuch/notmuch.go +++ b/

[PATCH 1/7] go: Use iota in enum bindings

Using iota is the correct way to get the values in the enum increment automatically. The old code would just set all the enum values to 0. --- bindings/go/src/notmuch/notmuch.go |6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bind

[PATCH 4/7] go: Make Destroy functions safe to call several times

Those methods were already checking if the underlying C object was NULL, but they were not setting the pointer to NULL after destroying it. --- bindings/go/src/notmuch/notmuch.go |6 ++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bin

[PATCH 5/7] go: Partially bind notmuch_database_upgrade

This binding does not handle the progress callback, but at least allows opening and upgrading a database if needed. --- bindings/go/src/notmuch/notmuch.go | 13 - 1 files changed, 12 insertions(+), 1 deletions(-) diff --git a/bindings/go/src/notmuch/notmuch.go b/bindings/go/src/not

[PATCH] Makefile: specify libnotmuch.so location with -rpath

On Sun, 25 Apr 2010 16:38:40 +0100, Chris Wilson wrote: > In order to handle installation into user directories, it is convenient > to encode the library location into the search path for the notmuch > executable. This is achieved for the GNU linker with the -rpath > argument. > --- > Makefile.lo

[PATCH] Makefile: specify libnotmuch.so location with -rpath

On Sun, 25 Apr 2010 16:38:40 +0100, Chris Wilson wrote: > In order to handle installation into user directories, it is convenient > to encode the library location into the search path for the notmuch > executable. This is achieved for the GNU linker with the -rpath > argument. > --- > Makefile.lo

Re: [PATCH] Makefile: specify libnotmuch.so location with -rpath

On Sun, 25 Apr 2010 16:38:40 +0100, Chris Wilson wrote: > In order to handle installation into user directories, it is convenient > to encode the library location into the search path for the notmuch > executable. This is achieved for the GNU linker with the -rpath > argument. > --- > Makefile.lo

Re: [PATCH] Makefile: specify libnotmuch.so location with -rpath

On Sun, 25 Apr 2010 16:38:40 +0100, Chris Wilson wrote: > In order to handle installation into user directories, it is convenient > to encode the library location into the search path for the notmuch > executable. This is achieved for the GNU linker with the -rpath > argument. > --- > Makefile.lo

[PATCH] Makefile.local: Fix LDFLAGS for notmuch-shared

This commit adds GMIME_LDFLAGS and TALLOC_LDFLAGS to the linker flags when linking notmuch-shared. Without these flags, linking fails because of undefined symbols. --- Makefile.local |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/Makefile.local b/Makefile.local index 5bb

[PATCH] Makefile.local: Fix LDFLAGS for notmuch-shared

This commit adds GMIME_LDFLAGS and TALLOC_LDFLAGS to the linker flags when linking notmuch-shared. Without these flags, linking fails because of undefined symbols. --- Makefile.local |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/Makefile.local b/Makefile.local index 5bb

[PATCH] configure: Add support for GMime 2.6

Just sending the patch again for the 0.3 merge window... Cheers Adrien On Thu, 15 Apr 2010 19:41:55 -0400, Adrien Bustany wrote: > Notmuch compiles just fine with GMime 2.6, so accept it in the configure > script. > --- > configure |5 + > 1 files changed, 5 insertions(

Re: [PATCH] configure: Add support for GMime 2.6

Just sending the patch again for the 0.3 merge window... Cheers Adrien On Thu, 15 Apr 2010 19:41:55 -0400, Adrien Bustany wrote: > Notmuch compiles just fine with GMime 2.6, so accept it in the configure > script. > --- > configure |5 + > 1 files changed, 5 insertions(

[PATCH] configure: Add support for GMime 2.6

Notmuch compiles just fine with GMime 2.6, so accept it in the configure script. --- configure |5 + 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/configure b/configure index eebe075..d4d462f 100755 --- a/configure +++ b/configure @@ -188,6 +188,11 @@ if pkg-config --modve

[PATCH] configure: Add support for GMime 2.6

Notmuch compiles just fine with GMime 2.6, so accept it in the configure script. --- configure |5 + 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/configure b/configure index eebe075..d4d462f 100755 --- a/configure +++ b/configure @@ -188,6 +188,11 @@ if pkg-config --modve

[notmuch] vala, this is notmuch. notmuch, this is vala

On Mon, 5 Apr 2010 14:50:04 +0100, Enrico Zini wrote: > On Mon, Apr 05, 2010 at 02:49:23PM +0200, Sebastian Spaeth wrote: > >> So I bound notmuch.so to vala (at least what I needed) and played with >> the code a bit. The resulting 100 lines of vala code are here: > > Ooh, a .vapi for notmuch, th

Re: [notmuch] vala, this is notmuch. notmuch, this is vala

On Mon, 5 Apr 2010 14:50:04 +0100, Enrico Zini wrote: > On Mon, Apr 05, 2010 at 02:49:23PM +0200, Sebastian Spaeth wrote: > >> So I bound notmuch.so to vala (at least what I needed) and played with >> the code a bit. The resulting 100 lines of vala code are here: > > Ooh, a .vapi for notmuch, th

[notmuch] keeping a copy of sent mail locally

On Mon, 21 Dec 2009 09:27:04 -0800, Carl Worth wrote: > On Sun, 20 Dec 2009 15:29:15 +1100, Alex Ghitza wrote: >> It looks like we need a way to get the primary email address from the >> config file. > > Yes, we definitely need that. > >> Actually, while we're at it, we can consider making this

Re: [notmuch] keeping a copy of sent mail locally

On Mon, 21 Dec 2009 09:27:04 -0800, Carl Worth wrote: > On Sun, 20 Dec 2009 15:29:15 +1100, Alex Ghitza wrote: >> It looks like we need a way to get the primary email address from the >> config file. > > Yes, we definitely need that. > >> Actually, while we're at it, we can consider making this