sup

community/pipermail-archives/sup-devel/2010-07.txt (81417B) - raw

 From sascha-pgp@silbe.org  Thu Jul  1 09:48:13 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Thu,  1 Jul 2010 13:48:13 +0000
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 Message-ID: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 
 If Message-ID contains only invalid characters (or none at all),
 sanitize_message_id will return an empty string. As Message.parse_header
 only checked for header existance, not validity, sup-dump could produce
 a dump file that sup-sync would choke on.
 
 We now fall back to a fake id for both missing and invalid Message-ID
 headers.
 ---
  lib/sup/message.rb |   13 +++++++------
  1 files changed, 7 insertions(+), 6 deletions(-)
 
 diff --git a/lib/sup/message.rb b/lib/sup/message.rb
 index a13fc0c..f273988 100644
 --- a/lib/sup/message.rb
 +++ b/lib/sup/message.rb
 @@ -77,14 +77,15 @@ class Message
    def parse_header encoded_header
      header = SavingHash.new { |k| decode_header_field encoded_header[k] }
  
 -    @id = if header["message-id"]
 +    @id = ''
 +    if header["message-id"]
        mid = header["message-id"] =~ /<(.+?)>/ ? $1 : header["message-id"]
 -      sanitize_message_id mid
 -    else
 -      id = "sup-faked-" + Digest::MD5.hexdigest(raw_header)
 -      from = header["from"]
 +      @id = sanitize_message_id mid
 +    end
 +    if (not @id.include? '@') || @id.length < 6
 +      @id = "sup-faked-" + Digest::MD5.hexdigest(raw_header)
 +      #from = header["from"]
        #debug "faking non-existent message-id for message from #{from}: #{id}"
 -      id
      end
  
      @from = Person.from_address(if header["from"]
 -- 
 1.7.1
 
 
 From bwalton@artsci.utoronto.ca  Thu Jul  1 10:33:22 2010
 From: bwalton@artsci.utoronto.ca (Ben Walton)
 Date: Thu, 01 Jul 2010 10:33:22 -0400
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1277994760-sup-4189@pinkfloyd.chass.utoronto.ca>
 
 Excerpts from Sascha Silbe's message of Thu Jul 01 09:48:13 -0400 2010:
 
 Hi Sascha,
 
 > +      #from = header["from"]
 
 This patch looks good, but why include the commented line?
 
 Thanks
 -Ben
 --
 Ben Walton
 Systems Programmer - CHASS
 University of Toronto
 C:416.407.5610 | W:416.978.4302
 
 
 From sascha-ml-email-sup-devel@silbe.org  Thu Jul  1 15:34:47 2010
 From: sascha-ml-email-sup-devel@silbe.org (Sascha Silbe)
 Date: Thu, 01 Jul 2010 19:34:47 +0000
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1277994760-sup-4189@pinkfloyd.chass.utoronto.ca>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 	<1277994760-sup-4189@pinkfloyd.chass.utoronto.ca>
 Message-ID: <1278012648-sup-257@twin.sascha.silbe.org>
 
 Excerpts from Ben Walton's message of Thu Jul 01 14:33:22 +0000 2010:
 
 > > +      #from = header["from"]
 > 
 > This patch looks good, but why include the commented line?
 I just commented out a line that was only useful in combination with the
 line directly below it that was already commented out. If you prefer
 I can remove both of them instead (or do the change in a separate patch).
 
 Sascha
 
 --
 http://sascha.silbe.org/
 http://www.infra-silbe.de/
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 490 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100701/2157f666/attachment.bin>
 
 From bwalton@artsci.utoronto.ca  Thu Jul  1 16:11:08 2010
 From: bwalton@artsci.utoronto.ca (Ben Walton)
 Date: Thu, 01 Jul 2010 16:11:08 -0400
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1278012648-sup-257@twin.sascha.silbe.org>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 	<1277994760-sup-4189@pinkfloyd.chass.utoronto.ca>
 	<1278012648-sup-257@twin.sascha.silbe.org>
 Message-ID: <1278015019-sup-7327@pinkfloyd.chass.utoronto.ca>
 
 Excerpts from Sascha Silbe's message of Thu Jul 01 15:34:47 -0400 2010:
 
 > I just commented out a line that was only useful in combination with
 > the line directly below it that was already commented out. If you
 > prefer I can remove both of them instead (or do the change in a
 > separate patch).
 
 I'd resubmit with the lines removed.  It keeps the history cleaner.
 Up to you of course, it's just my own preference I'm putting forward.
 
 Thanks
 -Ben
 --
 Ben Walton
 Systems Programmer - CHASS
 University of Toronto
 C:416.407.5610 | W:416.978.4302
 
 
 From cam+sup@cammunism.org  Thu Jul  1 17:32:37 2010
 From: cam+sup@cammunism.org (Cameron Matheson)
 Date: Thu, 01 Jul 2010 15:32:37 -0600
 Subject: [sup-devel] =?utf-8?b?W1BBVENIXSBEb27igJl0IGRpc3BsYXkgIi4uLiIg?=
 	=?utf-8?q?after_snippets_which_are_displayed_completely?=
 In-Reply-To: <1276711781-sup-4102@midna.zekjur.net>
 References: <1276711781-sup-4102@midna.zekjur.net>
 Message-ID: <1278019896-sup-9354@cammunism.org>
 
 Excerpts from Michael Stapelberg's message of Wed Jun 16 12:11:47 -0600 2010:
 > quote from the commit message:
 >     Don?t display "..." after snippets which are displayed completely
 >     
 >     Short mails (for example: "Yes, the date works for me.") often can
 >     be displayed completely in the snippet. However, before this patch,
 >     sup abbreviated the snippet even though it was not abbreviated.
 
 I haven't seen any one else comment on this yet, but this is a nagging
 issue that I have meant to fix for ages.  Thanks!
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:31:32 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:31:32 -0400
 Subject: [sup-devel] [PATCH] fix reference to EncodingUnsupportedError
 In-Reply-To: <1277797825-8181-1-git-send-email-sascha-pgp@silbe.org>
 References: <1277797825-8181-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278123679-sup-5970@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:31:53 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:31:53 -0400
 Subject: [sup-devel] [PATCH] Don't choke when scanning message with
 	unknown encoding
 In-Reply-To: <1277798694-8642-1-git-send-email-sascha-pgp@silbe.org>
 References: <1277798694-8642-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278123779-sup-6393@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:31:59 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:31:59 -0400
 Subject: [sup-devel] [PATCH] fix crash in sup-dump if the default sent
 	source is used
 In-Reply-To: <1277799125-8696-1-git-send-email-sascha-pgp@silbe.org>
 References: <1277799125-8696-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278123805-sup-3778@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:32:23 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:32:23 -0400
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278124230-sup-174@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:33:21 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:33:21 -0400
 Subject: [sup-devel] =?utf-8?b?W1BBVENIXSBEb27igJl0IGRpc3BsYXkgIi4uLiIg?=
 	=?utf-8?q?after_snippets_which_are_displayed_completely?=
 In-Reply-To: <1276711781-sup-4102@midna.zekjur.net>
 References: <1276711781-sup-4102@midna.zekjur.net>
 Message-ID: <1278124387-sup-5603@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Fri Jul  2 22:33:48 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Fri, 02 Jul 2010 22:33:48 -0400
 Subject: [sup-devel] [PATCH] inline-gpg: call text_to_chunks on the text
 	before/after the GPG part
 In-Reply-To: <1277221190-sup-946@midna.zekjur.net>
 References: <1277221190-sup-946@midna.zekjur.net>
 Message-ID: <1278124417-sup-4659@zyrg.net>
 
 Applied to master.
 
 From sascha-ml-email-sup-devel@silbe.org  Sat Jul  3 06:35:50 2010
 From: sascha-ml-email-sup-devel@silbe.org (Sascha Silbe)
 Date: Sat, 03 Jul 2010 10:35:50 +0000
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1278124230-sup-174@zyrg.net>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 	<1278124230-sup-174@zyrg.net>
 Message-ID: <1278153268-sup-6159@twin.sascha.silbe.org>
 
 [several patches from this and other threads]
 > Applied to master.
 Thanks!
 
 Sascha
 
 --
 http://sascha.silbe.org/
 http://www.infra-silbe.de/
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 490 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100703/14dfde2c/attachment.bin>
 
 From sascha-ml-email-sup-devel@silbe.org  Sat Jul  3 06:39:21 2010
 From: sascha-ml-email-sup-devel@silbe.org (Sascha Silbe)
 Date: Sat, 03 Jul 2010 10:39:21 +0000
 Subject: [sup-devel] [PATCH] parse_header: don't use empty or invalid
 	Message-ID header
 In-Reply-To: <1278015019-sup-7327@pinkfloyd.chass.utoronto.ca>
 References: <1277992093-10600-1-git-send-email-sascha-pgp@silbe.org>
 	<1277994760-sup-4189@pinkfloyd.chass.utoronto.ca>
 	<1278012648-sup-257@twin.sascha.silbe.org>
 	<1278015019-sup-7327@pinkfloyd.chass.utoronto.ca>
 Message-ID: <1278153353-sup-1102@twin.sascha.silbe.org>
 
 Excerpts from Ben Walton's message of Thu Jul 01 20:11:08 +0000 2010:
 
 > I'd resubmit with the lines removed.  It keeps the history cleaner.
 Since the patch has already been merged in its current form, I won't post
 a new one. It's no change in program behaviour or output and the comments
 are not incorrect. Maybe they even help someone while debugging one day. ;)
 
 Sascha
 
 --
 http://sascha.silbe.org/
 http://www.infra-silbe.de/
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 490 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100703/3531f528/attachment.bin>
 
 From sascha-pgp@silbe.org  Sat Jul  3 07:06:22 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Sat,  3 Jul 2010 13:06:22 +0200
 Subject: [sup-devel] [PATCH] don't leak fds for mbox sources
 Message-ID: <1278155182-8545-1-git-send-email-sascha-pgp@silbe.org>
 
 Only open the mbox when it's actually used and close it again once we're
 finished iterating over it.
 
 Adds a method go_idle to Source to inform the source we're unlikely to
 use it in the near future so it can do internal clean-up as appropriate.
 
