[PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Tomi Ollila]

On Sun, Nov 04 2012, Jani Nikula wrote:

> On Sun, 04 Nov 2012, Blake Jones  wrote:
>> Linux, FreeBSD, and Solaris all expect to find the prototype for
>> "index()" in .  On some operating systems, including
>>  is sufficient to get the prototype, but that's not the case
>> on Solaris.  This patch just modifies notmuch-config.c to include
>>  to get the prototype.
>
> We should probably just nuke index() and use strchr() instead.

indeed!

>
> Jani.
>

Tomi


>
> diff --git a/notmuch-config.c b/notmuch-config.c
> index 3e37a2d..47eb743 100644
> --- a/notmuch-config.c
> +++ b/notmuch-config.c
> @@ -688,7 +688,7 @@ _item_split (char *item, char **group, char **key)
>  
>  *group = item;
>  
> -period = index (item, '.');
> +period = strchr (item, '.');
>  if (period == NULL || *(period+1) == '\0') {
>   fprintf (stderr,
>"Invalid configuration name: %s\n"
> ___
> notmuch mailing list
> notmuch at notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch

[PATCH 00/10] Solaris support

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones  wrote:
> Hi all,
>
> This patch series fixes several issues which are needed to allow notmuch
> to build on Solaris 11.  I've been "testing" it for a month or so by
> using Karel Zak's fork of mutt along with a copy of notmuch-0.13.2 that
> I got to compile.  After a friend asked for a copy of my setup, I
> decided to try to get my patches cleaned up enough to submit, so that he
> wouldn't need to deal with them if he wanted to hack on it.

Thanks for doing this. I had a *very* brief look at the patches, and
commented on what I thought stood out. Otherwise I didn't spot anything
out of the ordinary, but I'd like it if someone else had a glance too.

BR,
Jani.

[PATCH 05/10] install: check for non-SysV version (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones  wrote:
> Solaris ships a program called "install" in /usr/sbin, which performs a
> task that's fairly similar to the GNU and BSD "install" programs but
> which uses very different command line arguments.  In particular, if it
> is invoked without "-c", "-f", or "-n", it will search the target
> directory for a file with the same name as the one being installed, and
> it will only install the file if it finds a matching name.  More
> excitingly, if it doesn't find a match, it will look in /bin, /usr/bin,
> /etc, /lib, and /usr/lib and try to do the same there.
>
> The standard workaround for this is to use GNU install.
> It is available via the standard Solaris packaging system (in
> "file/gnu-coreutils"), and installs itself as /usr/bin/ginstall.
>
> This patch adds a check to "configure" to see if "install" behaves in a
> way that's compatible with GNU and BSD install, and if not, it uses a
> program called "ginstall" instead.  It also modifies "configure" to set
> the $(INSTALL) variable, and changes various Makefiles to use it.
> ---
>  Makefile.local|2 +-
>  completion/Makefile.local |4 ++--
>  configure |   19 +++
>  emacs/Makefile.local  |6 +++---
>  lib/Makefile.local|4 ++--
>  man/Makefile.local|6 +++---
>  vim/Makefile  |6 ++
>  7 files changed, 32 insertions(+), 15 deletions(-)
>
> diff --git a/Makefile.local b/Makefile.local
> index 2b91946..7ccb1cd 100644
> --- a/Makefile.local
> +++ b/Makefile.local
> @@ -286,7 +286,7 @@ notmuch-shared: $(notmuch_client_modules) 
> lib/$(LINKER_NAME)
>  .PHONY: install
>  install: all install-man
>   mkdir -p "$(DESTDIR)$(prefix)/bin/"
> - install notmuch-shared "$(DESTDIR)$(prefix)/bin/notmuch"
> + $(INSTALL) notmuch-shared "$(DESTDIR)$(prefix)/bin/notmuch"
>  ifeq ($(MAKECMDGOALS), install)
>   @echo ""
>   @echo "Notmuch is now installed to $(DESTDIR)$(prefix)"
> diff --git a/completion/Makefile.local b/completion/Makefile.local
> index dfc1271..a648a78 100644
> --- a/completion/Makefile.local
> +++ b/completion/Makefile.local
> @@ -14,9 +14,9 @@ install-$(dir):
>   @echo $@
>  ifeq ($(WITH_BASH),1)
>   mkdir -p "$(DESTDIR)$(bash_completion_dir)"
> - install -m0644 $(bash_script) "$(DESTDIR)$(bash_completion_dir)/notmuch"
> + $(INSTALL) -m0644 $(bash_script) 
> "$(DESTDIR)$(bash_completion_dir)/notmuch"
>  endif
>  ifeq ($(WITH_ZSH),1)
>   mkdir -p "$(DESTDIR)$(zsh_completion_dir)"
> - install -m0644 $(zsh_script) "$(DESTDIR)$(zsh_completion_dir)/_notmuch"
> + $(INSTALL) -m0644 $(zsh_script) 
> "$(DESTDIR)$(zsh_completion_dir)/_notmuch"
>  endif
> diff --git a/configure b/configure
> index 5c5139f..dae837e 100755
> --- a/configure
> +++ b/configure
> @@ -591,6 +591,21 @@ for flag in -Wmissing-declarations; do
>  done
>  printf "\n\t${WARN_CFLAGS}\n"
>  
> +INSTALL="install"
> +printf "Checking for working \"install\" program... "
> +mkdir _tmp_

This doesn't feel like a hot idea. Don't tell me you'd need to create a
compatibility script for using mktemp --tmpdir too...

Or how about just always using ginstall on Solaris?

BR,
Jani.

> +cd _tmp_
> +echo 1 > 1
> +mkdir dest
> +if install 1 dest > /dev/null 2>&1 ; then
> +  printf "\"install\" works fine.\n"
> +else
> +  INSTALL="ginstall"
> +  printf "using \"ginstall\".\n"
> +fi
> +cd ..
> +rm -rf _tmp_
> +
>  rm -f minimal minimal.c
>  
>  cat < @@ -777,4 +792,8 @@ CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) 
> \$(GMIME_CFLAGS)\\
>-DSTD_ASCTIME=\$(STD_ASCTIME)
>  CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS) 
> \\
>\$(LIBNSL_LDFLAGS)
> +
> +# Which "install" program to use
> +INSTALL = ${INSTALL}
> +
>  EOF
> diff --git a/emacs/Makefile.local b/emacs/Makefile.local
> index fb82247..ee778cb 100644
> --- a/emacs/Makefile.local
> +++ b/emacs/Makefile.local
> @@ -36,11 +36,11 @@ endif
>  .PHONY: install-emacs
>  install-emacs:
>   mkdir -p "$(DESTDIR)$(emacslispdir)"
> - install -m0644 $(emacs_sources) "$(DESTDIR)$(emacslispdir)"
> + $(INSTALL) -m0644 $(emacs_sources) "$(DESTDIR)$(emacslispdir)"
>  ifeq ($(HAVE_EMACS),1)
> - install -m0644 $(emacs_bytecode) "$(DESTDIR)$(emacslispdir)"
> + $(INSTALL) -m0644 $(emacs_bytecode) "$(DESTDIR)$(emacslispdir)"
>  endif
>   mkdir -p "$(DESTDIR)$(emacsetcdir)"
> - install -m0644 $(emacs_images) "$(DESTDIR)$(emacsetcdir)"
> + $(INSTALL) -m0644 $(emacs_images) "$(DESTDIR)$(emacsetcdir)"
>  
>  CLEAN := $(CLEAN) $(emacs_bytecode)
> diff --git a/lib/Makefile.local b/lib/Makefile.local
> index 7785944..0c6b258 100644
> --- a/lib/Makefile.local
> +++ b/lib/Makefile.local
> @@ -89,11 +89,11 @@ install: install-$(dir)
>  
>  install-$(dir): $(dir)/$(LIBNAME)
>   mkdir -p "$(DESTDIR)$(libdir)/"
> - install -m0644 "$(lib)/$(LIBNAME)" "$(DESTDIR)$(libdir)/"
> + 

Automatic suppression of non-duplicate messages

[Austin Clements]

Quoth Jani Nikula on Nov 05 at 12:34 am:
> On Sat, 03 Nov 2012, David Bremner  wrote:
> > Eirik Byrkjeflot Anonsen  writes:
> >
> >> That's not what I see.  If I search for a term that only appears in
> >> one of the "copies", none of the copies are included in the search
> >> result.
> >
> > The offending code is at line 1813 of lib/database.cc; the message is
> > only indexed if the message-id is new.
> >
> > It might be sensible to move _notmuch_message_index_file into the other
> > branch of the if, but even if that works fine, something more
> > sophisticated is needed for the call to
> > __notmuch_message_set_header_values; the invariant that each message has
> > a single subject seems reasonable.
> >
> > Offhand I'm not sure of a good method of automatically deciding what is
> > the same message (with e.g. headers and footer text added by a mailing
> > list).
> 
> Assuming there was good method, what would you do with two different
> messages that have the same message id? That is the unique id we use to
> identify messages (which should be fine per RFC 5322 and its
> predecessors; we're talking about messages from broken systems here).
> 
> It might be helpful to have a configuration option similar to new.tags
> that would define the tags to be assigned to messages with duplicate
> message ids. (This could be done in the
> NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID case near line 516 of
> notmuch-new.c). This could be used to assign a "dupe" tag, for example,
> so the user could do whatever they want in the post-new hook or the user
> interface. A sufficiently clever post-new hook could compare the files
> of a message, and drop the tag or add another, as the case may
> be. Surely not a perfect solution, but keeps the implementation simple.

This would also trigger on message flag changes and folder moves
performed outside of notmuch, since notmuch sees those as a duplicate
message ID followed by a deletion.  The only way to do something for
every received message even if it has the same message ID as an
existing message is to do it in whatever delivers mail.  Currently, we
don't have a good story for integrating on-delivery operations with
notmuch.

[PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones  wrote:
> Linux, FreeBSD, and Solaris all expect to find the prototype for
> "index()" in .  On some operating systems, including
>  is sufficient to get the prototype, but that's not the case
> on Solaris.  This patch just modifies notmuch-config.c to include
>  to get the prototype.

We should probably just nuke index() and use strchr() instead.

Jani.


diff --git a/notmuch-config.c b/notmuch-config.c
index 3e37a2d..47eb743 100644
--- a/notmuch-config.c
+++ b/notmuch-config.c
@@ -688,7 +688,7 @@ _item_split (char *item, char **group, char **key)

 *group = item;

-period = index (item, '.');
+period = strchr (item, '.');
 if (period == NULL || *(period+1) == '\0') {
fprintf (stderr,
 "Invalid configuration name: %s\n"

[PATCH 10/10] timegm: add portable implementation (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones  wrote:
> Hi Jani,
>
>> I'd prefer to use timegm() where available, and the suggested
>> alternative [1] elsewhere.
>> 
>> [1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html
>
> I considered this alternative, but decided against it because it's
> completely MT-unsafe.  I don't know whether libnotmuch itself is
> MT-safe, but a process which called this routine in one thread would
> temporarily throw off any timezone-related work that any other threads
> were doing, even if they weren't using libnotmuch.

