From dave at badgers-in-foil.co.uk  Sun Sep  7 00:22:14 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Sun, 7 Sep 2003 00:22:14 +0000
Subject: [cvsspam-devel] Re: [Saw your post on RubyForge...]
In-Reply-To: <200309021031.03985.sdmitry@lrn.ru>
References: <20030901145610.GC18527@vhost.badgers-in-foil.co.uk> <200309021031.03985.sdmitry@lrn.ru>
Message-ID: <20030907002214.GA24229@vhost.badgers-in-foil.co.uk>

On Tue, Sep 02, 2003 at 10:31:03AM +0800, Dmitry V. Sabanin wrote:
> For cvsweb integration i had to fix 'cvsroot' problem, on rubyforge, cvsweb 
> URL's have additional variable 'cvsroot', like in this URL:
> http://rubyforge.org/cgi-bin/cvsweb.cgi/photokeeper/?cvsroot=muravey-tools
> And i fixed that with overwriting some code inside cvsspam.rb. 
> You can look at patch I've attached.

So, I'm adding a $repository_name configuration option that can be set
to either be a string, or the special constant, GUESS.

I notice that your regex removes any spaces from the last segment of
the CVSROOT path when working out the repository:  Do you know if this
something that Gforge does to the viewcvs configuration?  I was thinking
of just url_encode()ing the string, rather than removing spaces, but
maybe that's incorrect?

Also, would you be able to test the new version?  (I don't have ViewCVS
set up with multiple repositories.)


ta,
dave


From matt at raibledesigns.com  Wed Sep 10 16:25:05 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Wed, 10 Sep 2003 09:25:05 -0700 (PDT)
Subject: [cvsspam-devel] record_lastdir.rb: No such file or directory
Message-ID: <20030910162505.64654.qmail@web14306.mail.yahoo.com>

I'm getting the following error when trying to install CVSSpam:

cvs server: cannot exec /usr/local/cvsroot/CVSROOT/record_lastdir.rb: No such
file or directory
cvs server: Pre-commit check failed
cvs [server aborted]: correct above errors first!
cvs commit: saving log message in /tmp/cvsvoAWAR

I tried the solution mentioned at http://tinyurl.com/mvuk by changing my
CVSROOT/cvsspam.conf file to have:

$cvs_prog = "/usr/bin/cvs"

No dice though.  Your help is appreciated.

Thanks,

Matt


__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com


From dave at badgers-in-foil.co.uk  Wed Sep 10 16:37:34 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Wed, 10 Sep 2003 16:37:34 +0000
Subject: [cvsspam-devel] record_lastdir.rb: No such file or directory
In-Reply-To: <20030910162505.64654.qmail@web14306.mail.yahoo.com>
References: <20030910162505.64654.qmail@web14306.mail.yahoo.com>
Message-ID: <20030910163734.GA860@vhost.badgers-in-foil.co.uk>

Hello there,

On Wed, Sep 10, 2003 at 09:25:05AM -0700, Matt Raible wrote:
> cvs server: cannot exec /usr/local/cvsroot/CVSROOT/record_lastdir.rb: No such
> file or directory

If '/usr/local/cvsroot/CVSROOT/record_lastdir.rb' exists, check the
first line of the script and see if the ruby it specifies matches the
location ruby is installed in on your system.


dave


From matt at raibledesigns.com  Wed Sep 10 18:43:26 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Wed, 10 Sep 2003 11:43:26 -0700 (PDT)
Subject: [cvsspam-devel] record_lastdir.rb: No such file or directory
In-Reply-To: <20030910163734.GA860@vhost.badgers-in-foil.co.uk>
Message-ID: <20030910184327.55472.qmail@web14304.mail.yahoo.com>

That was it - ruby was in /usr/local/bin - so I created a symlink to /usr/bin.
I'm getting e-mails now - man this is a sweet product.  Nice job!

I'm getting the following warnings - should I just ignore them?

/usr/local/cvsroot/CVSROOT/cvsspam.rb:919: warning: global variable
`$viewcvsURL' not initialized
/usr/local/cvsroot/CVSROOT/cvsspam.rb:922: warning: global variable `$choraURL'
not initialized
/usr/local/cvsroot/CVSROOT/cvsspam.rb:924: warning: global variable
`$cvswebURL' not initialized

Thanks,

Matt


--- David Holroyd <dave badgers-in-foil.co.uk> wrote:
> Hello there,
> 
> On Wed, Sep 10, 2003 at 09:25:05AM -0700, Matt Raible wrote:
> > cvs server: cannot exec /usr/local/cvsroot/CVSROOT/record_lastdir.rb: No
> such
> > file or directory
> 
> If '/usr/local/cvsroot/CVSROOT/record_lastdir.rb' exists, check the
> first line of the script and see if the ruby it specifies matches the
> location ruby is installed in on your system.
> 
> 
> dave
> 
> _______________________________________________
> cvsspam-devel mailing list
> cvsspam-devel@lists.badgers-in-foil.co.uk
> http://lists.badgers-in-foil.co.uk/mailman/listinfo/cvsspam-devel
> 
> 