 Signed-off-by: Sascha Silbe <sascha-pgp at silbe.org>
 ---
  lib/sup/mbox.rb   |   28 ++++++++++++++++++++++++++--
  lib/sup/poll.rb   |    7 ++++++-
  lib/sup/source.rb |    6 ++++++
  3 files changed, 38 insertions(+), 3 deletions(-)
 
 
 This is my first non-trivial patch to a Ruby project. Please mention
 even minor style mistakes.
 
 
 diff --git a/lib/sup/mbox.rb b/lib/sup/mbox.rb
 index 3f7ae3e..2bf32ea 100644
 --- a/lib/sup/mbox.rb
 +++ b/lib/sup/mbox.rb
 @@ -23,7 +23,7 @@ class MBox < Source
        raise ArgumentError, "not an mbox uri" unless uri.scheme == "mbox"
        raise ArgumentError, "mbox URI ('#{uri}') cannot have a host: #{uri.host}" if uri.host
        raise ArgumentError, "mbox URI must have a path component" unless uri.path
 -      @f = File.open uri.path, 'rb'
 +      @f = nil
        @path = uri.path
      else
        @f = uri_or_fp
 @@ -53,12 +53,31 @@ class MBox < Source
      end
    end
  
 +  def ensure_open
 +    @f = File.open @path, 'rb' if @f.nil?
 +  end
 +  private :ensure_open
 +
 +  def go_idle
 +    @mutex.synchronize do
 +      return if @f.nil? || @path.nil?
 +      @f.close
 +      @f = nil
 +    end
 +  end
 +
    def start_offset; 0; end
 -  def end_offset; File.size @f; end
 +  def end_offset
 +    @mutex.synchronize do
 +      ensure_open
 +      File.size @f
 +    end
 +  end
  
    def load_header offset
      header = nil
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        l = @f.gets
        unless MBox::is_break_line? l
 @@ -71,6 +90,7 @@ class MBox < Source
  
    def load_message offset
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        begin
          ## don't use RMail::Mailbox::MBoxReader because it doesn't properly ignore
 @@ -88,6 +108,7 @@ class MBox < Source
    ## scan forward until we're at the valid start of a message
    def correct_offset!
      @mutex.synchronize do
 +      ensure_open
        @f.seek cur_offset
        string = ""
        until @f.eof? || MBox::is_break_line?(l = @f.gets)
 @@ -100,6 +121,7 @@ class MBox < Source
    def raw_header offset
      ret = ""
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        until @f.eof? || (l = @f.gets) =~ /^\r*$/
          ret << l
 @@ -131,6 +153,7 @@ class MBox < Source
    ## sup-sync-back has to do it.
    def each_raw_message_line offset
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        yield @f.gets
        until @f.eof? || MBox::is_break_line?(l = @f.gets)
 @@ -145,6 +168,7 @@ class MBox < Source
  
      begin
        @mutex.synchronize do
 +        ensure_open
          @f.seek cur_offset
  
          ## cur_offset could be at one of two places here:
 diff --git a/lib/sup/poll.rb b/lib/sup/poll.rb
 index 49e456b..3ce867c 100644
 --- a/lib/sup/poll.rb
 +++ b/lib/sup/poll.rb
 @@ -159,7 +159,10 @@ EOS
    ## this is the primary mechanism for iterating over messages from a source.
    def each_message_from source, opts={}
      begin
 -      return if source.done? || source.has_errors?
 +      if source.done? || source.has_errors?
 +        source.go_idle
 +        return
 +      end
  
        source.each do |offset, source_labels|
          if source.has_errors?
 @@ -175,6 +178,8 @@ EOS
          HookManager.run "before-add-message", :message => m
          yield m
        end
 +
 +      source.go_idle
      rescue SourceError => e
        warn "problem getting messages from #{source}: #{e.message}"
        Redwood::report_broken_sources :force_to_top => true
 diff --git a/lib/sup/source.rb b/lib/sup/source.rb
 index 2f2e5df..13dc443 100644
 --- a/lib/sup/source.rb
 +++ b/lib/sup/source.rb
 @@ -40,6 +40,7 @@ class Source
    ## - raw_header offset
    ## - raw_message offset
    ## - check (optional)
 +  ## - go_idle (optional)
    ## - next (or each, if you prefer): should return a message and an
    ##   array of labels.
    ##
 @@ -94,6 +95,11 @@ class Source
    ## to proactively notify the user of any source problems.
    def check; end
  
 +  ## release resources that are easy to reacquire. it is called
 +  ## after processing a source (e.g. polling) to prevent resource
 +  ## leaks (esp. file descriptors).
 +  def go_idle; end
 +
    ## yields successive offsets and labels, starting at #cur_offset.
    ##
    ## when implementing a source, you can overwrite either #each or #next. the
 -- 
 1.7.1
 
 
 From damien.leone@fensalir.fr  Sat Jul  3 13:22:42 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Sat, 03 Jul 2010 19:22:42 +0200
 Subject: [sup-devel] [PATCHES] reply-mode,
 	signatures and account selector
 In-Reply-To: <1276089233-sup-510@mailer>
 References: <1276087951-sup-7816@mailer> <1276089233-sup-510@mailer>
 Message-ID: <1278177699-sup-5324@mailer>
 
 Any reviewer for these patches?
 
 Thanks.
 
 Excerpts from Damien Leone's message of mer. juin 09 15:15:11 +0200 2010:
 > I noticed an error in the third patch.
 > Please consider the one attached to *this* email.
 > 
 > Regards,
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 
 From sascha-pgp@silbe.org  Thu Jul  8 08:28:07 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Thu,  8 Jul 2010 12:28:07 +0000
 Subject: [sup-devel] [PATCH] mbox: fix date format
 Message-ID: <1278592087-614-1-git-send-email-sascha-pgp@silbe.org>
 
 The date format used in all (known/common) variations is that of Unix/POSIX
 asctime(), not the RFC 822 / RFC 2822 Date: header format.
 
 Using an incorrect date format causes at least mutt, mb2md and Python
 UnixMailbox (but not PortableUnixMailbox) to either reject the mailbox or show
 it as empty.
 
 References:
 http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/mail-mbox-formats.html
 http://www.qmail.org/man/man5/mbox.html
 
 Signed-off-by: Sascha Silbe <sascha-pgp at silbe.org>
 ---
  lib/sup/mbox.rb |    2 +-
  1 files changed, 1 insertions(+), 1 deletions(-)
 
 diff --git a/lib/sup/mbox.rb b/lib/sup/mbox.rb
 index 2bf32ea..452519d 100644
 --- a/lib/sup/mbox.rb
 +++ b/lib/sup/mbox.rb
 @@ -140,7 +140,7 @@ class MBox < Source
      need_blank = File.exists?(@filename) && !File.zero?(@filename)
      File.open(@filename, "ab") do |f|
        f.puts if need_blank
 -      f.puts "From #{from_email} #{date.rfc2822}"
 +      f.puts "From #{from_email} #{date.asctime}"
        yield f
      end
    end
 -- 
 1.7.1
 
 
 From sascha-pgp@silbe.org  Thu Jul  8 08:28:31 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Thu,  8 Jul 2010 12:28:31 +0000
 Subject: [sup-devel] [PATCH] fix crash in sup-sync if the default sent
 	source is used
 Message-ID: <1278592111-654-1-git-send-email-sascha-pgp@silbe.org>
 
 This fixes a crash in sup-sync when syncing a folder that contains a copy of a
 "sent" message and no "sent" folder has been explicitly configured in the
 config file (so it hasn't been added to sources.yaml).
 
 Signed-off-by: Sascha Silbe <sascha-pgp at silbe.org>
 ---
  bin/sup-sync |    6 ++++++
  1 files changed, 6 insertions(+), 0 deletions(-)
 
 diff --git a/bin/sup-sync b/bin/sup-sync
 index 6eb8f80..7fb2a87 100755
 --- a/bin/sup-sync
 +++ b/bin/sup-sync
 @@ -116,6 +116,12 @@ index.lock_interactively or exit
  begin
    index.load
  
 +  if(s = Redwood::SourceManager.source_for Redwood::SentManager.source_uri)
 +    Redwood::SentManager.source = s
 +  else
 +    Redwood::SourceManager.add_source Redwood::SentManager.default_source
 +  end
 +
    sources = if opts[:all_sources]
      Redwood::SourceManager.sources
    elsif ARGV.empty?
 -- 
 1.7.1
 
 
 From sascha-pgp@silbe.org  Thu Jul  8 11:26:37 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Thu,  8 Jul 2010 15:26:37 +0000
 Subject: [sup-devel] [PATCH] fix crash during poll if Maildir has been
 	emptied
 Message-ID: <1278602797-2553-1-git-send-email-sascha-pgp@silbe.org>
 
 We need to do the rescan in start_offset instead of end_offset so that
 end_offset won't be called if the Maildir is empty after the rescan.
 
 Fixes this crash:
 
 NoMethodError from thread: user-invoked poll
 undefined method `+' for nil:NilClass
 /home/sascha.silbe/src/sup/lib/sup/maildir.rb:171:in `end_offset'
 /home/sascha.silbe/src/sup/lib/sup/source.rb:90:in `done?'
 [...]
 
 Signed-off-by: Sascha Silbe <sascha-pgp at silbe.org>
 ---
  lib/sup/maildir.rb |    4 ++--
  1 files changed, 2 insertions(+), 2 deletions(-)
 
 diff --git a/lib/sup/maildir.rb b/lib/sup/maildir.rb
 index ef3b318..0b2fd5e 100644
 --- a/lib/sup/maildir.rb
 +++ b/lib/sup/maildir.rb
 @@ -162,12 +162,12 @@ class Maildir < Source
    end
  
    def start_offset
 -    scan_mailbox
 +    scan_mailbox :rescan => true
      @ids.first
    end
  
    def end_offset
 -    scan_mailbox :rescan => true
 +    scan_mailbox
      @ids.last + 1
    end
  
 -- 
 1.7.1
 
 
 From sascha-pgp@silbe.org  Thu Jul  8 13:15:22 2010
 From: sascha-pgp@silbe.org (Sascha Silbe)
 Date: Thu,  8 Jul 2010 17:15:22 +0000
 Subject: [sup-devel] [PATCH] don't leak fds for mbox sources
 Message-ID: <1278609322-1401-1-git-send-email-sascha-pgp@silbe.org>
 
 Only open the mbox when it's actually used and close it again once we're
 finished iterating over it.
 