That is a valid point. Yet it doesn't change the fact that I'd prefer to
use timegm() where available. Internally, glibc uses the same code to
implement both timegm() and mktime(), and I'd hate it if the results
were subtly different depending on whether the time zone was specified
in the input or not. That said, I'm not opposed to using your simple
timegm() alternative in the compat code if you think it's good enough to
get you going on Solaris.

As to solving the compat linking problem, I think the patch at the end
of this message should fix it. Please try that with the regular notmuch
approach to portability. The general idea is to keep parse-time-string
as independent as possible from the rest of notmuch (possibly turning it
into a dynamic library and a package of its own eventually), but I think
including compat.h is an acceptable exception to make.

HTH,
Jani.


diff --git a/parse-time-string/Makefile.local b/parse-time-string/Makefile.local
index 53534f3..b3e5385 100644
--- a/parse-time-string/Makefile.local
+++ b/parse-time-string/Makefile.local
@@ -1,7 +1,9 @@
 dir := parse-time-string
 extra_cflags += -I$(srcdir)/$(dir)

-libparse-time-string_c_srcs := $(dir)/parse-time-string.c
+libparse-time-string_c_srcs =  \
+   $(notmuch_compat_srcs)  \
+   $(dir)/parse-time-string.c

 libparse-time-string_modules := $(libparse-time-string_c_srcs:.c=.o)

diff --git a/parse-time-string/parse-time-string.c 
b/parse-time-string/parse-time-string.c
index 584067d3..ccad422 100644
--- a/parse-time-string/parse-time-string.c
+++ b/parse-time-string/parse-time-string.c
@@ -32,6 +32,7 @@
 #include 
 #include 

+#include "compat.h"
 #include "parse-time-string.h"

 /*

[PATCH 05/10] install: check for non-SysV version (Solaris support)

[Blake Jones]

> > +INSTALL="install"
> > +printf "Checking for working \"install\" program... "
> > +mkdir _tmp_
> 
> This doesn't feel like a hot idea.

Out of curiosity, why not?  An "install" that behaves as expected is
one of the first things that an autoconf-generated "configure" looks
for.  Now, autoconf-configure implements that check using some
assumptions about where things are on different operating systems,
but that sort of check runs the risk of becoming stale (see below).

> Don't tell me you'd need to create a compatibility script for using
> mktemp --tmpdir too...

Yes I would, if it were used; not all the world's a GNU.  But in the
case of mktemp, the widely available "-p" and "-t" options seem to get
you most of the way there.  The SVR4 "install" in Solaris' /usr/sbin is
different enough from the BSD/GNU versions that I wouldn't want to try
to emulate it with a wrapper.

> Or how about just always using ginstall on Solaris?

I'd rather not do that.  With the old UCB tools having been EOL'ed [1],
/usr/ucb/install (which would have worked) will be going away.  There is
an open bug in the Sun bug tracking system about moving GNU install to
/usr/bin/install, specifically motivated by this change.  So while I
don't know if/when that bug will be fixed, I would guess [2] that a
future release of Solaris may have a BSD/GNU-compatible version of
"install" in the default $PATH.

Blake

[1] 
http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html
[2] Caveat: I work for Oracle in the Solaris kernel group, but I am not
speaking for my employer.

[PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Blake Jones]

>> On Sun, 04 Nov 2012, Blake Jones  wrote:
>>> Linux, FreeBSD, and Solaris all expect to find the prototype for
>>> "index()" in .  On some operating systems, including
>>>  is sufficient to get the prototype, but that's not the case
>>> on Solaris.  This patch just modifies notmuch-config.c to include
>>>  to get the prototype.
>>
>> We should probably just nuke index() and use strchr() instead.
> 
> indeed!

That was my initial preference, but I didn't know if there was anyone
who was committed to the BSD name.  Given that two more people think
it's a good idea, I'll do that instead.

Blake

[PATCH 10/10] timegm: add portable implementation (Solaris support)

[Blake Jones]

> That is a valid point. Yet it doesn't change the fact that I'd prefer
> to use timegm() where available. Internally, glibc uses the same code
> to implement both timegm() and mktime(), and I'd hate it if the
> results were subtly different depending on whether the time zone was
> specified in the input or not.

That's fine with me.

> That said, I'm not opposed to using your simple timegm() alternative
> in the compat code if you think it's good enough to get you going on
> Solaris.

I think it is, assuming you don't plan to use tm_wday or tm_yday in your
parse-time-string code, and that you don't plan to depend on the side
effect of timegm() canonicalizing the passed-in "struct tm".

> As to solving the compat linking problem, I think the patch at the end
> of this message should fix it. Please try that with the regular
> notmuch approach to portability. The general idea is to keep
> parse-time-string as independent as possible from the rest of notmuch
> (possibly turning it into a dynamic library and a package of its own
> eventually), but I think including compat.h is an acceptable exception
> to make.

Yeah, this seems to work.  I'll update my patch set accordingly.

Blake

[PATCH 10/10] timegm: add portable implementation (Solaris support)

[Jani Nikula]

On Nov 4, 2012 11:30 AM, "Blake Jones"  wrote:
>
> The timegm(3) function is a non-standard extension to libc which is
> available in GNU libc and on some BSDs.  Although SunOS had this
> function in its libc, Solaris (unfortunately) removed it.  This patch
> implements a very simple version of timegm() which is good enough for
> parse-time-string.c.
>
> Although notmuch's idiom for portability is to test for native
> availability and put alternate versions in compat/, that approach led to
> a compilation problem in this case.  libnotmuch.a includes a call to
> parse_time_string() from parse-time-vrp.o, and parse_time_string() in
> libparse-time-string.a needs to call timegm().  An attempt to create
> compat/timegm.c caused the link to fail, because libparse-time-string.a
> acquired a dependency on the new timegm.o in libnotmuch.a, and the
> linker only does a single pass on each ".a" looking for dependencies.
> This seems to be the case both for the GNU linker and the Solaris
> linker.  A different possible workaround would have been to include
> libnotmuch.a multiple times on the link line, but that seemed like a
> brittle way to track this dependency.

I'd prefer to use timegm() where available, and the suggested alternative
[1] elsewhere. I'll look into the compat build issues when I have a moment.

Jani.

[1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html

> ---
>  parse-time-string/parse-time-string.c |   37
-
>  1 file changed, 36 insertions(+), 1 deletion(-)
>
> diff --git a/parse-time-string/parse-time-string.c
b/parse-time-string/parse-time-string.c
> index 584067d..28901af 100644
> --- a/parse-time-string/parse-time-string.c
> +++ b/parse-time-string/parse-time-string.c
> @@ -1315,6 +1315,41 @@ fixup_ampm (struct state *state)
>  return 0;
>  }
>
> +static int
> +leapyear (int year)
> +{
> +return ((year % 4) == 0 && ((year % 100) != 0 || (year % 400) == 0));
> +}
> +
> +/*
> + * This is a simple implementation of timegm() which does what is needed
> + * by create_output() -- just turns the "struct tm" into a GMT time_t.
> + * It does not normalize any of the fields of the "struct tm", nor does
> + * it set tm_wday or tm_yday.
> + */
> +static time_t
> +local_timegm (struct tm *tm)
> +{
> +intmonthlen[2][12] = {
> +   { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
> +   { 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
> +};
> +intyear, month, days;
> +
> +days = 365 * (tm->tm_year - 70);
> +for (year = 70; year < tm->tm_year; year++) {
> +   if (leapyear(1900 + year)) {
> +   days++;
> +   }
> +}
> +for (month = 0; month < tm->tm_mon; month++) {
> +   days += monthlen[leapyear(1900 + year)][month];
> +}
> +days += tm->tm_mday - 1;
> +
> +return days * 24) + tm->tm_hour) * 60 + tm->tm_min) * 60 +
tm->tm_sec);
> +}
> +
>  /* Combine absolute and relative fields, and round. */
>  static int
>  create_output (struct state *state, time_t *t_out, const time_t *ref,
> @@ -1465,7 +1500,7 @@ create_output (struct state *state, time_t *t_out,
const time_t *ref,
>  if (is_field_set (state, TM_TZ)) {
> /* tm is in specified TZ, convert to UTC for timegm(3). */
> tm.tm_min -= get_field (state, TM_TZ);
> -   t = timegm ();
> +   t = local_timegm ();
>  } else {
> /* tm is in local time. */
> t = mktime ();
> --
> 1.7.9.2
>
> ___
> notmuch mailing list
> notmuch at notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch
-- next part --
An HTML attachment was scrubbed...
URL: 
<http://notmuchmail.org/pipermail/notmuch/attachments/20121104/511b837a/attachment-0001.html>

Automatic suppression of non-duplicate messages

[Eirik Byrkjeflot Anonsen]

David Bremner  writes:

> Eirik Byrkjeflot Anonsen  writes:
>
>> That's not what I see.  If I search for a term that only appears in
>> one of the "copies", none of the copies are included in the search
>> result.
>
> The offending code is at line 1813 of lib/database.cc; the message is
> only indexed if the message-id is new.
>
> It might be sensible to move _notmuch_message_index_file into the other
> branch of the if, but even if that works fine, something more
> sophisticated is needed for the call to
> __notmuch_message_set_header_values; the invariant that each message has
> a single subject seems reasonable.

Hmm, depends.  Assuming indexing is intended to be used for searching,
one might want to search for something that occurs in one subject but
not the other.  In practice I doubt it matters.


> Offhand I'm not sure of a good method of automatically deciding what is
> the same message (with e.g. headers and footer text added by a mailing
> list).

I don't think the real problem here is the duplicate detection algorithm
itself.  It is rather that notmuch forces a particular duplicate
detection algorithm on its users.  Duplicate detection should really be
delegated to a different application, thus allowing people to experiment
with whatever algorithm works best for them.  (Just like notmuch
delegates the choice of initial tags on messages to an external
application.)

But first notmuch must be modified so it can sensibly treat multiple
instances having the same message-id as separate messages.  That seems
to me to be the hard part.  (And some way for external applications to
join and split copies, of course.)




However, if you want an algorithm that is likely to get rid of most
duplicates while keeping most non-duplicates separate, here's a quick
suggestion:


Just to clarify: The goal is to suppress most copies of the same message
while not suppressing a single instance of a different message.  It
isn't important if a few duplicate messages makes it through, but it is
imperative that no "real" message is dropped.

To check whether two instances are duplicates, I suspect something like
this algorithm would be "good enough":

- Message-Id must be the same.  This isn't actually necessary, but it
  makes sense to require it anyway.

- From and Date must be the same.  These form important context that may
  change the meaning of the message (e.g. "me too" depends heavily on
  From, and "let's meet tomorrow" depends heavily on Date).  (Are there
  more context-supplying headers we should worry about?)

- If Subject and body are also the same, the instances are duplicates.

- Otherwise, if neither of the messages come from a mailing list,
  they're probably not duplicates.

- Otherwise, grab a few other (recent) mails from the same mailing list.
  If all the bodies end with the same text, ignore that text when
  comparing the bodies.

- For the Subject, again use a few other (recent) mails from the same
  mailing list for comparison.  But this time only look for one of the
  well-known common patterns.  If all the mails matches the same
  pattern, ignore that pattern when comparing the Subject.

- For both of the above, it would be good to pick messages from
  different threads, to avoid accidental similarities.  I suspect this
  is more important for subjects than bodies, though.

- Also, leading and trailing whitespace should probably be dropped.

- (Some other transformations may make sense, such as reflowing text or
  converting between character sets.  In practice I doubt that will make
  much of a difference.)

- If the "canonicalized" body and Subject are the same, the messages are
  duplicates.  At least there's now pretty much no chance that there is
  anything interesting that will be missed by dropping one of the
  messages.


(I'm assuming that identifying mailing lists are usually
straightforward, e.g. using the List-Id header).

eirik

[PATCH 10/10] timegm: add portable implementation (Solaris support)

[Blake Jones]

Hi Jani,

> I'd prefer to use timegm() where available, and the suggested
> alternative [1] elsewhere.
> 
> [1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html

I considered this alternative, but decided against it because it's
completely MT-unsafe.  I don't know whether libnotmuch itself is
MT-safe, but a process which called this routine in one thread would
temporarily throw off any timezone-related work that any other threads
were doing, even if they weren't using libnotmuch.

> I'll look into the compat build issues when I have a moment.

If you do, here's a boiled-down version of the problem that I came up
with while investigating it:

$ echo 'int main() { extern int foo1(); return foo1(); }' > main.c
$ echo 'int foo1() { extern int bar();  return bar();  }' > foo1.c
$ echo 'int bar()  { extern int foo2(); return foo2(); }' > bar.c
$ echo 'int foo2() { return 0; }' > foo2.c
$ gcc -c main.c foo1.c bar.c foo2.c
$ ar rcs libfoo.a foo1.o foo2.o
$ ar rcs libbar.a bar.o
$ gcc -o main main.o libfoo.a libbar.a
  [fails]
$ gcc -o main main.o libbar.a libfoo.a
  [fails]

Another alternative would be to just include parse-time-string.o in
libnotmuch.a directly; I think that would solve the problem.

Blake

[PATCH v2] contrib: notmuch-pick: add tests

[Tomi Ollila]

On Thu, Nov 01 2012, Mark Walters  wrote:

> The test should be run using the wrapper run-tests.sh.  This links
> the tests into the normal notmuch TEST_DIRECTORY and runs them from
> there. After the test is complete then the links are removed.
> ---

2 things:

The PICK_DIR variable needs export in run-tests.sh, otherwise it is
not (always??!!) available for test scripts (all tests failed
for me until I added that export (fatal).

git am informs  test/pick.expected-output/notmuch-pick-show-window 
has trailing whitespace when applied. Could you arrange your test
to use other message there (wish).


Tomi


>
> This is version 2 of the patches to add some tests for notmuch pick. 
>
> The main change is that I have added Tomi's suggested wrapper script
> (id:m2ip9srpd6.fsf at guru.guru-group.fi) which is much more robust and
> generally better than the one I had.
>
> The other changes are to split out the tests for the default async
> parser from those for the sync parser. My guess is that the sync
> parser should be removed but it can be useful for debugging so is
> probably worth keeping until pick is closer to being ready for proper
> mainline.
>
> Best wishes
>
> Mark
>
>
>
>
>  contrib/notmuch-pick/README|5 ++
>  contrib/notmuch-pick/run-tests.sh  |   46 
>  contrib/notmuch-pick/test/emacs-pick   |   76 
> 
>  contrib/notmuch-pick/test/emacs-pick-sync  |   64 
>  .../pick.expected-output/notmuch-pick-show-window  |   32 
>  .../notmuch-pick-single-thread |6 ++
>  .../pick.expected-output/notmuch-pick-tag-inbox|   53 ++
>  7 files changed, 282 insertions(+), 0 deletions(-)
>  create mode 100755 contrib/notmuch-pick/run-tests.sh
>  create mode 100755 contrib/notmuch-pick/test/emacs-pick
>  create mode 100755 contrib/notmuch-pick/test/emacs-pick-sync
>  create mode 100644 
> contrib/notmuch-pick/test/pick.expected-output/notmuch-pick-show-window
>  create mode 100644 
> contrib/notmuch-pick/test/pick.expected-output/notmuch-pick-single-thread
>  create mode 100644 
> contrib/notmuch-pick/test/pick.expected-output/notmuch-pick-tag-inbox
>
> diff --git a/contrib/notmuch-pick/README b/contrib/notmuch-pick/README
> index 8eed974..4200824 100644
> --- a/contrib/notmuch-pick/README
> +++ b/contrib/notmuch-pick/README
> @@ -15,6 +15,11 @@ Then after the "(require 'notmuch)" line in your .emacs 
> file add
>  the line "(require 'notmuch-pick nil t)". This will load notmuch-pick on
>  your next emacs start.
>  
> +TEST
> +
> +Just execute run-tests.sh and it should all work (it does require that
> +notmuch has already been built).
> +
>  USING PICK
>  
>  The main key entries to notmuch pick are
> diff --git a/contrib/notmuch-pick/run-tests.sh 
> b/contrib/notmuch-pick/run-tests.sh
> new file mode 100755
> index 000..97cc9e8
> --- /dev/null
> +++ b/contrib/notmuch-pick/run-tests.sh
> @@ -0,0 +1,46 @@
> +#!/usr/bin/env bash
> +
> +set -eu
> +
> +fail() {
> +echo ERROR $1
> +exit 1
> +}
> +
> +TESTS="emacs-pick emacs-pick-sync"
> +TESTFILES="$TESTS pick.expected-output"
> +
> +PICK_DIR="`cd \`dirname "$0"\` && pwd`"
> +PICK_TEST_DIR="$PICK_DIR/test"
> +
> +
> +for f in $TESTFILES
> +do
> +test -f "$PICK_TEST_DIR/$f" || test -d "$PICK_TEST_DIR/$f" || fail 
> "$PICK_TEST_DIR/$f does not exist"
> +done
> +
> +cd "$PICK_DIR/../../test"
> +
> +test -x ../notmuch || fail "`cd .. && pwd`/notmuch has not been built"
> +
> +for f in $TESTFILES
> +do
> +if test -f "$f"
> +then
> + fail "$f exists"
> +fi
> +done
> +
> +trap "rm -f $TESTFILES" 0
> +
> +for f in $TESTFILES
> +do
> +ln -s "$PICK_TEST_DIR/$f" .
> +done
> +
> +#don't exec -- traps would not run.
> +for f in $TESTS
> +do
> +echo $f
> +./$f
> +done
> diff --git a/contrib/notmuch-pick/test/emacs-pick 
> b/contrib/notmuch-pick/test/emacs-pick
> new file mode 100755
> index 000..9db4b34
> --- /dev/null
> +++ b/contrib/notmuch-pick/test/emacs-pick
> @@ -0,0 +1,76 @@
> +#!/usr/bin/env bash
> +
> +test_description="emacs pick interface"
> +. test-lib.sh
> +
> +EXPECTED=$TEST_DIRECTORY/pick.expected-output
> +
> +add_email_corpus
> +test_begin_subtest "Do we have emacs"
> +test_emacs '(insert "hello\n")
> + (test-output)'
> +cat  +hello
> +EOF
> +test_expect_equal_file OUTPUT EXPECTED
> +
> +test_begin_subtest "Basic notmuch-pick view in emacs"
> +test_emacs '(add-to-list (quote load-path) "'$PICK_DIR'")
> + (require (quote notmuch-pick))
> + (notmuch-pick "tag:inbox")
> + (notmuch-test-wait)
> + (test-output)
> + (delete-other-windows)'
> +test_expect_equal_file OUTPUT $EXPECTED/notmuch-pick-tag-inbox
> +
> +test_begin_subtest "Navigation of notmuch-hello to search results"
> +test_emacs '(notmuch-hello)
> + (goto-char (point-min))
> + (re-search-forward "inbox")
> + 

[PATCH 09/10] debugger.c: correct return type from getppid() (Solaris support)

[Blake Jones]

Cast the return value of getppid() to int from pid_t in debugger.c,
since it is being passed to sprintf(%d), which wants an int
argument.  On Solaris, pid_t is a long for 32-bit programs.
---
 debugger.c |2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/debugger.c b/debugger.c
index e8b9378..8ff13d6 100644
--- a/debugger.c
+++ b/debugger.c
@@ -36,7 +36,7 @@ debugger_is_active (void)
 if (RUNNING_ON_VALGRIND)
return TRUE;
 
-sprintf (buf, /proc/%d/exe, getppid ());
+sprintf (buf, /proc/%d/exe, (int) getppid ());
 if (readlink (buf, buf, sizeof (buf)) != -1 
strncmp (basename (buf), gdb, 3) == 0)
 {
-- 
1.7.9.2

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

[PATCH 04/10] configure: check for -Wl,-rpath (Solaris support)

[Blake Jones]

Add a check to configure to see whether -Wl,-rpath can be used without
--enable-new-dtags.  Solaris needs the former and doesn't know about the
latter.
---
 configure |4 
 1 file changed, 4 insertions(+)

diff --git a/configure b/configure
index 28d4110..5c5139f 100755
--- a/configure
+++ b/configure
@@ -552,6 +552,10 @@ if ${CC} -Wl,--enable-new-dtags -Wl,-rpath,/tmp/ -o 
minimal minimal.c /dev/null
 then
 printf Yes.\n
 rpath_ldflags=-Wl,--enable-new-dtags -Wl,-rpath,\$(libdir)
+elif ${CC} -Wl,-rpath,/tmp/ -o minimal minimal.c /dev/null 21
+then
+printf Yes.\n
+rpath_ldflags=-Wl,-rpath,\$(libdir)
 else
 printf No (nothing to worry about).\n
 rpath_ldflags=
-- 
1.7.9.2

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

[PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Blake Jones]

Linux, FreeBSD, and Solaris all expect to find the prototype for
index() in strings.h.  On some operating systems, including
string.h is sufficient to get the prototype, but that's not the case
on Solaris.  This patch just modifies notmuch-config.c to include
strings.h to get the prototype.
---
 notmuch-config.c |1 +
 1 file changed, 1 insertion(+)

diff --git a/notmuch-config.c b/notmuch-config.c
index 3e37a2d..2537a65 100644
--- a/notmuch-config.c
+++ b/notmuch-config.c
@@ -23,6 +23,7 @@
 #include pwd.h
 #include netdb.h
 #include assert.h
+#include strings.h
 
 static const char toplevel_config_comment[] =
  .notmuch-config - Configuration file for the notmuch mail system\n
-- 
1.7.9.2

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

[PATCH 03/10] gethostbyname: check for libnsl (Solaris support)

[Blake Jones]