 Adds a method go_idle to Source to inform the source we're unlikely to
 use it in the near future so it can do internal clean-up as appropriate.
 
 Signed-off-by: Sascha Silbe <sascha-pgp at silbe.org>
 ---
  lib/sup/mbox.rb   |   28 ++++++++++++++++++++++++++--
  lib/sup/poll.rb   |    7 ++++++-
  lib/sup/source.rb |    6 ++++++
  3 files changed, 38 insertions(+), 3 deletions(-)
 
 This patch allows configuring more than RLIMIT_NOFILE mbox sources. I have
 been testing it for several days.
 
 Please do point out even minor style mistakes.
 
 
 diff --git a/lib/sup/mbox.rb b/lib/sup/mbox.rb
 index 3f7ae3e..2bf32ea 100644
 --- a/lib/sup/mbox.rb
 +++ b/lib/sup/mbox.rb
 @@ -23,7 +23,7 @@ class MBox < Source
        raise ArgumentError, "not an mbox uri" unless uri.scheme == "mbox"
        raise ArgumentError, "mbox URI ('#{uri}') cannot have a host: #{uri.host}" if uri.host
        raise ArgumentError, "mbox URI must have a path component" unless uri.path
 -      @f = File.open uri.path, 'rb'
 +      @f = nil
        @path = uri.path
      else
        @f = uri_or_fp
 @@ -53,12 +53,31 @@ class MBox < Source
      end
    end
  
 +  def ensure_open
 +    @f = File.open @path, 'rb' if @f.nil?
 +  end
 +  private :ensure_open
 +
 +  def go_idle
 +    @mutex.synchronize do
 +      return if @f.nil? || @path.nil?
 +      @f.close
 +      @f = nil
 +    end
 +  end
 +
    def start_offset; 0; end
 -  def end_offset; File.size @f; end
 +  def end_offset
 +    @mutex.synchronize do
 +      ensure_open
 +      File.size @f
 +    end
 +  end
  
    def load_header offset
      header = nil
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        l = @f.gets
        unless MBox::is_break_line? l
 @@ -71,6 +90,7 @@ class MBox < Source
  
    def load_message offset
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        begin
          ## don't use RMail::Mailbox::MBoxReader because it doesn't properly ignore
 @@ -88,6 +108,7 @@ class MBox < Source
    ## scan forward until we're at the valid start of a message
    def correct_offset!
      @mutex.synchronize do
 +      ensure_open
        @f.seek cur_offset
        string = ""
        until @f.eof? || MBox::is_break_line?(l = @f.gets)
 @@ -100,6 +121,7 @@ class MBox < Source
    def raw_header offset
      ret = ""
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        until @f.eof? || (l = @f.gets) =~ /^\r*$/
          ret << l
 @@ -131,6 +153,7 @@ class MBox < Source
    ## sup-sync-back has to do it.
    def each_raw_message_line offset
      @mutex.synchronize do
 +      ensure_open
        @f.seek offset
        yield @f.gets
        until @f.eof? || MBox::is_break_line?(l = @f.gets)
 @@ -145,6 +168,7 @@ class MBox < Source
  
      begin
        @mutex.synchronize do
 +        ensure_open
          @f.seek cur_offset
  
          ## cur_offset could be at one of two places here:
 diff --git a/lib/sup/poll.rb b/lib/sup/poll.rb
 index 49e456b..3ce867c 100644
 --- a/lib/sup/poll.rb
 +++ b/lib/sup/poll.rb
 @@ -159,7 +159,10 @@ EOS
    ## this is the primary mechanism for iterating over messages from a source.
    def each_message_from source, opts={}
      begin
 -      return if source.done? || source.has_errors?
 +      if source.done? || source.has_errors?
 +        source.go_idle
 +        return
 +      end
  
        source.each do |offset, source_labels|
          if source.has_errors?
 @@ -175,6 +178,8 @@ EOS
          HookManager.run "before-add-message", :message => m
          yield m
        end
 +
 +      source.go_idle
      rescue SourceError => e
        warn "problem getting messages from #{source}: #{e.message}"
        Redwood::report_broken_sources :force_to_top => true
 diff --git a/lib/sup/source.rb b/lib/sup/source.rb
 index 2f2e5df..13dc443 100644
 --- a/lib/sup/source.rb
 +++ b/lib/sup/source.rb
 @@ -40,6 +40,7 @@ class Source
    ## - raw_header offset
    ## - raw_message offset
    ## - check (optional)
 +  ## - go_idle (optional)
    ## - next (or each, if you prefer): should return a message and an
    ##   array of labels.
    ##
 @@ -94,6 +95,11 @@ class Source
    ## to proactively notify the user of any source problems.
    def check; end
  
 +  ## release resources that are easy to reacquire. it is called
 +  ## after processing a source (e.g. polling) to prevent resource
 +  ## leaks (esp. file descriptors).
 +  def go_idle; end
 +
    ## yields successive offsets and labels, starting at #cur_offset.
    ##
    ## when implementing a source, you can overwrite either #each or #next. the
 -- 
 1.7.1
 
 
 From damien.leone@fensalir.fr  Fri Jul  9 12:44:01 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Fri, 09 Jul 2010 18:44:01 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 Message-ID: <1278693172-sup-6247@mailer>
 
 Sup guys,
 
 I have been working on backward synchronization for Maildir sources
 last week and I need your help to test and review the code.
 
 I pushed my commits (based on the 'next' branch) to my repo which you
 can find here [0] for the web interface and here [1] for the git
 address.
 
 The following is related to Maildir sources ONLY.
 
 So what does work and what does not (if you don't want to read this,
 scroll to the quick howto below):
 
 - Sup will now stay synchronized with your remote sources by detecting
 if a message has been remotely updated (ie: you change a flag from
 another client), these updates will be applied to your local index and
 your thread-view-mode buffers should be properly refreshed;
 
 - When a message is remotely deleted, it should now disappear from
 your index;
 
 - There is a new configuration option called "sync_back_to_maildir"
 which is false by default. When true, this option updates your actual
 Maildir files when you change a label, this is done in real time
 before being saved to xapian. So for instance if you use offlineimap
 the changes should be synched back to your IMAP server by the next
 poll in Sup (assuming that your before-poll hook runs offlineimap);
 
 - I wrote a "sup-sync-back-maildir" script that will synchronize all
 messages from your sources at once. This should be executed the first
 time BEFORE any polling, otherwise your IMAP server will be synched to
 your Sup index and if you did not used another client to mark your
 emails as read etc. (like me) and to update your IMAP server then you
 will lose all what you did in Sup (labels, etc.);
 
 - I added two new hidden labels in Sup: 'replied' and 'forwarded',
 they are automatically added when replying, forwarding or bouncing a
 message, this is to bring a better Maildir support to Sup, this should
 be invisible for users;
 
 - However you will probably lose all your remote 'replied' and
 'forwarded' flags after your first backward synchronization to Maildir
 since Sup dropped these labels when it added your messages to xapian.
 
 - Moving a message from a Maildir source to another is not (yet?)
 supported.
 
 So, you should now be able to use multiple clients to handle your
 emails, for instance I can use gmail to mark messages as starred or
 so, it will be reflected in Sup. In the same way, if I mark a message
 as read in Sup it will appear as read in the gmail web interface.
 
 I tested this code with my main account (2 GB and like 15 Maildir
 sources) it works fine so far.
 
 Be warned that it is highly experimental, but if you use offlineimap
 it has a realdelete option so you *should* not lose any email, in the
 worst case it might mess your flags/labels up.
 
 I tested it only with offlineimap working with a gmail account.
 
 Quick howto:
 
 1. Close sup
 2. Backup your emails and your xapian index or use another Sup session
 3. Clone the branch [1]
 4. IMPORTANT: run "bin/sup-sync-back-maildir" to synchronize the
    Maildirs you wish, check the help
 5. Add ":sync_back_to_maildir: true" to your config.yaml
 6. Run and use sup
 
 Please test it and make reviews! :)
 
 NOTE: if you want this to work, get sure that you are not calling
 offlineimap with the '-q' option (as suggested in the wiki) otherwise
 it will ignore the flag updates on IMAP server, so check your
 before-poll hook.
 
 [0] http://git.fensalir.fr/?p=dleone/sup.git
 [1] $ git clone git://fensalir.fr/dleone/sup.git -b maildir-sync
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100709/775ef421/attachment.bin>
 
 From ezyang@MIT.EDU  Fri Jul  9 12:56:31 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Fri, 09 Jul 2010 12:56:31 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1278694430-sup-914@ezyang>
 
 Hey Damien,
 
 This is great!  A few quick questions for you:
 
 1. Suppose we decide to test your branch, and run into some show-stopper
    bugs.  What?s the procedure for backing out changes?
 