Add a check to configure to see whether -lnsl is needed for programs
that are using gethostbyname().  This change also adds the file
compat/check_ghbn.c, which configure uses to perform its check.
---
 compat/check_ghbn.c |8 
 configure   |   17 -
 2 files changed, 24 insertions(+), 1 deletion(-)
 create mode 100644 compat/check_ghbn.c

diff --git a/compat/check_ghbn.c b/compat/check_ghbn.c
new file mode 100644
index 000..ec5f227
--- /dev/null
+++ b/compat/check_ghbn.c
@@ -0,0 +1,8 @@
+#include netdb.h
+
+int main()
+{
+(void) gethostbyname(NULL);
+
+return (0);
+}
diff --git a/configure b/configure
index 047c011..28d4110 100755
--- a/configure
+++ b/configure
@@ -534,6 +534,17 @@ else
 fi
 rm -f compat/check_asctime
 
+printf Checking whether libnsl is needed for gethostbyname... 
+if ${CC} -o compat/check_ghbn $srcdir/compat/check_ghbn.c  /dev/null 21
+then
+printf No.\n
+libnsl_ldflags=
+else
+printf Yes.\n
+libnsl_ldflags=-lnsl
+fi
+rm -f compat/check_ghbn
+
 printf int main(void){return 0;}\n  minimal.c
 
 printf Checking for rpath support... 
@@ -723,6 +734,9 @@ GMIME_LDFLAGS = ${gmime_ldflags}
 TALLOC_CFLAGS = ${talloc_cflags}
 TALLOC_LDFLAGS = ${talloc_ldflags}
 
+# Flags needed to get gethostbyname() at link time
+LIBNSL_LDFLAGS = ${libnsl_ldflags}
+
 # Flags needed to have linker set rpath attribute
 RPATH_LDFLAGS = ${rpath_ldflags}
 
@@ -757,5 +771,6 @@ CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) 
\$(GMIME_CFLAGS)\\
 -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR)   \\
 -DSTD_GETPWUID=\$(STD_GETPWUID) \\
 -DSTD_ASCTIME=\$(STD_ASCTIME)
-CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS)
+CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS) \\
+\$(LIBNSL_LDFLAGS)
 EOF
-- 
1.7.9.2

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

[PATCH 10/10] timegm: add portable implementation (Solaris support)

[Blake Jones]

The timegm(3) function is a non-standard extension to libc which is
available in GNU libc and on some BSDs.  Although SunOS had this
function in its libc, Solaris (unfortunately) removed it.  This patch
implements a very simple version of timegm() which is good enough for
parse-time-string.c.

Although notmuch's idiom for portability is to test for native
availability and put alternate versions in compat/, that approach led to
a compilation problem in this case.  libnotmuch.a includes a call to
parse_time_string() from parse-time-vrp.o, and parse_time_string() in
libparse-time-string.a needs to call timegm().  An attempt to create
compat/timegm.c caused the link to fail, because libparse-time-string.a
acquired a dependency on the new timegm.o in libnotmuch.a, and the
linker only does a single pass on each .a looking for dependencies.
This seems to be the case both for the GNU linker and the Solaris
linker.  A different possible workaround would have been to include
libnotmuch.a multiple times on the link line, but that seemed like a
brittle way to track this dependency.
---
 parse-time-string/parse-time-string.c |   37 -
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/parse-time-string/parse-time-string.c 
b/parse-time-string/parse-time-string.c
index 584067d..28901af 100644
--- a/parse-time-string/parse-time-string.c
+++ b/parse-time-string/parse-time-string.c
@@ -1315,6 +1315,41 @@ fixup_ampm (struct state *state)
 return 0;
 }
 
+static int
+leapyear (int year)
+{
+return ((year % 4) == 0  ((year % 100) != 0 || (year % 400) == 0));
+}
+
+/*
+ * This is a simple implementation of timegm() which does what is needed
+ * by create_output() -- just turns the struct tm into a GMT time_t.
+ * It does not normalize any of the fields of the struct tm, nor does
+ * it set tm_wday or tm_yday.
+ */
+static time_t
+local_timegm (struct tm *tm)
+{
+intmonthlen[2][12] = {
+   { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
+   { 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
+};
+intyear, month, days;
+
+days = 365 * (tm-tm_year - 70);
+for (year = 70; year  tm-tm_year; year++) {
+   if (leapyear(1900 + year)) {
+   days++;
+   }
+}
+for (month = 0; month  tm-tm_mon; month++) {
+   days += monthlen[leapyear(1900 + year)][month];
+}
+days += tm-tm_mday - 1;
+
+return days * 24) + tm-tm_hour) * 60 + tm-tm_min) * 60 + tm-tm_sec);
+}
+
 /* Combine absolute and relative fields, and round. */
 static int
 create_output (struct state *state, time_t *t_out, const time_t *ref,
@@ -1465,7 +1500,7 @@ create_output (struct state *state, time_t *t_out, const 
time_t *ref,
 if (is_field_set (state, TM_TZ)) {
/* tm is in specified TZ, convert to UTC for timegm(3). */
tm.tm_min -= get_field (state, TM_TZ);
-   t = timegm (tm);
+   t = local_timegm (tm);
 } else {
/* tm is in local time. */
t = mktime (tm);
-- 
1.7.9.2

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

[PATCH 01/10] getpwuid: check for standards compliance (Solaris support)

[Blake Jones]

Add checks to configure to see whether _POSIX_PTHREAD_SEMANTICS needs
to be defined to get the right number of arguments in the prototypes for
getpwuid_r().  Solaris' default implementation conforms to POSIX.1c
Draft 6, rather than the final POSIX.1c spec.  The standards-compliant
version can be used by defining _POSIX_PTHREAD_SEMANTICS.

This change also adds the file compat/check_getpwuid.c, which
configure uses to perform its check, and modifies compat/compat.h to
define _POSIX_PTHREAD_SEMANTICS if configure detected it was needed.
---
 compat/check_getpwuid.c |   10 ++
 compat/compat.h |4 
 configure   |   23 +--
 3 files changed, 35 insertions(+), 2 deletions(-)
 create mode 100644 compat/check_getpwuid.c

diff --git a/compat/check_getpwuid.c b/compat/check_getpwuid.c
new file mode 100644
index 000..5da2590
--- /dev/null
+++ b/compat/check_getpwuid.c
@@ -0,0 +1,10 @@
+#include pwd.h
+
+int main()
+{
+struct passwd passwd, *ignored;
+
+(void) getpwuid_r (0, passwd, NULL, 0, ignored);
+
+return (0);
+}
diff --git a/compat/compat.h b/compat/compat.h
index b2e2736..8374d2f 100644
--- a/compat/compat.h
+++ b/compat/compat.h
@@ -54,6 +54,10 @@ char* strcasestr(const char *haystack, const char *needle);
 #define IGNORE_RESULT(x) x
 #endif /* __GNUC__ */
 
+#if !STD_GETPWUID
+#define _POSIX_PTHREAD_SEMANTICS
+#endif
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/configure b/configure
index ea8a1ad..3c18a45 100755
--- a/configure
+++ b/configure
@@ -512,6 +512,17 @@ else
 fi
 rm -f compat/have_strcasestr
 
+printf Checking for standard version of getpwuid_r... 
+if ${CC} -o compat/check_getpwuid $srcdir/compat/check_getpwuid.c  
/dev/null 21
+then
+printf Yes.\n
+std_getpwuid=1
+else
+printf No (will define _POSIX_PTHREAD_SEMANTICS to get it).\n
+std_getpwuid=0
+fi
+rm -f compat/check_getpwuid
+
 printf int main(void){return 0;}\n  minimal.c
 
 printf Checking for rpath support... 
@@ -671,6 +682,11 @@ HAVE_GETLINE = ${have_getline}
 # build its own version)
 HAVE_STRCASESTR = ${have_strcasestr}
 
+# Whether the getpwuid_r function is standards-compliant
+# (if not, then notmuch will compile with -D_POSIX_PTHREAD_SEMANTICS
+# to enable the standards-compliant version -- needed for Solaris)
+STD_GETPWUID = ${std_getpwuid}
+
 # Supported platforms (so far) are: LINUX, MACOSX, SOLARIS, FREEBSD, OPENBSD
 PLATFORM = ${platform}
 
@@ -715,10 +731,13 @@ WITH_ZSH = ${WITH_ZSH}
 # Combined flags for compiling and linking against all of the above
 CONFIGURE_CFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) \$(GMIME_CFLAGS)  \\
   \$(TALLOC_CFLAGS) -DHAVE_VALGRIND=\$(HAVE_VALGRIND)   \\
-  \$(VALGRIND_CFLAGS) -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR)
+  \$(VALGRIND_CFLAGS)   \\
+  -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR) \\
+  -DSTD_GETPWUID=\$(STD_GETPWUID)
 CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) \$(GMIME_CFLAGS)\\
 \$(TALLOC_CFLAGS) -DHAVE_VALGRIND=\$(HAVE_VALGRIND) \\
 \$(VALGRIND_CFLAGS) \$(XAPIAN_CXXFLAGS) \\
- -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR)
+-DHAVE_STRCASESTR=\$(HAVE_STRCASESTR)   \\
+-DSTD_GETPWUID=\$(STD_GETPWUID)
 CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS)
 EOF
-- 
1.7.9.2

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

[PATCH 07/10] gen-version-script: parse Solaris nm output (Solaris support)

[Blake Jones]

The output of nm on Solaris is substantially different from that on
Linux, and the current version of gen-version-script is tied to the
Linux nm output.  This patch separates the parts of nm processing
which are dependent on the output format into a couple shell functions,
and makes another shell function to use the appropriate version of
c++filt to demangle symbols.  It also modifies lib/Makefile.local
to pass the generated symbol table correctly to the Solaris linker.
---
 lib/Makefile.local|4 
 lib/gen-version-script.sh |   50 -
 2 files changed, 49 insertions(+), 5 deletions(-)

diff --git a/lib/Makefile.local b/lib/Makefile.local
index 0c6b258..2068e4a 100644
--- a/lib/Makefile.local
+++ b/lib/Makefile.local
@@ -30,7 +30,11 @@ LIBRARY_SUFFIX = so
 LINKER_NAME = libnotmuch.$(LIBRARY_SUFFIX)
 SONAME = $(LINKER_NAME).$(LIBNOTMUCH_VERSION_MAJOR)
 LIBNAME = $(SONAME).$(LIBNOTMUCH_VERSION_MINOR).$(LIBNOTMUCH_VERSION_RELEASE)
+ifeq ($(PLATFORM),SOLARIS)
+LIBRARY_LINK_FLAG = -shared -Wl,-M notmuch.sym -Wl,-soname=$(SONAME) 
-Wl,--no-undefined -lc
+else
 LIBRARY_LINK_FLAG = -shared -Wl,--version-script=notmuch.sym,-soname=$(SONAME) 
-Wl,--no-undefined
+endif
 ifeq ($(PLATFORM),OPENBSD)
 LIBRARY_LINK_FLAG += -lc
 endif
diff --git a/lib/gen-version-script.sh b/lib/gen-version-script.sh
index 76670d5..d7d96da 100644
--- a/lib/gen-version-script.sh
+++ b/lib/gen-version-script.sh
@@ -1,3 +1,4 @@
+#!/bin/sh
 
 # we go through a bit of work to get the unmangled names of the
 # typeinfo symbols because of
@@ -11,10 +12,44 @@ fi
 HEADER=$1
 shift
 
+if [ `uname -s` == SunOS ] ; then
+#
+# Using Solaris nm, a defined symbol looks like this:
+#
+# [Index]ValueSize Type  Bind  Other Shndx   Name
+# [15]|128| 16|FUNC |GLOB |0|1  |notmuch_tags_get
+#
+# and an undefined symbol looks like this:
+#
+# [Index]ValueSize Type  Bind  Other Shndx   Name
+# [35]|  0|  0|NOTY |GLOB |0|UNDEF  |notmuch_tags_get
+#
+find_xapian_error() {
+   nawk -F'\|' '$7 !~ UNDEF  $8 ~ Xapian.*Error { print $8 }'
+}
+find_compat_syms() {
+   nawk -F'\|' '$7 !~ UNDEF  $8 ~ get(line|delim) { print $8 ; }'
+}
+demangle() {
+   gc++filt $@
+}
+else
+find_xapian_error() {
+   awk '$1 ~ ^[0-9a-fA-F][0-9a-fA-F]*$  $3 ~ Xapian.*Error {print 
$3}'
+}
+find_compat_syms() {
+   awk '$1 ~ ^[0-9a-fA-F][0-9a-fA-F]*$  $2 == T  $3 ~ 
^get(line|delim)$ {print $3 ;}'
+}
+demangle() {
+   c++filt $@
+}
+fi
+
 printf '{\nglobal:\n'
-nm  $* | awk '$1 ~ ^[0-9a-fA-F][0-9a-fA-F]*$  $3 ~ Xapian.*Error {print 
$3}' | sort | uniq | \
-while read sym; do
-demangled=$(c++filt $sym)
+
+# Find the typeinfo for Xapian::*Errors.
+nm $* | find_xapian_error | sort | uniq | while read sym; do
+demangled=$(demangle $sym)
 case $demangled in
typeinfo*) 
printf \t$sym;\n
@@ -23,6 +58,11 @@ while read sym; do
;;
 esac
 done
-nm $* | awk '$1 ~ ^[0-9a-fA-F][0-9a-fA-F]*$  $2 == T  $3 ~ 
^get(line|delim)$ {print $3 ;}'
-sed  -n 's/^[[:space:]]*\(notmuch_[a-z_]*\)[[:space:]]*(.*/ \1;/p' $HEADER
+
+# Find the compat syms that we need to export.
+nm $* | find_compat_syms
+
+# Finally, get the real notmuch symbols.
+sed -n 's/^[   ]*\(notmuch_[a-z_]*\)[  ]*(.*/ \1;/p' $HEADER
+
 printf local: *;\n};\n
-- 
1.7.9.2

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

[PATCH 02/10] asctime: check for standards compliance (Solaris support)

[Blake Jones]

Add checks to configure to see whether _POSIX_PTHREAD_SEMANTICS needs
to be defined to get the right number of arguments in the prototypes for
asctime_r().  Solaris' default implementation conforms to POSIX.1c
Draft 6, rather than the final POSIX.1c spec.  The standards-compliant
version can be used by defining _POSIX_PTHREAD_SEMANTICS.

This change also adds the file compat/check_asctime.c, which
configure uses to perform its check, and modifies compat/compat.h to
define _POSIX_PTHREAD_SEMANTICS if configure detected it was needed.
---
 compat/check_asctime.c |   17 +
 compat/compat.h|3 +++
 configure  |   22 --
 3 files changed, 40 insertions(+), 2 deletions(-)
 create mode 100644 compat/check_asctime.c

diff --git a/compat/check_asctime.c b/compat/check_asctime.c
new file mode 100644
index 000..a595110
--- /dev/null
+++ b/compat/check_asctime.c
@@ -0,0 +1,17 @@
+/*
+ * This compatibility check actually succeeds (on Solaris) if
+ * _POSIX_PTHREAD_SEMANTICS is not defined.  But we need to define that to get
+ * the right version of getpwuid_r(), so we define it here to ensure that the
+ * compatibility check ends up doing the same thing as the rest of the code.
+ */
+#define_POSIX_PTHREAD_SEMANTICS
+#include time.h
+
+int main()
+{
+struct tm tm;
+
+(void) asctime_r (tm, NULL, 0);
+
+return (0);
+}
diff --git a/compat/compat.h b/compat/compat.h
index 8374d2f..b750501 100644
--- a/compat/compat.h
+++ b/compat/compat.h
@@ -57,6 +57,9 @@ char* strcasestr(const char *haystack, const char *needle);
 #if !STD_GETPWUID
 #define _POSIX_PTHREAD_SEMANTICS
 #endif
+#if !STD_ASCTIME
+#define _POSIX_PTHREAD_SEMANTICS
+#endif
 
 #ifdef __cplusplus
 }
diff --git a/configure b/configure
index 3c18a45..047c011 100755
--- a/configure
+++ b/configure
@@ -523,6 +523,17 @@ else
 fi
 rm -f compat/check_getpwuid
 
+printf Checking for standard version of asctime_r... 
+if ${CC} -o compat/check_asctime $srcdir/compat/check_asctime.c  /dev/null 
21
+then
+printf Yes.\n
+std_asctime=1
+else
+printf No (will define _POSIX_PTHREAD_SEMANTICS to get it).\n
+std_asctime=0
+fi
+rm -f compat/check_asctime
+
 printf int main(void){return 0;}\n  minimal.c
 
 printf Checking for rpath support... 
@@ -687,6 +698,11 @@ HAVE_STRCASESTR = ${have_strcasestr}
 # to enable the standards-compliant version -- needed for Solaris)
 STD_GETPWUID = ${std_getpwuid}
 
+# Whether the asctime_r function is standards-compliant
+# (if not, then notmuch will compile with -D_POSIX_PTHREAD_SEMANTICS
+# to enable the standards-compliant version -- needed for Solaris)
+STD_ASCTIME = ${std_asctime}
+
 # Supported platforms (so far) are: LINUX, MACOSX, SOLARIS, FREEBSD, OPENBSD
 PLATFORM = ${platform}
 
@@ -733,11 +749,13 @@ CONFIGURE_CFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) 
\$(GMIME_CFLAGS)  \\
   \$(TALLOC_CFLAGS) -DHAVE_VALGRIND=\$(HAVE_VALGRIND)   \\
   \$(VALGRIND_CFLAGS)   \\
   -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR) \\
-  -DSTD_GETPWUID=\$(STD_GETPWUID)
+  -DSTD_GETPWUID=\$(STD_GETPWUID)   \\
+  -DSTD_ASCTIME=\$(STD_ASCTIME)
 CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) \$(GMIME_CFLAGS)\\
 \$(TALLOC_CFLAGS) -DHAVE_VALGRIND=\$(HAVE_VALGRIND) \\
 \$(VALGRIND_CFLAGS) \$(XAPIAN_CXXFLAGS) \\
 -DHAVE_STRCASESTR=\$(HAVE_STRCASESTR)   \\
--DSTD_GETPWUID=\$(STD_GETPWUID)
+-DSTD_GETPWUID=\$(STD_GETPWUID) \\
+-DSTD_ASCTIME=\$(STD_ASCTIME)
 CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS)
 EOF
-- 
1.7.9.2

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

[PATCH 06/10] strsep: check for availability (Solaris support)

[Blake Jones]

Solaris does not ship a version of the strsep() function.  This change
adds a check to configure to see whether notmuch needs to provide its
own implementation, and if so, it uses the new version in
compat/strsep.c (which was copied from Mutt, and apparently before
that from glibc).
---
 compat/Makefile.local |4 +++
 compat/compat.h   |4 +++
 compat/have_strsep.c  |   10 
 compat/strsep.c   |   65 +
 configure |   17 +
 5 files changed, 100 insertions(+)
 create mode 100644 compat/have_strsep.c
 create mode 100644 compat/strsep.c

diff --git a/compat/Makefile.local b/compat/Makefile.local
index 13f16cd..2c4f65f 100644
--- a/compat/Makefile.local
+++ b/compat/Makefile.local
@@ -13,4 +13,8 @@ ifneq ($(HAVE_STRCASESTR),1)
 notmuch_compat_srcs += $(dir)/strcasestr.c
 endif
 
+ifneq ($(HAVE_STRSEP),1)
+notmuch_compat_srcs += $(dir)/strsep.c
+endif
+
 SRCS := $(SRCS) $(notmuch_compat_srcs)
diff --git a/compat/compat.h b/compat/compat.h
index b750501..1fd2723 100644
--- a/compat/compat.h
+++ b/compat/compat.h
@@ -46,6 +46,10 @@ getdelim (char **lineptr, size_t *n, int delimiter, FILE 
*fp);
 char* strcasestr(const char *haystack, const char *needle);
 #endif /* !HAVE_STRCASESTR */
 
+#if !HAVE_STRSEP
+char *strsep(char **stringp, const char *delim);
+#endif /* !HAVE_STRSEP */
+
 /* Silence gcc warnings about unused results.  These warnings exist
  * for a reason; any use of this needs to be justified. */
 #ifdef __GNUC__
diff --git a/compat/have_strsep.c b/compat/have_strsep.c
new file mode 100644
index 000..5bd396c
--- /dev/null
+++ b/compat/have_strsep.c
@@ -0,0 +1,10 @@
+#define _GNU_SOURCE
+#include string.h
+
+int main()
+{
+char *found;
+char **stringp, const char *delim;
+
+found = strsep(stringp, delim);
+}
diff --git a/compat/strsep.c b/compat/strsep.c
new file mode 100644
index 000..78ab9e7
--- /dev/null
+++ b/compat/strsep.c
@@ -0,0 +1,65 @@
+/* Copyright (C) 1992, 93, 96, 97, 98, 99, 2004 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include string.h
+
+/* Taken from glibc 2.6.1 */
+
+char *strsep (char **stringp, const char *delim)
+{
+  char *begin, *end;
+
+  begin = *stringp;
+  if (begin == NULL)
+return NULL;
+
+  /* A frequent case is when the delimiter string contains only one
+ character.  Here we don't need to call the expensive `strpbrk'
+ function and instead work using `strchr'.  */
+  if (delim[0] == '\0' || delim[1] == '\0')
+{
+  char ch = delim[0];
+
+  if (ch == '\0')
+   end = NULL;
+  else
+   {
+ if (*begin == ch)
+   end = begin;
+ else if (*begin == '\0')
+   end = NULL;
+ else
+   end = strchr (begin + 1, ch);
+   }
+}
+  else
+/* Find the end of the token.  */
+end = strpbrk (begin, delim);
+
+  if (end)
+{
+  /* Terminate the token and set *STRINGP past NUL character.  */
+  *end++ = '\0';
+  *stringp = end;
+}
+  else
+/* No more delimiters; this is the last token.  */
+*stringp = NULL;
+
+  return begin;
+}
diff --git a/configure b/configure
index dae837e..e519abc 100755
--- a/configure
+++ b/configure
@@ -512,6 +512,17 @@ else
 fi
 rm -f compat/have_strcasestr
 
+printf Checking for strsep... 
+if ${CC} -o compat/have_strsep $srcdir/compat/have_strsep.c  /dev/null 21
+then
+printf Yes.\n
+have_strsep=1
+else
+printf No (will use our own instead).\n
+have_strsep=0
+fi
+rm -f compat/have_strsep
+
 printf Checking for standard version of getpwuid_r... 
 if ${CC} -o compat/check_getpwuid $srcdir/compat/check_getpwuid.c  
/dev/null 21
 then
@@ -723,6 +734,10 @@ HAVE_GETLINE = ${have_getline}
 # build its own version)
 HAVE_STRCASESTR = ${have_strcasestr}
 
+# Whether the strsep function is available (if not, then notmuch will
+# build its own version)
+HAVE_STRSEP = ${have_strsep}
+
 # Whether the getpwuid_r function is standards-compliant
 # (if not, then notmuch will compile with -D_POSIX_PTHREAD_SEMANTICS
 # to enable the standards-compliant version -- needed for Solaris)
@@ -782,12 +797,14 @@ CONFIGURE_CFLAGS = 

[PATCH 05/10] install: check for non-SysV version (Solaris support)

[Blake Jones]

Solaris ships a program called install in /usr/sbin, which performs a
task that's fairly similar to the GNU and BSD install programs but
which uses very different command line arguments.  In particular, if it
is invoked without -c, -f, or -n, it will search the target
directory for a file with the same name as the one being installed, and
it will only install the file if it finds a matching name.  More
excitingly, if it doesn't find a match, it will look in /bin, /usr/bin,
/etc, /lib, and /usr/lib and try to do the same there.