 2. What precisely are the semantics for handling Sup labels?
 
 Thanks,
 Edward
 
 Excerpts from Damien Leone's message of Fri Jul 09 12:44:01 -0400 2010:
 > Sup guys,
 > 
 > I have been working on backward synchronization for Maildir sources
 > last week and I need your help to test and review the code.
 > 
 > I pushed my commits (based on the 'next' branch) to my repo which you
 > can find here [0] for the web interface and here [1] for the git
 > address.
 > 
 > The following is related to Maildir sources ONLY.
 > 
 > So what does work and what does not (if you don't want to read this,
 > scroll to the quick howto below):
 > 
 > - Sup will now stay synchronized with your remote sources by detecting
 > if a message has been remotely updated (ie: you change a flag from
 > another client), these updates will be applied to your local index and
 > your thread-view-mode buffers should be properly refreshed;
 > 
 > - When a message is remotely deleted, it should now disappear from
 > your index;
 > 
 > - There is a new configuration option called "sync_back_to_maildir"
 > which is false by default. When true, this option updates your actual
 > Maildir files when you change a label, this is done in real time
 > before being saved to xapian. So for instance if you use offlineimap
 > the changes should be synched back to your IMAP server by the next
 > poll in Sup (assuming that your before-poll hook runs offlineimap);
 > 
 > - I wrote a "sup-sync-back-maildir" script that will synchronize all
 > messages from your sources at once. This should be executed the first
 > time BEFORE any polling, otherwise your IMAP server will be synched to
 > your Sup index and if you did not used another client to mark your
 > emails as read etc. (like me) and to update your IMAP server then you
 > will lose all what you did in Sup (labels, etc.);
 > 
 > - I added two new hidden labels in Sup: 'replied' and 'forwarded',
 > they are automatically added when replying, forwarding or bouncing a
 > message, this is to bring a better Maildir support to Sup, this should
 > be invisible for users;
 > 
 > - However you will probably lose all your remote 'replied' and
 > 'forwarded' flags after your first backward synchronization to Maildir
 > since Sup dropped these labels when it added your messages to xapian.
 > 
 > - Moving a message from a Maildir source to another is not (yet?)
 > supported.
 > 
 > So, you should now be able to use multiple clients to handle your
 > emails, for instance I can use gmail to mark messages as starred or
 > so, it will be reflected in Sup. In the same way, if I mark a message
 > as read in Sup it will appear as read in the gmail web interface.
 > 
 > I tested this code with my main account (2 GB and like 15 Maildir
 > sources) it works fine so far.
 > 
 > Be warned that it is highly experimental, but if you use offlineimap
 > it has a realdelete option so you *should* not lose any email, in the
 > worst case it might mess your flags/labels up.
 > 
 > I tested it only with offlineimap working with a gmail account.
 > 
 > Quick howto:
 > 
 > 1. Close sup
 > 2. Backup your emails and your xapian index or use another Sup session
 > 3. Clone the branch [1]
 > 4. IMPORTANT: run "bin/sup-sync-back-maildir" to synchronize the
 >    Maildirs you wish, check the help
 > 5. Add ":sync_back_to_maildir: true" to your config.yaml
 > 6. Run and use sup
 > 
 > Please test it and make reviews! :)
 > 
 > NOTE: if you want this to work, get sure that you are not calling
 > offlineimap with the '-q' option (as suggested in the wiki) otherwise
 > it will ignore the flag updates on IMAP server, so check your
 > before-poll hook.
 > 
 > [0] http://git.fensalir.fr/?p=dleone/sup.git
 > [1] $ git clone git://fensalir.fr/dleone/sup.git -b maildir-sync
 > 
 
 From damien.leone@fensalir.fr  Fri Jul  9 13:14:58 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Fri, 09 Jul 2010 19:14:58 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278694430-sup-914@ezyang>
 References: <1278693172-sup-6247@mailer> <1278694430-sup-914@ezyang>
 Message-ID: <1278695023-sup-8869@mailer>
 
 Hello Edward,
 
 > 1. Suppose we decide to test your branch, and run into some show-stopper
 >    bugs.  What?s the procedure for backing out changes?
 
 You should backup your Maildir directories and your xapian index, this
 is what I did and if something goes wrong I can still restore my
 backups and run offlineimap to resynch everything as it was before on
 the IMAP server.
 
 But for my initial tests I used a test Maildir containing nothing
 important so maybe this is what you should do. ;)
 
 > 2. What precisely are the semantics for handling Sup labels?
 
 I followed the Maildir documentation [0]. When synching a message, Sup
 builds the flags string and relink the file to the disk, moving it
 from new/ to cur/ or vice versa if needed.
 
 Related labels are: unread, replied, forwarded, deleted, draft (but it
 won't work for now because draft messages in Sup are handled
 separately), flagged (aka starred).
 
 Every other labels are ignored because not supported.
 
 [0] http://cr.yp.to/proto/maildir.html
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100709/eef91e70/attachment.bin>
 
 From ezyang@MIT.EDU  Fri Jul  9 15:46:08 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Fri, 09 Jul 2010 15:46:08 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278695023-sup-8869@mailer>
 References: <1278693172-sup-6247@mailer> <1278694430-sup-914@ezyang>
 	<1278695023-sup-8869@mailer>
 Message-ID: <1278704722-sup-2256@ezyang>
 
 Excerpts from Damien Leone's message of Fri Jul 09 13:14:58 -0400 2010:
 > > 1. Suppose we decide to test your branch, and run into some show-stopper
 > >    bugs.  What?s the procedure for backing out changes?
 > 
 > You should backup your Maildir directories and your xapian index, this
 > is what I did and if something goes wrong I can still restore my
 > backups and run offlineimap to resynch everything as it was before on
 > the IMAP server.
 > 
 > But for my initial tests I used a test Maildir containing nothing
 > important so maybe this is what you should do. ;)
 
 I doubt a test Maildir will give you the sort of test pressure that you?re
 looking for. Go go real datasets.
 
 > > 2. What precisely are the semantics for handling Sup labels?
 > 
 > I followed the Maildir documentation [0]. When synching a message, Sup
 > builds the flags string and relink the file to the disk, moving it
 > from new/ to cur/ or vice versa if needed.
 > 
 > Related labels are: unread, replied, forwarded, deleted, draft (but it
 > won't work for now because draft messages in Sup are handled
 > separately), flagged (aka starred).
 
 Great, thanks!  I?ll give it a spin this weekend.
 
 Edward
 
 From rlane@club.cc.cmu.edu  Mon Jul 12 22:10:51 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Mon, 12 Jul 2010 22:10:51 -0400
 Subject: [sup-devel] [PATCH] mbox: fix date format
 In-Reply-To: <1278592087-614-1-git-send-email-sascha-pgp@silbe.org>
 References: <1278592087-614-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278987038-sup-162@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Mon Jul 12 22:13:26 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Mon, 12 Jul 2010 22:13:26 -0400
 Subject: [sup-devel] [PATCH] fix crash in sup-sync if the default sent
 	source is used
 In-Reply-To: <1278592111-654-1-git-send-email-sascha-pgp@silbe.org>
 References: <1278592111-654-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278987191-sup-868@zyrg.net>
 
 Applied to master.
 
 From rlane@club.cc.cmu.edu  Mon Jul 12 22:19:28 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Mon, 12 Jul 2010 22:19:28 -0400
 Subject: [sup-devel] [PATCH] fix crash during poll if Maildir has been
 	emptied
 In-Reply-To: <1278602797-2553-1-git-send-email-sascha-pgp@silbe.org>
 References: <1278602797-2553-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278987290-sup-1833@zyrg.net>
 
 I'm going to be merging the source rewrite into master soon, so 
 this patch won't apply. I'll make sure empty maildirs work.
 
 From rlane@club.cc.cmu.edu  Mon Jul 12 22:23:46 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Mon, 12 Jul 2010 22:23:46 -0400
 Subject: [sup-devel] [PATCH] don't leak fds for mbox sources
 In-Reply-To: <1278609322-1401-1-git-send-email-sascha-pgp@silbe.org>
 References: <1278609322-1401-1-git-send-email-sascha-pgp@silbe.org>
 Message-ID: <1278987576-sup-9788@zyrg.net>
 
 This patch also won't apply once the source rewrite is merged. It's
 definitely a good idea though, so I'll (eventually) adapt this patch to
 the new codebase. Or feel free to rebase against next.
 
 The code itself looks good, I have no style nits with it.
 
 From sgoldman@teza.com  Fri Jul  9 16:39:37 2010
 From: sgoldman@teza.com (Steve Goldman)
 Date: Fri, 09 Jul 2010 16:39:37 -0400
 Subject: [sup-devel] Character encoding crash
 Message-ID: <1278707910-sup-3979@sodium.teza.com>
 
 
 I would send the characters from the message I was replying to, but
 then sup would crash.
 
 Running from master, latest commit June 9th.
 
 Thanks.
 
 
 --- ArgumentError from thread: main
 invalid byte sequence in US-ASCII
 /teza/home-h/sgoldman/sup-src/mainline/lib/sup/modes/edit-message-mode.rb:369:in `build_message'
 /teza/home-h/sgoldman/sup-src/mainline/lib/sup/modes/edit-message-mode.rb:343:in `send_message'
 /teza/home-h/sgoldman/sup-src/mainline/lib/sup/mode.rb:59:in `handle_input'
 /teza/home-h/sgoldman/sup-src/mainline/lib/sup/buffer.rb:277:in `handle_input'
 bin/sup:260:in `<module:Redwood>'
 bin/sup:69:in `<main>'
 
 From viroteck@viroteck.net  Sat Jul 10 14:53:20 2010
 From: viroteck@viroteck.net (Robin Burchell)
 Date: Sat, 10 Jul 2010 19:53:20 +0100
 Subject: [sup-devel] [PATCH] Fetch the account found from the to/cc info,
 	don't just use it as-is.
 Message-ID: <1278788000-23738-1-git-send-email-robin.burchell@collabora.co.uk>
 
 This prevents people's To: info screwing up my replies to them which is
 immensely irritating to me.
 ---
  lib/sup/modes/reply-mode.rb |    2 +-
  1 files changed, 1 insertions(+), 1 deletions(-)
 
 diff --git a/lib/sup/modes/reply-mode.rb b/lib/sup/modes/reply-mode.rb
 index bbac922..294bcfa 100644
 --- a/lib/sup/modes/reply-mode.rb
 +++ b/lib/sup/modes/reply-mode.rb
 @@ -76,7 +76,7 @@ EOS
      ## otherwise, try and find an account somewhere in the list of to's
      ## and cc's.
      elsif(b = (@m.to + @m.cc).find { |p| AccountManager.is_account? p })
 -      b
 +      AccountManager.account_for(b.email)
      ## if all else fails, use the default
      else
        AccountManager.default_account
 -- 
 1.7.0.4
 
 
 From rlane@club.cc.cmu.edu  Wed Jul 14 23:34:58 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Wed, 14 Jul 2010 23:34:58 -0400
 Subject: [sup-devel] maildir branch merged to master
 Message-ID: <1279164331-sup-2007@zyrg.net>
 
 The test suite isn't done yet, but I've been running this code for a
 while and I think it's stable enough to merge. This will also make
 merging the larger patches posted recently easier since next won't be so
 far diverged.
 