The standard workaround for this is to use GNU install.
It is available via the standard Solaris packaging system (in
file/gnu-coreutils), and installs itself as /usr/bin/ginstall.

This patch adds a check to configure to see if install behaves in a
way that's compatible with GNU and BSD install, and if not, it uses a
program called ginstall instead.  It also modifies configure to set
the $(INSTALL) variable, and changes various Makefiles to use it.
---
 Makefile.local|2 +-
 completion/Makefile.local |4 ++--
 configure |   19 +++
 emacs/Makefile.local  |6 +++---
 lib/Makefile.local|4 ++--
 man/Makefile.local|6 +++---
 vim/Makefile  |6 ++
 7 files changed, 32 insertions(+), 15 deletions(-)

diff --git a/Makefile.local b/Makefile.local
index 2b91946..7ccb1cd 100644
--- a/Makefile.local
+++ b/Makefile.local
@@ -286,7 +286,7 @@ notmuch-shared: $(notmuch_client_modules) lib/$(LINKER_NAME)
 .PHONY: install
 install: all install-man
mkdir -p $(DESTDIR)$(prefix)/bin/
-   install notmuch-shared $(DESTDIR)$(prefix)/bin/notmuch
+   $(INSTALL) notmuch-shared $(DESTDIR)$(prefix)/bin/notmuch
 ifeq ($(MAKECMDGOALS), install)
@echo 
@echo Notmuch is now installed to $(DESTDIR)$(prefix)
diff --git a/completion/Makefile.local b/completion/Makefile.local
index dfc1271..a648a78 100644
--- a/completion/Makefile.local
+++ b/completion/Makefile.local
@@ -14,9 +14,9 @@ install-$(dir):
@echo $@
 ifeq ($(WITH_BASH),1)
mkdir -p $(DESTDIR)$(bash_completion_dir)
-   install -m0644 $(bash_script) $(DESTDIR)$(bash_completion_dir)/notmuch
+   $(INSTALL) -m0644 $(bash_script) 
$(DESTDIR)$(bash_completion_dir)/notmuch
 endif
 ifeq ($(WITH_ZSH),1)
mkdir -p $(DESTDIR)$(zsh_completion_dir)
-   install -m0644 $(zsh_script) $(DESTDIR)$(zsh_completion_dir)/_notmuch
+   $(INSTALL) -m0644 $(zsh_script) 
$(DESTDIR)$(zsh_completion_dir)/_notmuch
 endif
diff --git a/configure b/configure
index 5c5139f..dae837e 100755
--- a/configure
+++ b/configure
@@ -591,6 +591,21 @@ for flag in -Wmissing-declarations; do
 done
 printf \n\t${WARN_CFLAGS}\n
 
+INSTALL=install
+printf Checking for working \install\ program... 
+mkdir _tmp_
+cd _tmp_
+echo 1  1
+mkdir dest
+if install 1 dest  /dev/null 21 ; then
+  printf \install\ works fine.\n
+else
+  INSTALL=ginstall
+  printf using \ginstall\.\n
+fi
+cd ..
+rm -rf _tmp_
+
 rm -f minimal minimal.c
 
 cat EOF
@@ -777,4 +792,8 @@ CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) 
\$(GMIME_CFLAGS)\\
 -DSTD_ASCTIME=\$(STD_ASCTIME)
 CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS) \\
 \$(LIBNSL_LDFLAGS)
+
+# Which install program to use
+INSTALL = ${INSTALL}
+
 EOF
diff --git a/emacs/Makefile.local b/emacs/Makefile.local
index fb82247..ee778cb 100644
--- a/emacs/Makefile.local
+++ b/emacs/Makefile.local
@@ -36,11 +36,11 @@ endif
 .PHONY: install-emacs
 install-emacs:
mkdir -p $(DESTDIR)$(emacslispdir)
-   install -m0644 $(emacs_sources) $(DESTDIR)$(emacslispdir)
+   $(INSTALL) -m0644 $(emacs_sources) $(DESTDIR)$(emacslispdir)
 ifeq ($(HAVE_EMACS),1)
-   install -m0644 $(emacs_bytecode) $(DESTDIR)$(emacslispdir)
+   $(INSTALL) -m0644 $(emacs_bytecode) $(DESTDIR)$(emacslispdir)
 endif
mkdir -p $(DESTDIR)$(emacsetcdir)
-   install -m0644 $(emacs_images) $(DESTDIR)$(emacsetcdir)
+   $(INSTALL) -m0644 $(emacs_images) $(DESTDIR)$(emacsetcdir)
 
 CLEAN := $(CLEAN) $(emacs_bytecode)
diff --git a/lib/Makefile.local b/lib/Makefile.local
index 7785944..0c6b258 100644
--- a/lib/Makefile.local
+++ b/lib/Makefile.local
@@ -89,11 +89,11 @@ install: install-$(dir)
 
 install-$(dir): $(dir)/$(LIBNAME)
mkdir -p $(DESTDIR)$(libdir)/
-   install -m0644 $(lib)/$(LIBNAME) $(DESTDIR)$(libdir)/
+   $(INSTALL) -m0644 $(lib)/$(LIBNAME) $(DESTDIR)$(libdir)/
ln -sf $(LIBNAME) $(DESTDIR)$(libdir)/$(SONAME)
ln -sf $(LIBNAME) $(DESTDIR)$(libdir)/$(LINKER_NAME)
mkdir -p $(DESTDIR)$(includedir)
-   install -m0644 $(srcdir)/$(lib)/notmuch.h $(DESTDIR)$(includedir)/
+   $(INSTALL) -m0644 $(srcdir)/$(lib)/notmuch.h 
$(DESTDIR)$(includedir)/
$(LIBRARY_INSTALL_POST_COMMAND)
 
 SRCS  := $(SRCS) $(libnotmuch_c_srcs) $(libnotmuch_cxx_srcs)
diff --git a/man/Makefile.local 

[PATCH 00/10] Solaris support

[Blake Jones]

Hi all,

This patch series fixes several issues which are needed to allow notmuch
to build on Solaris 11.  I've been testing it for a month or so by
using Karel Zak's fork of mutt along with a copy of notmuch-0.13.2 that
I got to compile.  After a friend asked for a copy of my setup, I
decided to try to get my patches cleaned up enough to submit, so that he
wouldn't need to deal with them if he wanted to hack on it.

I don't have access to any machines running a modern version of Linux
(much less *BSD), so I haven't been able to check whether these changes
have broken any other platforms.  I've tried to follow the prevailing
idiom of compatibility checks and conditionally-set variables, to
minimize the impact on other platforms, but it's possible that I
overlooked something.

I'm really pleased with my experience using notmuch; using it along with
the fork of mutt has given me real control of my email for the first
time in many years.  So I'm happy to offer these changes up in the hopes
that they can be useful to others.

I'm not subscribed to the mailing list, so please CC me on any comments.
I'll also try to watch the web archives of the list just in case.

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

Re: Automatic suppression of non-duplicate messages

[Eirik Byrkjeflot Anonsen]

David Bremner da...@tethera.net writes:

 Eirik Byrkjeflot Anonsen ei...@eirikba.org writes:

 That's not what I see.  If I search for a term that only appears in
 one of the copies, none of the copies are included in the search
 result.

 The offending code is at line 1813 of lib/database.cc; the message is
 only indexed if the message-id is new.

 It might be sensible to move _notmuch_message_index_file into the other
 branch of the if, but even if that works fine, something more
 sophisticated is needed for the call to
 __notmuch_message_set_header_values; the invariant that each message has
 a single subject seems reasonable.

Hmm, depends.  Assuming indexing is intended to be used for searching,
one might want to search for something that occurs in one subject but
not the other.  In practice I doubt it matters.


 Offhand I'm not sure of a good method of automatically deciding what is
 the same message (with e.g. headers and footer text added by a mailing
 list).

I don't think the real problem here is the duplicate detection algorithm
itself.  It is rather that notmuch forces a particular duplicate
detection algorithm on its users.  Duplicate detection should really be
delegated to a different application, thus allowing people to experiment
with whatever algorithm works best for them.  (Just like notmuch
delegates the choice of initial tags on messages to an external
application.)

But first notmuch must be modified so it can sensibly treat multiple
instances having the same message-id as separate messages.  That seems
to me to be the hard part.  (And some way for external applications to
join and split copies, of course.)




However, if you want an algorithm that is likely to get rid of most
duplicates while keeping most non-duplicates separate, here's a quick
suggestion:


Just to clarify: The goal is to suppress most copies of the same message
while not suppressing a single instance of a different message.  It
isn't important if a few duplicate messages makes it through, but it is
imperative that no real message is dropped.

To check whether two instances are duplicates, I suspect something like
this algorithm would be good enough:

- Message-Id must be the same.  This isn't actually necessary, but it
  makes sense to require it anyway.

- From and Date must be the same.  These form important context that may
  change the meaning of the message (e.g. me too depends heavily on
  From, and let's meet tomorrow depends heavily on Date).  (Are there
  more context-supplying headers we should worry about?)

- If Subject and body are also the same, the instances are duplicates.

- Otherwise, if neither of the messages come from a mailing list,
  they're probably not duplicates.

- Otherwise, grab a few other (recent) mails from the same mailing list.
  If all the bodies end with the same text, ignore that text when
  comparing the bodies.

- For the Subject, again use a few other (recent) mails from the same
  mailing list for comparison.  But this time only look for one of the
  well-known common patterns.  If all the mails matches the same
  pattern, ignore that pattern when comparing the Subject.

- For both of the above, it would be good to pick messages from
  different threads, to avoid accidental similarities.  I suspect this
  is more important for subjects than bodies, though.

- Also, leading and trailing whitespace should probably be dropped.

- (Some other transformations may make sense, such as reflowing text or
  converting between character sets.  In practice I doubt that will make
  much of a difference.)

- If the canonicalized body and Subject are the same, the messages are
  duplicates.  At least there's now pretty much no chance that there is
  anything interesting that will be missed by dropping one of the
  messages.


(I'm assuming that identifying mailing lists are usually
straightforward, e.g. using the List-Id header).

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

Re: [PATCH 10/10] timegm: add portable implementation (Solaris support)

[Jani Nikula]

On Nov 4, 2012 11:30 AM, Blake Jones bla...@foo.net wrote:

 The timegm(3) function is a non-standard extension to libc which is
 available in GNU libc and on some BSDs.  Although SunOS had this
 function in its libc, Solaris (unfortunately) removed it.  This patch
 implements a very simple version of timegm() which is good enough for
 parse-time-string.c.

 Although notmuch's idiom for portability is to test for native
 availability and put alternate versions in compat/, that approach led to
 a compilation problem in this case.  libnotmuch.a includes a call to
 parse_time_string() from parse-time-vrp.o, and parse_time_string() in
 libparse-time-string.a needs to call timegm().  An attempt to create
 compat/timegm.c caused the link to fail, because libparse-time-string.a
 acquired a dependency on the new timegm.o in libnotmuch.a, and the
 linker only does a single pass on each .a looking for dependencies.
 This seems to be the case both for the GNU linker and the Solaris
 linker.  A different possible workaround would have been to include
 libnotmuch.a multiple times on the link line, but that seemed like a
 brittle way to track this dependency.

I'd prefer to use timegm() where available, and the suggested alternative
[1] elsewhere. I'll look into the compat build issues when I have a moment.

Jani.

[1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html

 ---
  parse-time-string/parse-time-string.c |   37
-
  1 file changed, 36 insertions(+), 1 deletion(-)

 diff --git a/parse-time-string/parse-time-string.c
b/parse-time-string/parse-time-string.c
 index 584067d..28901af 100644
 --- a/parse-time-string/parse-time-string.c
 +++ b/parse-time-string/parse-time-string.c
 @@ -1315,6 +1315,41 @@ fixup_ampm (struct state *state)
  return 0;
  }

 +static int
 +leapyear (int year)
 +{
 +return ((year % 4) == 0  ((year % 100) != 0 || (year % 400) == 0));
 +}
 +
 +/*
 + * This is a simple implementation of timegm() which does what is needed
 + * by create_output() -- just turns the struct tm into a GMT time_t.
 + * It does not normalize any of the fields of the struct tm, nor does
 + * it set tm_wday or tm_yday.
 + */
 +static time_t
 +local_timegm (struct tm *tm)
 +{
 +intmonthlen[2][12] = {
 +   { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
 +   { 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
 +};
 +intyear, month, days;
 +
 +days = 365 * (tm-tm_year - 70);
 +for (year = 70; year  tm-tm_year; year++) {
 +   if (leapyear(1900 + year)) {
 +   days++;
 +   }
 +}
 +for (month = 0; month  tm-tm_mon; month++) {
 +   days += monthlen[leapyear(1900 + year)][month];
 +}
 +days += tm-tm_mday - 1;
 +
 +return days * 24) + tm-tm_hour) * 60 + tm-tm_min) * 60 +
tm-tm_sec);
 +}
 +
  /* Combine absolute and relative fields, and round. */
  static int
  create_output (struct state *state, time_t *t_out, const time_t *ref,
 @@ -1465,7 +1500,7 @@ create_output (struct state *state, time_t *t_out,
const time_t *ref,
  if (is_field_set (state, TM_TZ)) {
 /* tm is in specified TZ, convert to UTC for timegm(3). */
 tm.tm_min -= get_field (state, TM_TZ);
 -   t = timegm (tm);
 +   t = local_timegm (tm);
  } else {
 /* tm is in local time. */
 t = mktime (tm);
 --
 1.7.9.2

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

Re: [PATCH 10/10] timegm: add portable implementation (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Hi Jani,

 I'd prefer to use timegm() where available, and the suggested
 alternative [1] elsewhere.
 
 [1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html

 I considered this alternative, but decided against it because it's
 completely MT-unsafe.  I don't know whether libnotmuch itself is
 MT-safe, but a process which called this routine in one thread would
 temporarily throw off any timezone-related work that any other threads
 were doing, even if they weren't using libnotmuch.

That is a valid point. Yet it doesn't change the fact that I'd prefer to
use timegm() where available. Internally, glibc uses the same code to
implement both timegm() and mktime(), and I'd hate it if the results
were subtly different depending on whether the time zone was specified
in the input or not. That said, I'm not opposed to using your simple
timegm() alternative in the compat code if you think it's good enough to
get you going on Solaris.

As to solving the compat linking problem, I think the patch at the end
of this message should fix it. Please try that with the regular notmuch
approach to portability. The general idea is to keep parse-time-string
as independent as possible from the rest of notmuch (possibly turning it
into a dynamic library and a package of its own eventually), but I think
including compat.h is an acceptable exception to make.

HTH,
Jani.


diff --git a/parse-time-string/Makefile.local b/parse-time-string/Makefile.local
index 53534f3..b3e5385 100644
--- a/parse-time-string/Makefile.local
+++ b/parse-time-string/Makefile.local
@@ -1,7 +1,9 @@
 dir := parse-time-string
 extra_cflags += -I$(srcdir)/$(dir)
 
-libparse-time-string_c_srcs := $(dir)/parse-time-string.c
+libparse-time-string_c_srcs =  \
+   $(notmuch_compat_srcs)  \
+   $(dir)/parse-time-string.c
 
 libparse-time-string_modules := $(libparse-time-string_c_srcs:.c=.o)
 
diff --git a/parse-time-string/parse-time-string.c 
b/parse-time-string/parse-time-string.c
index 584067d3..ccad422 100644
--- a/parse-time-string/parse-time-string.c
+++ b/parse-time-string/parse-time-string.c
@@ -32,6 +32,7 @@
 #include sys/time.h
 #include sys/types.h
 
+#include compat.h
 #include parse-time-string.h
 
 /*
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Linux, FreeBSD, and Solaris all expect to find the prototype for
 index() in strings.h.  On some operating systems, including
 string.h is sufficient to get the prototype, but that's not the case
 on Solaris.  This patch just modifies notmuch-config.c to include
 strings.h to get the prototype.

We should probably just nuke index() and use strchr() instead.

Jani.


diff --git a/notmuch-config.c b/notmuch-config.c
index 3e37a2d..47eb743 100644
--- a/notmuch-config.c
+++ b/notmuch-config.c
@@ -688,7 +688,7 @@ _item_split (char *item, char **group, char **key)
 
 *group = item;
 
-period = index (item, '.');
+period = strchr (item, '.');
 if (period == NULL || *(period+1) == '\0') {
fprintf (stderr,
 Invalid configuration name: %s\n
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH 10/10] timegm: add portable implementation (Solaris support)

[Blake Jones]

Hi Jani,

 I'd prefer to use timegm() where available, and the suggested
 alternative [1] elsewhere.
 
 [1] http://www.kernel.org/doc/man-pages/online/pages/man3/timegm.3.html

I considered this alternative, but decided against it because it's
completely MT-unsafe.  I don't know whether libnotmuch itself is
MT-safe, but a process which called this routine in one thread would
temporarily throw off any timezone-related work that any other threads
were doing, even if they weren't using libnotmuch.

 I'll look into the compat build issues when I have a moment.

If you do, here's a boiled-down version of the problem that I came up
with while investigating it:

$ echo 'int main() { extern int foo1(); return foo1(); }'  main.c
$ echo 'int foo1() { extern int bar();  return bar();  }'  foo1.c
$ echo 'int bar()  { extern int foo2(); return foo2(); }'  bar.c
$ echo 'int foo2() { return 0; }'  foo2.c
$ gcc -c main.c foo1.c bar.c foo2.c
$ ar rcs libfoo.a foo1.o foo2.o
$ ar rcs libbar.a bar.o
$ gcc -o main main.o libfoo.a libbar.a
  [fails]
$ gcc -o main main.o libbar.a libfoo.a
  [fails]

Another alternative would be to just include parse-time-string.o in
libnotmuch.a directly; I think that would solve the problem.

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

Re: [PATCH 05/10] install: check for non-SysV version (Solaris support)

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Solaris ships a program called install in /usr/sbin, which performs a
 task that's fairly similar to the GNU and BSD install programs but
 which uses very different command line arguments.  In particular, if it
 is invoked without -c, -f, or -n, it will search the target
 directory for a file with the same name as the one being installed, and
 it will only install the file if it finds a matching name.  More
 excitingly, if it doesn't find a match, it will look in /bin, /usr/bin,
 /etc, /lib, and /usr/lib and try to do the same there.

 The standard workaround for this is to use GNU install.
 It is available via the standard Solaris packaging system (in
 file/gnu-coreutils), and installs itself as /usr/bin/ginstall.

 This patch adds a check to configure to see if install behaves in a
 way that's compatible with GNU and BSD install, and if not, it uses a
 program called ginstall instead.  It also modifies configure to set
 the $(INSTALL) variable, and changes various Makefiles to use it.
 ---
  Makefile.local|2 +-
  completion/Makefile.local |4 ++--
  configure |   19 +++
  emacs/Makefile.local  |6 +++---
  lib/Makefile.local|4 ++--
  man/Makefile.local|6 +++---
  vim/Makefile  |6 ++
  7 files changed, 32 insertions(+), 15 deletions(-)

 diff --git a/Makefile.local b/Makefile.local
 index 2b91946..7ccb1cd 100644
 --- a/Makefile.local
 +++ b/Makefile.local
 @@ -286,7 +286,7 @@ notmuch-shared: $(notmuch_client_modules) 
 lib/$(LINKER_NAME)
  .PHONY: install
  install: all install-man
   mkdir -p $(DESTDIR)$(prefix)/bin/
 - install notmuch-shared $(DESTDIR)$(prefix)/bin/notmuch
 + $(INSTALL) notmuch-shared $(DESTDIR)$(prefix)/bin/notmuch
  ifeq ($(MAKECMDGOALS), install)
   @echo 
   @echo Notmuch is now installed to $(DESTDIR)$(prefix)
 diff --git a/completion/Makefile.local b/completion/Makefile.local
 index dfc1271..a648a78 100644
 --- a/completion/Makefile.local
 +++ b/completion/Makefile.local
 @@ -14,9 +14,9 @@ install-$(dir):
   @echo $@
  ifeq ($(WITH_BASH),1)
   mkdir -p $(DESTDIR)$(bash_completion_dir)
 - install -m0644 $(bash_script) $(DESTDIR)$(bash_completion_dir)/notmuch
 + $(INSTALL) -m0644 $(bash_script) 
 $(DESTDIR)$(bash_completion_dir)/notmuch
  endif
  ifeq ($(WITH_ZSH),1)
   mkdir -p $(DESTDIR)$(zsh_completion_dir)
 - install -m0644 $(zsh_script) $(DESTDIR)$(zsh_completion_dir)/_notmuch
 + $(INSTALL) -m0644 $(zsh_script) 
 $(DESTDIR)$(zsh_completion_dir)/_notmuch
  endif
 diff --git a/configure b/configure
 index 5c5139f..dae837e 100755
 --- a/configure
 +++ b/configure
 @@ -591,6 +591,21 @@ for flag in -Wmissing-declarations; do
  done
  printf \n\t${WARN_CFLAGS}\n
  
 +INSTALL=install
 +printf Checking for working \install\ program... 
 +mkdir _tmp_

This doesn't feel like a hot idea. Don't tell me you'd need to create a
compatibility script for using mktemp --tmpdir too...