 From damien.leone@fensalir.fr  Sat Jul 17 08:55:56 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Sat, 17 Jul 2010 14:55:56 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1279371142-sup-3470@mailer>
 
 So, any reviewer? I thought backward synchronization was a wanted
 feature.
 
 For information I have been testing it with my real account for more
 than a week now without noticing any serious issue, I often check my
 gmail web interface and everything is synched.
 
 I had once a location trouble after having sent an email but I could
 not reproduce and it has been resolved after a sup restart.
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100717/436ae663/attachment.bin>
 
 From bwalton@artsci.utoronto.ca  Sat Jul 17 13:05:44 2010
 From: bwalton@artsci.utoronto.ca (Ben Walton)
 Date: Sat, 17 Jul 2010 13:05:44 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279371142-sup-3470@mailer>
 References: <1278693172-sup-6247@mailer> <1279371142-sup-3470@mailer>
 Message-ID: <1279386307-sup-360@pinkfloyd.chass.utoronto.ca>
 
 Excerpts from Damien Leone's message of Sat Jul 17 08:55:56 -0400 2010:
 
 > So, any reviewer? I thought backward synchronization was a wanted
 > feature.
 
 I believe it is (not by me personally as I'm not affected).  How does
 it play with the recent maildir branch merge from Rich?
 
 Thanks
 -Ben
 --
 Ben Walton
 Systems Programmer - CHASS
 University of Toronto
 C:416.407.5610 | W:416.978.4302
 
 
 From damien.leone@fensalir.fr  Sat Jul 17 13:26:16 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Sat, 17 Jul 2010 19:26:16 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279386307-sup-360@pinkfloyd.chass.utoronto.ca>
 References: <1278693172-sup-6247@mailer> <1279371142-sup-3470@mailer>
 	<1279386307-sup-360@pinkfloyd.chass.utoronto.ca>
 Message-ID: <1279387408-sup-5931@mailer>
 
 Hi Ben,
 
 Excerpts from Ben Walton's message of Sat Jul 17 19:05:44 +0200 2010:
 > I believe it is (not by me personally as I'm not affected).  How does
 > it play with the recent maildir branch merge from Rich?
 
 It should play well since I based my branch on the 'next' branch which
 already had the 'maildir' merge.
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100717/130153bb/attachment.bin>
 
 From tero@tilus.net  Sun Jul 18 07:19:55 2010
 From: tero@tilus.net (Tero Tilus)
 Date: Sun, 18 Jul 2010 14:19:55 +0300
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279371142-sup-3470@mailer>
 References: <1278693172-sup-6247@mailer> <1279371142-sup-3470@mailer>
 Message-ID: <1279451667-sup-3500@tilus.net>
 
 Damien Leone, 2010-07-17 15:55:
 > So, any reviewer?
 
 I haven't had the time yet.  I definitely will once the temperature
 here in Finland drops below 20 degrees Celcius. ;)  For now I'm almost
 totally off the net and on the beach.
 
 > I thought backward synchronization was a wanted
 > feature.
 
 And you were _very_ much right!
 
 > I often check my gmail web interface and everything is synched.
 
 My setup (once I get there) will be mainly sup + occasional mutt (on
 maildirs) and evolution (via dovecot).
 
 -- 
 Tero Tilus ## 050 3635 235 ## http://tero.tilus.net/
 
 From ezyang@MIT.EDU  Sun Jul 18 13:53:39 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Sun, 18 Jul 2010 13:53:39 -0400
 Subject: [sup-devel] [PATCH] Return nothing between if end index is not
 	found.
 Message-ID: <1279475566-sup-7173@ezyang>
 
 This patch fixed Nil error that occurred when I had a GPG signed
 mail with a starting marker but no ending marker.
 
 >From 22d12ffe7caa0d11c91876e2c8c7806854d0b3ab Mon Sep 17 00:00:00 2001
 From: Edward Z. Yang <ezyang at mit.edu>
 Date: Sun, 18 Jul 2010 03:38:52 -0400
 Subject: [PATCH] Return nothing between if end index is not found.
 
 Signed-off-by: Edward Z. Yang <ezyang at mit.edu>
 ---
  lib/sup/util.rb |    1 +
  1 files changed, 1 insertions(+), 0 deletions(-)
 
 diff --git a/lib/sup/util.rb b/lib/sup/util.rb
 index d19caca..dad479f 100644
 --- a/lib/sup/util.rb
 +++ b/lib/sup/util.rb
 @@ -479,6 +479,7 @@ module Enumerable
  
    ## returns all the entries which are equal to startline up to endline
    def between startline, endline
 +    return [] if not index(endline)
      select { |l| true if l == startline .. l == endline }
    end
  end
 -- 
 1.7.0.4
 
 From ezyang@MIT.EDU  Sun Jul 18 18:46:33 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Sun, 18 Jul 2010 18:46:33 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1279492496-sup-3638@ezyang>
 
 Sup Damien,
 
 I cut over to your branch, plus one patch of my own making.
 
 THE GOOD NEWS
 
 It works!  Whereby, messages are marked read in my webmail client and
 deleted messages go away.
 
 THE BAD NEWS
 
 It makes Sup very crashy and there's some strange behavior.  This is probably
 the sixth time I'm composing this message (yeah, haven't learned my
 lesson.) In more detail...
 
 * You want the following OfflineImap patch, because otherwise the massive
   changes to labels that will surely be induced by backwards syncing will
   cause OfflineImap to CRAWL.
 
 diff --git a/offlineimap/folder/LocalStatus.py b/offlineimap/folder/LocalStatus.py
 index 157989d..7f4dc5b 100644
 --- a/offlineimap/folder/LocalStatus.py
 +++ b/offlineimap/folder/LocalStatus.py
 @@ -33,7 +33,7 @@ class LocalStatusFolder(BaseFolder):
          self.messagelist = None
          self.repository = repository
          self.savelock = threading.Lock()
 -        self.doautosave = 1
 +        self.doautosave = 0
          self.accountname = accountname
          BaseFolder.__init__(self)
 
   Though, you probably want to undo this patch when you're done.  Autosaving
   means partial progress is saved.
 
 * Backwards syncing appears to make old deleted mail show up again; I'm not
   sure if it's because touching the Maildir causes Sup to pick up on deleted
   tags again or something, but I had to go back and delete lots of deleted mail.
 
 * When Sup is deleting messages, it is /extremely/ unsafe to iterate over
   a list of all messages in the index and query them for info: the threading means
   that any given message might not exist anymore.  I had a hook that kept triggering
   this bug, and earlier on in the process drawing functions would also frequently
   Nil error.  This is probably under-tested codepaths in Sup that we should track
   down; unfortunately I didn't have the presence of mind to record traces until
   later (and the ones I got all look like my hooks fault).
 
 * In a similar vein, while messages are going away, '@' will occasionally return
   no messages (pressing again usually fixes it).  Messages that are labelled inbox
   will show up in 'U', but not in the inbox view without a Sup restart.
 
 I haven't done a code review yet.
 
 Cheers,
 Edward
 
 From ezyang@MIT.EDU  Sun Jul 18 18:50:39 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Sun, 18 Jul 2010 18:50:39 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1279493200-sup-3095@ezyang>
 
 Here are some suggestions...
 
 Excerpts from Damien Leone's message of Fri Jul 09 12:44:01 -0400 2010:
 > - I wrote a "sup-sync-back-maildir" script that will synchronize all
 > messages from your sources at once. This should be executed the first
 > time BEFORE any polling, otherwise your IMAP server will be synched to
 > your Sup index and if you did not used another client to mark your
 > emails as read etc. (like me) and to update your IMAP server then you
 > will lose all what you did in Sup (labels, etc.);
 
 I would recommend having some sort of flag like 'sync-back'd' programatically
 generated by sup-sync-back-maildir so that Sup will refuse to start if
 you turned on sync-backs without running this command.
 
 > - However you will probably lose all your remote 'replied' and
 > 'forwarded' flags after your first backward synchronization to Maildir
 > since Sup dropped these labels when it added your messages to xapian.
 
 It seems to me that merging the Maildir flags and the Sup flags would
 let you preserve this, but for heavy Sup users this isn't a big deal.
 
 > - Moving a message from a Maildir source to another is not (yet?)
 > supported.
 
 I think this will be a pretty killer feature to have, because it means
 we can then assign "primary" tags and use those to have sync-backs move
 messages between Maildir folders (on the same account, I imagine), which
 would translate into a folder-sorted webmail view, which is about the
 best you can do without first-class label support in Maildir.  I know of
 at least one friend who this would be a deal-maker.
 
 Cheers,
 Edward
 
 From ezyang@MIT.EDU  Sun Jul 18 18:57:31 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Sun, 18 Jul 2010 18:57:31 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1279493829-sup-2931@ezyang>
 
 Here is one trace that I managed to scrounge up
 
 undefined method `content_width' for nil:NilClass
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:953:in `from_width'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:870:in `text_for_thread_at'
 /home/ezyang/Dev/sup/lib/sup/hook.rb:55:in `each_with_index'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:869:in `each'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:869:in `each_with_index'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:869:in `text_for_thread_at'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:811:in `regen_text'
 /home/ezyang/Dev/sup/lib/sup/util.rb:423:in `map_with_index'
 /home/ezyang/Dev/sup/lib/sup/hook.rb:55:in `each_with_index'
 /home/ezyang/Dev/sup/lib/sup/util.rb:423:in `each'
 /home/ezyang/Dev/sup/lib/sup/util.rb:423:in `each_with_index'
 /home/ezyang/Dev/sup/lib/sup/util.rb:423:in `map_with_index'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:811:in `regen_text'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:262:in `update'
 /home/ezyang/Dev/sup/lib/sup/modes/thread-index-mode.rb:217:in `handle_location_deleted_update'
 /home/ezyang/Dev/sup/lib/sup/update.rb:26:in `send'
 /home/ezyang/Dev/sup/lib/sup/update.rb:26:in `relay'
 /home/ezyang/Dev/sup/lib/sup/update.rb:26:in `each'
 /home/ezyang/Dev/sup/lib/sup/update.rb:26:in `relay'
 /home/ezyang/Dev/sup/lib/sup/util.rb:588:in `send'
 /home/ezyang/Dev/sup/lib/sup/util.rb:588:in `method_missing'
 /home/ezyang/Dev/sup/lib/sup/poll.rb:197:in `poll_from'
 /home/ezyang/Dev/sup/lib/sup/index.rb:258:in `each_message'
 /home/ezyang/Dev/sup/lib/sup/index.rb:245:in `each_id'
 /home/ezyang/Dev/sup/lib/sup/index.rb:245:in `each'
 /home/ezyang/Dev/sup/lib/sup/index.rb:245:in `each_id'
 /home/ezyang/Dev/sup/lib/sup/index.rb:257:in `each_message'
 /home/ezyang/Dev/sup/lib/sup/util.rb:588:in `send'
 /home/ezyang/Dev/sup/lib/sup/util.rb:588:in `method_missing'
 [snip]
 
 Cheers,
 Edward
 
 From michael+sup@stapelberg.de  Mon Jul 19 15:39:31 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Mon, 19 Jul 2010 21:39:31 +0200
 Subject: [sup-devel] [PATCH] Bugfix: Make sure to use UNIX line-endings for
 	GPG decrypted payloads
 Message-ID: <1279568308-sup-1044@midna.zekjur.net>
 
 Hi,
 
 please find attached a patch for the following issue (quuote from the commit
 message):
 
 	Bugfix: Make sure to use UNIX line-endings for GPG decrypted payloads
 
 	Otherwise, RMail will fail to correctly parse multipart/signed messages
 
 Best regards,
 Michael
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: 0001-Bugfix-Make-sure-to-use-UNIX-line-endings-for-GPG-de.patch
 Type: application/octet-stream
 Size: 891 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100719/6cef1d48/attachment.obj>
 
 From michael+sup@stapelberg.de  Mon Jul 19 17:38:18 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Mon, 19 Jul 2010 23:38:18 +0200
 Subject: [sup-devel] (REVIEW NEEDED) [PATCH] Recode attachments to UTF-8
 	when forwarding
 Message-ID: <1279575318-sup-7427@midna.zekjur.net>
 
 Hi,
 
 can you please review the attached patch? It recodes attachments from whichever
 charset they are encoded in to UTF-8 when forwarding messages in order to
 please rmail (it crashes with attachments encoded in Windows-1252 otherwise).
 
 FYI, the crash backtrace is the following (pathes shortened):
 --- Encoding::CompatibilityError from thread: main
 incompatible character encodings: UTF-8 and ASCII-8BIT
 rmail-1.0.0/lib/rmail/serialize.rb:112:in `serialize_low'
 rmail-1.0.0/lib/rmail/serialize.rb:99:in `block in serialize_low'
 rmail-1.0.0/lib/rmail/message.rb:155:in `block in each_part'
 rmail-1.0.0/lib/rmail/message.rb:154:in `each'
 rmail-1.0.0/lib/rmail/message.rb:154:in `each_part'
 rmail-1.0.0/lib/rmail/serialize.rb:96:in `serialize_low'
 rmail-1.0.0/lib/rmail/serialize.rb:63:in `serialize'
 rmail-1.0.0/lib/rmail/message.rb:146:in `to_s'
 sup/lib/sup/modes/edit-message-mode.rb:344:in `puts'
 sup/lib/sup/modes/edit-message-mode.rb:344:in `block in send_message'
 sup/lib/sup/modes/edit-message-mode.rb:344:in `popen'
 sup/lib/sup/modes/edit-message-mode.rb:344:in `send_message'
 sup/lib/sup/mode.rb:59:in `handle_input'
 sup/lib/sup/buffer.rb:277:in `handle_input'
 
 So, if you have a better way of handling this, please let us know.
 
 Best regards,
 Michael
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: 0001-Recode-attachments-to-UTF-8-when-forwarding.patch
 Type: application/octet-stream
 Size: 3970 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100719/2712c0ee/attachment.obj>
 
 From michael+sup@stapelberg.de  Mon Jul 19 18:50:37 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Tue, 20 Jul 2010 00:50:37 +0200
 Subject: [sup-devel] [PATCH] Return nothing between if end index is not
 	found.
 In-Reply-To: <1279475566-sup-7173@ezyang>
 References: <1279475566-sup-7173@ezyang>
 Message-ID: <1279579814-sup-4745@midna.zekjur.net>
 
 Hi Edward,
 
 Excerpts from Edward Z. Yang's message of 2010-07-18 19:53:39 +0200:
 > This patch fixed Nil error that occurred when I had a GPG signed
 > mail with a starting marker but no ending marker.
 Thanks for the patch, I can confirm it working and necessary.
 
 Best regards,
 Michael
 
 From michael+sup@stapelberg.de  Mon Jul 19 18:57:18 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Tue, 20 Jul 2010 00:57:18 +0200
 Subject: [sup-devel] [PATCH] Bugfix: Make sure to use UNIX line-endings
 	for GPG decrypted payloads
 In-Reply-To: <1279568308-sup-1044@midna.zekjur.net>
 References: <1279568308-sup-1044@midna.zekjur.net>
 Message-ID: <1279580212-sup-1480@midna.zekjur.net>
 
 Hi,
 
 Excerpts from Michael Stapelberg's message of 2010-07-19 21:39:31 +0200:
 > please find attached a patch for the following issue (quuote from the commit
 > message):
 > 
 >     Bugfix: Make sure to use UNIX line-endings for GPG decrypted payloads
 > 
 >     Otherwise, RMail will fail to correctly parse multipart/signed messages
 Attached you can find the fixed version of this patch.
 
 Best regards,
 Michael
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: 0001-Bugfix-Make-sure-to-use-UNIX-line-endings-for-GPG-de.patch
 Type: application/octet-stream
 Size: 957 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100720/ba0614ab/attachment.obj>
 
 From damien.leone@fensalir.fr  Tue Jul 20 05:39:41 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Tue, 20 Jul 2010 11:39:41 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279492496-sup-3638@ezyang>
 References: <1278693172-sup-6247@mailer> <1279492496-sup-3638@ezyang>
 Message-ID: <1279618205-sup-993@mailer>
 
 Sup Edward,
 
 Thank you for your feedback!
 
 Excerpts from Edward Z. Yang's message of Mon Jul 19 00:46:33 +0200 2010:
 > It makes Sup very crashy and there's some strange behavior.  This is probably
 > the sixth time I'm composing this message (yeah, haven't learned my
 > lesson.) In more detail...
 
 I'm impressed, I never experienced such crashes, it is quite stable
 here but I use only very basic hooks.
 
 > * Backwards syncing appears to make old deleted mail show up again; I'm not
 >   sure if it's because touching the Maildir causes Sup to pick up on deleted
 >   tags again or something, but I had to go back and delete lots of deleted mail.
 
 Hmm, strange, do you use the latest commits of the branch? I pushed a
 fix for deleted mails about 10 days ago.
 
 The mails that show up again don't have the deleted flag at all?
 
 > * When Sup is deleting messages, it is /extremely/ unsafe to iterate over
 >   a list of all messages in the index and query them for info: the threading means
 >   that any given message might not exist anymore.  I had a hook that kept triggering
 >   this bug, and earlier on in the process drawing functions would also frequently
 >   Nil error.  This is probably under-tested codepaths in Sup that we should track
 >   down; unfortunately I didn't have the presence of mind to record traces until
 >   later (and the ones I got all look like my hooks fault).
 
 I did not test the code with hooks, what do you suggest for this one?
 
 > * In a similar vein, while messages are going away, '@' will occasionally return
 >   no messages (pressing again usually fixes it).  Messages that are labelled inbox
 >   will show up in 'U', but not in the inbox view without a Sup restart.
 
 I never use this keybinding but indeed I noticed that pressing '@' two
 times in a row was not showing anything.
 
 I will try to track down these bugs.
 
 Thanks again!
 Cheers,
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100720/5d457344/attachment.bin>
 
 From damien.leone@fensalir.fr  Tue Jul 20 05:45:41 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Tue, 20 Jul 2010 11:45:41 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279493200-sup-3095@ezyang>
 References: <1278693172-sup-6247@mailer> <1279493200-sup-3095@ezyang>
 Message-ID: <1279618822-sup-4912@mailer>
 
 Hi,
 
 Excerpts from Edward Z. Yang's message of Mon Jul 19 00:50:39 +0200 2010:
 > I would recommend having some sort of flag like 'sync-back'd' programatically
 > generated by sup-sync-back-maildir so that Sup will refuse to start if
 > you turned on sync-backs without running this command.
 
 Good idea, that might prevent some bad suprises.
 
 > It seems to me that merging the Maildir flags and the Sup flags would
 > let you preserve this, but for heavy Sup users this isn't a big deal.
 
 True, and that could be done by sup-sync-back-maildir.
 
 > I think this will be a pretty killer feature to have, because it means
 > we can then assign "primary" tags and use those to have sync-backs move
 > messages between Maildir folders (on the same account, I imagine), which
 > would translate into a folder-sorted webmail view, which is about the
 > best you can do without first-class label support in Maildir.  I know of
 > at least one friend who this would be a deal-maker.
 
 And that doesn't seem hard to implement, we should definitely consider
 this feature, maybe after the primary code has been stabilized.
 
 Cheers,
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100720/3baa46e4/attachment.bin>
 
 From eric+sup-devel@gerlach.ca  Tue Jul 20 15:15:40 2010
 From: eric+sup-devel@gerlach.ca (Eric Gerlach)
 Date: Tue, 20 Jul 2010 15:15:40 -0400
 Subject: [sup-devel] Crash,
 	getting "wrong id on nil" when loading new messages
 Message-ID: <1279653097-sup-8302@gerlach.ca>
 
 I get the following crash when loading more messages (at a certain
 point).  I've created an issue for it (111), but I'd like some help
 debugging this further.  I'm pretty sure it's caused by some sort of
 problematic message in my Maildir, but I don't know which one.  I'm
 pretty good with Ruby, but not so much with the sup codebase.
 
 I've traced it down to ThreadIndexMode#update, particularly the part
 where it calls t.first.id.  I've used ruby-debug to be able to isolate
 the particular thread, and the thread has no children, but then I don't
 quite know what to do from there.  How can I trace a thread object back
 to its origins?
 