Or how about just always using ginstall on Solaris?

BR,
Jani.

 +cd _tmp_
 +echo 1  1
 +mkdir dest
 +if install 1 dest  /dev/null 21 ; then
 +  printf \install\ works fine.\n
 +else
 +  INSTALL=ginstall
 +  printf using \ginstall\.\n
 +fi
 +cd ..
 +rm -rf _tmp_
 +
  rm -f minimal minimal.c
  
  cat EOF
 @@ -777,4 +792,8 @@ CONFIGURE_CXXFLAGS = -DHAVE_GETLINE=\$(HAVE_GETLINE) 
 \$(GMIME_CFLAGS)\\
-DSTD_ASCTIME=\$(STD_ASCTIME)
  CONFIGURE_LDFLAGS =  \$(GMIME_LDFLAGS) \$(TALLOC_LDFLAGS) \$(XAPIAN_LDFLAGS) 
 \\
\$(LIBNSL_LDFLAGS)
 +
 +# Which install program to use
 +INSTALL = ${INSTALL}
 +
  EOF
 diff --git a/emacs/Makefile.local b/emacs/Makefile.local
 index fb82247..ee778cb 100644
 --- a/emacs/Makefile.local
 +++ b/emacs/Makefile.local
 @@ -36,11 +36,11 @@ endif
  .PHONY: install-emacs
  install-emacs:
   mkdir -p $(DESTDIR)$(emacslispdir)
 - install -m0644 $(emacs_sources) $(DESTDIR)$(emacslispdir)
 + $(INSTALL) -m0644 $(emacs_sources) $(DESTDIR)$(emacslispdir)
  ifeq ($(HAVE_EMACS),1)
 - install -m0644 $(emacs_bytecode) $(DESTDIR)$(emacslispdir)
 + $(INSTALL) -m0644 $(emacs_bytecode) $(DESTDIR)$(emacslispdir)
  endif
   mkdir -p $(DESTDIR)$(emacsetcdir)
 - install -m0644 $(emacs_images) $(DESTDIR)$(emacsetcdir)
 + $(INSTALL) -m0644 $(emacs_images) $(DESTDIR)$(emacsetcdir)
  
  CLEAN := $(CLEAN) $(emacs_bytecode)
 diff --git a/lib/Makefile.local b/lib/Makefile.local
 index 7785944..0c6b258 100644
 --- a/lib/Makefile.local
 +++ b/lib/Makefile.local
 @@ -89,11 +89,11 @@ install: install-$(dir)
  
  install-$(dir): $(dir)/$(LIBNAME)
   mkdir -p $(DESTDIR)$(libdir)/
 - install -m0644 $(lib)/$(LIBNAME) $(DESTDIR)$(libdir)/
 + $(INSTALL) -m0644 $(lib)/$(LIBNAME) $(DESTDIR)$(libdir)/
   ln -sf $(LIBNAME) $(DESTDIR)$(libdir)/$(SONAME)
   ln -sf $(LIBNAME) $(DESTDIR)$(libdir)/$(LINKER_NAME)
   mkdir -p 

Re: [PATCH 00/10] Solaris support

[Jani Nikula]

On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Hi all,

 This patch series fixes several issues which are needed to allow notmuch
 to build on Solaris 11.  I've been testing it for a month or so by
 using Karel Zak's fork of mutt along with a copy of notmuch-0.13.2 that
 I got to compile.  After a friend asked for a copy of my setup, I
 decided to try to get my patches cleaned up enough to submit, so that he
 wouldn't need to deal with them if he wanted to hack on it.

Thanks for doing this. I had a *very* brief look at the patches, and
commented on what I thought stood out. Otherwise I didn't spot anything
out of the ordinary, but I'd like it if someone else had a glance too.

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

Re: [PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Tomi Ollila]

On Sun, Nov 04 2012, Jani Nikula wrote:

 On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Linux, FreeBSD, and Solaris all expect to find the prototype for
 index() in strings.h.  On some operating systems, including
 string.h is sufficient to get the prototype, but that's not the case
 on Solaris.  This patch just modifies notmuch-config.c to include
 strings.h to get the prototype.

 We should probably just nuke index() and use strchr() instead.

indeed!


 Jani.


Tomi



 diff --git a/notmuch-config.c b/notmuch-config.c
 index 3e37a2d..47eb743 100644
 --- a/notmuch-config.c
 +++ b/notmuch-config.c
 @@ -688,7 +688,7 @@ _item_split (char *item, char **group, char **key)
  
  *group = item;
  
 -period = index (item, '.');
 +period = strchr (item, '.');
  if (period == NULL || *(period+1) == '\0') {
   fprintf (stderr,
Invalid configuration name: %s\n
 ___
 notmuch mailing list
 notmuch@notmuchmail.org
 http://notmuchmail.org/mailman/listinfo/notmuch
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

Re: Automatic suppression of non-duplicate messages

[Jani Nikula]

On Sat, 03 Nov 2012, David Bremner da...@tethera.net wrote:
 Eirik Byrkjeflot Anonsen ei...@eirikba.org writes:

 That's not what I see.  If I search for a term that only appears in
 one of the copies, none of the copies are included in the search
 result.

 The offending code is at line 1813 of lib/database.cc; the message is
 only indexed if the message-id is new.

 It might be sensible to move _notmuch_message_index_file into the other
 branch of the if, but even if that works fine, something more
 sophisticated is needed for the call to
 __notmuch_message_set_header_values; the invariant that each message has
 a single subject seems reasonable.

 Offhand I'm not sure of a good method of automatically deciding what is
 the same message (with e.g. headers and footer text added by a mailing
 list).

Assuming there was good method, what would you do with two different
messages that have the same message id? That is the unique id we use to
identify messages (which should be fine per RFC 5322 and its
predecessors; we're talking about messages from broken systems here).

It might be helpful to have a configuration option similar to new.tags
that would define the tags to be assigned to messages with duplicate
message ids. (This could be done in the
NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID case near line 516 of
notmuch-new.c). This could be used to assign a dupe tag, for example,
so the user could do whatever they want in the post-new hook or the user
interface. A sufficiently clever post-new hook could compare the files
of a message, and drop the tag or add another, as the case may
be. Surely not a perfect solution, but keeps the implementation simple.


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

Re: Automatic suppression of non-duplicate messages

[Austin Clements]

Quoth Jani Nikula on Nov 05 at 12:34 am:
 On Sat, 03 Nov 2012, David Bremner da...@tethera.net wrote:
  Eirik Byrkjeflot Anonsen ei...@eirikba.org writes:
 
  That's not what I see.  If I search for a term that only appears in
  one of the copies, none of the copies are included in the search
  result.
 
  The offending code is at line 1813 of lib/database.cc; the message is
  only indexed if the message-id is new.
 
  It might be sensible to move _notmuch_message_index_file into the other
  branch of the if, but even if that works fine, something more
  sophisticated is needed for the call to
  __notmuch_message_set_header_values; the invariant that each message has
  a single subject seems reasonable.
 
  Offhand I'm not sure of a good method of automatically deciding what is
  the same message (with e.g. headers and footer text added by a mailing
  list).
 
 Assuming there was good method, what would you do with two different
 messages that have the same message id? That is the unique id we use to
 identify messages (which should be fine per RFC 5322 and its
 predecessors; we're talking about messages from broken systems here).
 
 It might be helpful to have a configuration option similar to new.tags
 that would define the tags to be assigned to messages with duplicate
 message ids. (This could be done in the
 NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID case near line 516 of
 notmuch-new.c). This could be used to assign a dupe tag, for example,
 so the user could do whatever they want in the post-new hook or the user
 interface. A sufficiently clever post-new hook could compare the files
 of a message, and drop the tag or add another, as the case may
 be. Surely not a perfect solution, but keeps the implementation simple.

This would also trigger on message flag changes and folder moves
performed outside of notmuch, since notmuch sees those as a duplicate
message ID followed by a deletion.  The only way to do something for
every received message even if it has the same message ID as an
existing message is to do it in whatever delivers mail.  Currently, we
don't have a good story for integrating on-delivery operations with
notmuch.
___
notmuch mailing list
notmuch@notmuchmail.org
http://notmuchmail.org/mailman/listinfo/notmuch

Re: [PATCH 10/10] timegm: add portable implementation (Solaris support)

[Blake Jones]

 That is a valid point. Yet it doesn't change the fact that I'd prefer
 to use timegm() where available. Internally, glibc uses the same code
 to implement both timegm() and mktime(), and I'd hate it if the
 results were subtly different depending on whether the time zone was
 specified in the input or not.

That's fine with me.

 That said, I'm not opposed to using your simple timegm() alternative
 in the compat code if you think it's good enough to get you going on
 Solaris.

I think it is, assuming you don't plan to use tm_wday or tm_yday in your
parse-time-string code, and that you don't plan to depend on the side
effect of timegm() canonicalizing the passed-in struct tm.

 As to solving the compat linking problem, I think the patch at the end
 of this message should fix it. Please try that with the regular
 notmuch approach to portability. The general idea is to keep
 parse-time-string as independent as possible from the rest of notmuch
 (possibly turning it into a dynamic library and a package of its own
 eventually), but I think including compat.h is an acceptable exception
 to make.

Yeah, this seems to work.  I'll update my patch set accordingly.

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

Re: [PATCH 08/10] notmuch-config: header for index() prototype (Solaris support)

[Blake Jones]

 On Sun, 04 Nov 2012, Blake Jones bla...@foo.net wrote:
 Linux, FreeBSD, and Solaris all expect to find the prototype for
 index() in strings.h.  On some operating systems, including
 string.h is sufficient to get the prototype, but that's not the case
 on Solaris.  This patch just modifies notmuch-config.c to include
 strings.h to get the prototype.

 We should probably just nuke index() and use strchr() instead.
 
 indeed!

That was my initial preference, but I didn't know if there was anyone
who was committed to the BSD name.  Given that two more people think
it's a good idea, I'll do that instead.

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

notmuch search and tags

[Aneesh Kumar K.V]

We have the mails appearing as below now. Earlier the tags towards the
end used be enclosed in ( )

[21/21]   Today 10:57Blake Jones, Jani Nikula, Tomi Ollila[PATCH 05/10] 
install: check for non-SysV version (Solaris support) notmuch unread
[5/5] Today 09:58Eirik Byrkjeflot Anonsen, David Bremner, Jani Nikula, 
Austin Clements Automatic suppression of non-duplicate 
messagesnotmuch unread


-aneesh

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