 Help?
 
 Cheers,
 
 Eric
 
 wrong id called on nil
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup.rb:19:in `id'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `update'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/hook.rb:55:in `sort_by'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `each'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `sort_by'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `update'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:224:in `synchronize'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:224:in `update'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:642:in
 `__unprotected_load_n_threads'
 (eval):12:in `load_n_threads'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:614:in
 `load_n_threads_background'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup.rb:74:in `reporting_thread'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:613:in
 `load_n_threads_background'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:684:in
 `__unprotected_load_threads'
 (eval):12:in `load_threads'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/mode.rb:59:in `send'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/mode.rb:59:in `handle_input'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/buffer.rb:279:in `handle_input'
 /var/lib/gems/1.8/gems/sup-0.11/bin/sup:279
 /var/lib/gems/1.8/bin/sup:19:in `load'
 /var/lib/gems/1.8/bin/sup:19
 
 From sup-bugs@masanjin.net  Tue Jul 20 14:49:03 2010
 From: sup-bugs@masanjin.net (anonymous)
 Date: Tue, 20 Jul 2010 18:49:03 +0000
 Subject: [sup-devel] [issue111] wrong id called on nil
 In-Reply-To: <1279651743.08.0.100305852621.issue111@masanjin.net>
 Message-ID: <1279651743.08.0.100305852621.issue111@masanjin.net>
 
 
 New submission from anonymous:
 
 I get the following crash when loading more messages.  Will start a thread on
 sup-devel to get help with debugging.
 
 wrong id called on nil
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup.rb:19:in `id'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `update'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/hook.rb:55:in `sort_by'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `each'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `sort_by'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:226:in `update'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:224:in `synchronize'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:224:in `update'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:642:in
 `__unprotected_load_n_threads'
 (eval):12:in `load_n_threads'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:614:in `load_n_threads_background'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup.rb:74:in `reporting_thread'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:613:in `load_n_threads_background'
 /usr/lib/ruby/1.8/sup/modes/thread-index-mode.rb:684:in `__unprotected_load_threads'
 (eval):12:in `load_threads'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/mode.rb:59:in `send'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/mode.rb:59:in `handle_input'
 /var/lib/gems/1.8/gems/sup-0.11/lib/sup/buffer.rb:279:in `handle_input'
 /var/lib/gems/1.8/gems/sup-0.11/bin/sup:279
 /var/lib/gems/1.8/bin/sup:19:in `load'
 /var/lib/gems/1.8/bin/sup:19
 
 ----------
 messages: 256
 nosy: anonymous
 priority: bug
 ruby_version: 1.8.7
 status: unread
 sup_version: 0.11
 title: wrong id called on nil
 
 _________________________________________
 Sup issue tracker <sup-bugs at masanjin.net>
 <http://masanjin.net/sup-bugs/issue111>
 _________________________________________
 
 From bwalton@artsci.utoronto.ca  Tue Jul 20 15:37:10 2010
 From: bwalton@artsci.utoronto.ca (Ben Walton)
 Date: Tue, 20 Jul 2010 15:37:10 -0400
 Subject: [sup-devel] Crash,
 	getting "wrong id on nil" when loading new messages
 In-Reply-To: <1279653097-sup-8302@gerlach.ca>
 References: <1279653097-sup-8302@gerlach.ca>
 Message-ID: <1279654563-sup-816@pinkfloyd.chass.utoronto.ca>
 
 Excerpts from Eric Gerlach's message of Tue Jul 20 15:15:40 -0400 2010:
 
 Hi Eric,
 
 > I get the following crash when loading more messages (at a certain
 > point).  I've created an issue for it (111), but I'd like some help
 > debugging this further.  I'm pretty sure it's caused by some sort of
 > problematic message in my Maildir, but I don't know which one.  I'm
 > pretty good with Ruby, but not so much with the sup codebase.
 
 I'm pretty sure that issue is fixed in next.  I used to get hit by it
 but haven't for some time since Rich committed some fixes for it.
 
 Try running from the git codebase instead of the gem and see if it
 goes away.
 
 HTH.
 -Ben
 --
 Ben Walton
 Systems Programmer - CHASS
 University of Toronto
 C:416.407.5610 | W:416.978.4302
 
 
 From michael+sup@stapelberg.de  Tue Jul 20 16:09:39 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Tue, 20 Jul 2010 22:09:39 +0200
 Subject: [sup-devel] [PATCH] Return nothing between if end index is not
 	found.
 In-Reply-To: <1279579814-sup-4745@midna.zekjur.net>
 References: <1279475566-sup-7173@ezyang> <1279579814-sup-4745@midna.zekjur.net>
 Message-ID: <1279656544-sup-4931@midna.zekjur.net>
 
 Hi,
 
 Excerpts from Michael Stapelberg's message of 2010-07-20 00:50:37 +0200:
 > Thanks for the patch, I can confirm it working and necessary.
 I was wrong. I just opened an email and it was no longer verified (it is
 GPG-signed). I will post an updated patch soon.
 
 Best regards,
 Michael
 
 From eric+sup-devel@gerlach.ca  Tue Jul 20 16:11:33 2010
 From: eric+sup-devel@gerlach.ca (Eric Gerlach)
 Date: Tue, 20 Jul 2010 16:11:33 -0400
 Subject: [sup-devel] Crash,
 	getting "wrong id on nil" when loading new messages
 In-Reply-To: <1279654563-sup-816@pinkfloyd.chass.utoronto.ca>
 References: <1279653097-sup-8302@gerlach.ca>
 	<1279654563-sup-816@pinkfloyd.chass.utoronto.ca>
 Message-ID: <1279656659-sup-5582@gerlach.ca>
 
 Excerpts from Ben Walton's message of Tue Jul 20 15:37:10 -0400 2010:
 > Excerpts from Eric Gerlach's message of Tue Jul 20 15:15:40 -0400 2010:
 > 
 > Hi Eric,
 > 
 > > I get the following crash when loading more messages (at a certain
 > > point).  I've created an issue for it (111), but I'd like some help
 > > debugging this further.  I'm pretty sure it's caused by some sort of
 > > problematic message in my Maildir, but I don't know which one.  I'm
 > > pretty good with Ruby, but not so much with the sup codebase.
 > 
 > I'm pretty sure that issue is fixed in next.  I used to get hit by it
 > but haven't for some time since Rich committed some fixes for it.
 > 
 > Try running from the git codebase instead of the gem and see if it
 > goes away.
 
 It went away.  I'll go close that bug I created now.  Thanks for the
 help!
 
 Cheers,
 
 Eric
 
 From ezyang@MIT.EDU  Tue Jul 20 20:44:38 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Tue, 20 Jul 2010 20:44:38 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279618205-sup-993@mailer>
 References: <1278693172-sup-6247@mailer> <1279492496-sup-3638@ezyang>
 	<1279618205-sup-993@mailer>
 Message-ID: <1279672865-sup-6946@ezyang>
 
 Excerpts from Damien Leone's message of Tue Jul 20 05:39:41 -0400 2010:
 > Excerpts from Edward Z. Yang's message of Mon Jul 19 00:46:33 +0200 2010:
 > > It makes Sup very crashy and there's some strange behavior.  This is probably
 > > the sixth time I'm composing this message (yeah, haven't learned my
 > > lesson.) In more detail...
 > 
 > I'm impressed, I never experienced such crashes, it is quite stable
 > here but I use only very basic hooks.
 
 In retrospect, I'm fairly certain the majority of crashes were from user hooks.
 
 > > * Backwards syncing appears to make old deleted mail show up again; I'm not
 > >   sure if it's because touching the Maildir causes Sup to pick up on deleted
 > >   tags again or something, but I had to go back and delete lots of deleted mail.
 > 
 > Hmm, strange, do you use the latest commits of the branch? I pushed a
 > fix for deleted mails about 10 days ago.
 > 
 > The mails that show up again don't have the deleted flag at all?
 
 My Sup is running off of f150276f62e263d146fa81ede645b5e90e06a8a1.  They don't
 appear to have the deleted flag, but I didn't look too closely.  It's *possible*
 that a regular synchronization would have picked them up, if they accidentally
 got dropped from my Sup index but were still inside my Maildir.  I suggest not
 worrying too much about this unless someone else reproduces.
 
 > I did not test the code with hooks, what do you suggest for this one?
 
 I mentioned this in the id and Nil message; basically, when iterating over all
 messages messages might disappear, and some hooks need to check if a message is
 Nil before doing anything to it.
 
 > > * In a similar vein, while messages are going away, '@' will occasionally return
 > >   no messages (pressing again usually fixes it).  Messages that are labelled inbox
 > >   will show up in 'U', but not in the inbox view without a Sup restart.
 > 
 > I never use this keybinding but indeed I noticed that pressing '@' two
 > times in a row was not showing anything.
 
 It is possible that this is a bug in vanilla Sup.
 
 Cheers,
 Edward
 
 From damien.leone@fensalir.fr  Wed Jul 21 04:52:00 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Wed, 21 Jul 2010 10:52:00 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279672865-sup-6946@ezyang>
 References: <1278693172-sup-6247@mailer> <1279492496-sup-3638@ezyang>
 	<1279618205-sup-993@mailer> <1279672865-sup-6946@ezyang>
 Message-ID: <1279702065-sup-7856@mailer>
 
 Sup Edward,
 
 Excerpts from Edward Z. Yang's message of Wed Jul 21 02:44:38 +0200 2010:
 > Excerpts from Damien Leone's message of Tue Jul 20 05:39:41 -0400 2010:
 > In retrospect, I'm fairly certain the majority of crashes were from user hooks.
 
 Could you paste me these hooks please?
 
 > My Sup is running off of f150276f62e263d146fa81ede645b5e90e06a8a1.  They don't
 > appear to have the deleted flag, but I didn't look too closely.  It's *possible*
 > that a regular synchronization would have picked them up, if they accidentally
 > got dropped from my Sup index but were still inside my Maildir.  I suggest not
 > worrying too much about this unless someone else reproduces.
 
 What is your setup? IMAP + offlineimap? Do deleted mails work fine
 now?
 
 Cheers,
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100721/abd46324/attachment.bin>
 
 From damien.leone@fensalir.fr  Wed Jul 21 04:54:18 2010
 From: damien.leone@fensalir.fr (Damien Leone)
 Date: Wed, 21 Jul 2010 10:54:18 +0200
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1279702065-sup-7856@mailer>
 References: <1278693172-sup-6247@mailer> <1279492496-sup-3638@ezyang>
 	<1279618205-sup-993@mailer> <1279672865-sup-6946@ezyang>
 	<1279702065-sup-7856@mailer>
 Message-ID: <1279702378-sup-2954@mailer>
 
 Excerpts from Damien Leone's message of Wed Jul 21 10:52:00 +0200 2010:
 > What is your setup? IMAP + offlineimap?
 
 My bad, you mentioned offlineimap a few mails ago.
 
 --
 Damien Leone <damien.leone at fensalir.fr>
 
 Web: http://dleone.fensalir.fr/
 GPG: 0x82EB4DDF
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 198 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100721/2df79bdd/attachment.bin>
 
 From sup-bugs@masanjin.net  Thu Jul 22 13:34:51 2010
 From: sup-bugs@masanjin.net (anonymous)
 Date: Thu, 22 Jul 2010 17:34:51 +0000
 Subject: [sup-devel] [issue112] wildcard in from: field do not work
 In-Reply-To: <1279820091.87.0.35266318432.issue112@masanjin.net>
 Message-ID: <1279820091.87.0.35266318432.issue112@masanjin.net>
 
 
 New submission from anonymous:
 
 A search for from:robe* does not find all the expected emails (eg ones from
 people called Robert). 
 
 The translated query (line 399 of lib/sup/index.rb) is
 (from_email:robe OR from_name:robe)*
 
 I think it should expand to (from_email:robe* OR from_name:robe*)
 
 Indeed, a search for from_name:robe* yields the expected results.
 
 ----------
 messages: 257
 nosy: anonymous
 priority: bug
 ruby_version: 1.9.1
 status: unread
 sup_version: git/debian
 title: wildcard in from: field do not work
 
 _________________________________________
 Sup issue tracker <sup-bugs at masanjin.net>
 <http://masanjin.net/sup-bugs/issue112>
 _________________________________________
 
 From rlane@club.cc.cmu.edu  Sat Jul 24 17:30:16 2010
 From: rlane@club.cc.cmu.edu (Rich Lane)
 Date: Sat, 24 Jul 2010 17:30:16 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1278693172-sup-6247@mailer>
 References: <1278693172-sup-6247@mailer>
 Message-ID: <1280006159-sup-5717@zyrg.net>
 
 I've rebased this against master and merged it to branch maildir-sync in
 the official repo. This is a big change, so help with testing and code
 review would be much appreciated.
 
 From ezyang@MIT.EDU  Sun Jul 25 18:25:26 2010
 From: ezyang@MIT.EDU (Edward Z. Yang)
 Date: Sun, 25 Jul 2010 18:25:26 -0400
 Subject: [sup-devel] Adding backward synchronization for Maildir sources
 In-Reply-To: <1280006159-sup-5717@zyrg.net>
 References: <1278693172-sup-6247@mailer> <1280006159-sup-5717@zyrg.net>
 Message-ID: <1280096695-sup-2557@ezyang>
 
 Excerpts from Rich Lane's message of Sat Jul 24 17:30:16 -0400 2010:
 > I've rebased this against master and merged it to branch maildir-sync in
 > the official repo. This is a big change, so help with testing and code
 > review would be much appreciated.
 
 Hi Rich,
 
 I've moved my Sup checkout onto this branch. *test test test test*
 After a week of usage, I'd say the changes are pretty stable.
 
 Cheers,
 Edward
 
 From michael+sup@stapelberg.de  Tue Jul 27 08:54:57 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Tue, 27 Jul 2010 14:54:57 +0200
 Subject: [sup-devel] [PATCH] Use the configured real name combined with the
 	actual email address as Sender when replying
 Message-ID: <1280235249-sup-8730@midna.zekjur.net>
 
 Hi,
 
 quote from the commit message:
 	For example, mails with "To: michael+sup at stapelberg.de" don't get the
 	sender "michael+sup <michael+sup at stapelberg.de>" anymore but
 	"Michael Stapelberg <michael+sup at stapelberg.de>"
 
 Best regards,
 Michael
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: 0001-Use-the-configured-real-name-combined-with-the-actua.patch
 Type: application/octet-stream
 Size: 1313 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100727/b372afc4/attachment-0001.obj>
 
 From viroteck@viroteck.net  Tue Jul 27 12:16:48 2010
 From: viroteck@viroteck.net (Robin Burchell)
 Date: Tue, 27 Jul 2010 17:16:48 +0100
 Subject: [sup-devel] [PATCH] Use the configured real name combined with
 	the actual email address as Sender when replying
 In-Reply-To: <1280235249-sup-8730@midna.zekjur.net>
 References: <1280235249-sup-8730@midna.zekjur.net>
 Message-ID: <1280247338-sup-5612@virgin>
 
 Hi Michael,
 
 Excerpts from Michael Stapelberg's message of Tue Jul 27 13:54:57 +0100 2010:
 > Hi,
 > 
 > quote from the commit message:
 >     For example, mails with "To: michael+sup at stapelberg.de" don't get the
 >     sender "michael+sup <michael+sup at stapelberg.de>" anymore but
 >     "Michael Stapelberg <michael+sup at stapelberg.de>"
 
 I made a similar (though not perfect) fix which, I guess from your mail, never
 got through the moderation queue to sup-devel.
 
 See:
 http://github.com/rburchell/sup/commit/efd9752fc62d14fbabecf7719c42f13643c21db
 
 > Best regards,
 > Michael
 
 --
 Robin Burchell
 http://rburchell.com
 
 From michael+sup@stapelberg.de  Tue Jul 27 12:38:51 2010
 From: michael+sup@stapelberg.de (Michael Stapelberg)
 Date: Tue, 27 Jul 2010 18:38:51 +0200
 Subject: [sup-devel] [PATCH] Use the configured real name combined with
 	the actual email address as Sender when replying
 In-Reply-To: <1280247338-sup-5612@virgin>
 References: <1280235249-sup-8730@midna.zekjur.net> <1280247338-sup-5612@virgin>
 Message-ID: <1280248685-sup-4691@midna.zekjur.net>
 
 Hi Robin,
 
 Excerpts from Robin Burchell's message of 2010-07-27 18:16:48 +0200:
 > I made a similar (though not perfect) fix which, I guess from your mail, never
 > got through the moderation queue to sup-devel.
 Not sure, at least it did not get merged into mainline as of today. Also, your
 patch does not consider the regexes/alternates system but always just uses the
 main email address.
 
 Best regards,
 Michael
 
 From daniel.schoepe@googlemail.com  Thu Jul 29 19:30:19 2010
 From: daniel.schoepe@googlemail.com (Daniel Schoepe)
 Date: Fri, 30 Jul 2010 01:30:19 +0200
 Subject: [sup-devel] [PATCH] Fix crash when sending signed/encrypted
 	messages with attachments
 Message-ID: <1280446020-sup-4762@gilead>
 
 Currently, when trying to send an encrypted or signed message with an
 attachment, sup crashes, since it tries to .pack('m') the body of a
 multipart message, which is an array of RMail::Messages and thus
 crashes.
 
 The attached patch should fix that.
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: 0001-Fix-crash-when-sending-signed-encrypted-messages-wit.patch
 Type: application/octet-stream
 Size: 1415 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100730/98154b32/attachment.obj>
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: signature.asc
 Type: application/pgp-signature
 Size: 836 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100730/98154b32/attachment.bin>
 
 From alvherre@alvh.no-ip.org  Thu Jul 29 23:09:04 2010
 From: alvherre@alvh.no-ip.org (Alvaro Herrera)
 Date: Thu, 29 Jul 2010 23:09:04 -0400
 Subject: [sup-devel] Fix crash when opening attachment with slashes in
 	filename
 Message-ID: <1280459186-sup-8579@alvh.no-ip.org>
 
 Hi,
 
 This patch fixes a crash caused by trying to create a temp file with
 forward slashes.  The fix is simple -- replace the slashes with a
 different char.
 
 I reported this as issue 105 on masanjin.net previously, but that's
 apparently dead now.
 
 --
 ?lvaro Herrera <alvherre at alvh.no-ip.org>
 -------------- next part --------------
 A non-text attachment was scrubbed...
 Name: msgchunks.diff
 Type: application/octet-stream
 Size: 748 bytes
 Desc: not available
 URL: <http://rubyforge.org/pipermail/sup-devel/attachments/20100729/3402b837/attachment.obj>
 
 From sup-bugs@masanjin.net  Fri Jul 30 14:37:16 2010
 From: sup-bugs@masanjin.net (anonymous)
 Date: Fri, 30 Jul 2010 18:37:16 +0000
 Subject: [sup-devel] [issue113] Error at startup
 In-Reply-To: <1280515036.67.0.567575591319.issue113@masanjin.net>
 Message-ID: <1280515036.67.0.567575591319.issue113@masanjin.net>
 
 
 New submission from anonymous:
 
 I received a bug info when I tried to run sup in Archlinux. The contents from
 ~.sup/exception-log.txt.
 
 --- RuntimeError from thread: main
 can't modify frozen string
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/account.rb:53:in `force_encoding'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/account.rb:53:in `block in
 add_account'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/account.rb:53:in `each'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/account.rb:53:in `add_account'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/account.rb:37:in `initialize'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/util.rb:574:in `new'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup/util.rb:574:in `init'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/lib/sup.rb:126:in `start'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/bin/sup:153:in `<module:Redwood>'
 /usr/lib/ruby/gems/1.9.1/gems/sup-0.11/bin/sup:68:in `<top (required)>'
 /usr/bin/sup:19:in `load'
 /usr/bin/sup:19:in `<main>'
 
 ----------
 messages: 260
 nosy: anonymous
 priority: bug
 ruby_version: 1.9.1
 status: unread
 sup_version: 0.11.2
 title: Error at startup
 
 _________________________________________
 Sup issue tracker <sup-bugs at masanjin.net>
 <http://masanjin.net/sup-bugs/issue113>
 _________________________________________