__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com


From dave at badgers-in-foil.co.uk  Thu Sep 11 09:38:09 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Thu, 11 Sep 2003 09:38:09 +0000
Subject: [cvsspam-devel] record_lastdir.rb: No such file or directory
In-Reply-To: <20030910184327.55472.qmail@web14304.mail.yahoo.com>
References: <20030910163734.GA860@vhost.badgers-in-foil.co.uk> <20030910184327.55472.qmail@web14304.mail.yahoo.com>
Message-ID: <20030911093809.GA14492@vhost.badgers-in-foil.co.uk>

On Wed, Sep 10, 2003 at 11:43:26AM -0700, Matt Raible wrote:
> That was it - ruby was in /usr/local/bin - so I created a symlink to /usr/bin.

Great.


> I'm getting e-mails now - man this is a sweet product.  Nice job!

Thanks!


> I'm getting the following warnings - should I just ignore them?
> 
> /usr/local/cvsroot/CVSROOT/cvsspam.rb:919: warning: global variable `$viewcvsURL' not initialized
> /usr/local/cvsroot/CVSROOT/cvsspam.rb:922: warning: global variable `$choraURL' not initialized
> /usr/local/cvsroot/CVSROOT/cvsspam.rb:924: warning: global variable `$cvswebURL' not initialized

It's safe to ignore them, but I should fix this.  Not sure why I've not
seen the same messages, tbh.


Thanks for pointing out the problem!

dave


From matt at raibledesigns.com  Thu Sep 11 17:22:57 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Thu, 11 Sep 2003 11:22:57 -0600
Subject: [cvsspam-devel] Error in collect_diffs
Message-ID: <004801c37888$bc360c90$6400a8c0@fatbastard>

I'm having issues with my cvs spam install because CVSROOT is not an
environment variable on my server.  So I altered collect_diffs.rb to
have:

$cvsroot = ENV["CVSROOT"] || "/usr/bin/cvs"

And then I reference $cvsroot wherever it's making a call to get the
environment variable.  Now I'm getting:

Checking in log4j-1.2.8/cvsspam-test.txt;
/repos/libs/log4j-1.2.8/cvsspam-test.txt,v  <--  cvsspam-test.txt
new revision: 1.5; previous revision: 1.4
done
/usr/bin/ruby: invalid option -
  (-h will show valid options)

Thanks,

Matt

Here's my full collect_diffs.rb file:

#!/usr/bin/ruby -w

# Part of CVSspam
#   http://www.badgers-in-foil.co.uk/projects/cvsspam/
# Copyright (c) David Holroyd

#  CVSROOT is /var/lib/cvs
#  ARGV is 'CVSROOT/foo/space dir space file,NONE,1.1 some,file,1.4,1.5'
#  ----
#  Update of /var/lib/cvs/CVSROOT/foo/space dir
#  In directory foil:/tmp/cvs-serv13059
#  
#  Modified Files:
#  	some,file 
#  Added Files:
#  	space file 
#  Log Message:
#  msg
#  


# Assumptions
# - file names do not contain newlines or single quotes


$tmpdir = ENV["TMPDIR"] || "/tmp"
$cvsroot = ENV["CVSROOT"] || "/usr/bin/cvs"
$dirtemplate = "#cvsspam.#{Process.getpgrp}.#{Process.uid}"

def find_data_dir
  Dir["#{$tmpdir}/#{$dirtemplate}-*"].each do |dir|
    stat = File.stat(dir)
    return dir if stat.owned?
  end
  nil
end


def blah(msg)
  if $debug
    puts "collect_diffs.rb: #{msg}"
  end
end


# Like IO.popen, but accepts multiple arguments like Kernel.exec
# (So no need to escape shell metacharacters)
def safer_popen(*args)
  IO.popen("-") do |pipe|
    if pipe==nil
      exec(*args)
    else
      yield pipe
    end
  end
end

class ChangeInfo
  def initialize(file, fromVer, toVer)
    @file, @fromVer, @toVer = file, fromVer, toVer
    if fromVer == toVer
      fail "'from' and 'to' versions should be different ('#{fromVer}')"
    end
  end
  attr_reader :file, :fromVer, :toVer
  def to_s
    "<ChangeInfo \"#{@file}\" #{@toVer}<--#{@fromVer}>"
  end

  def isAddition ; fromVer == 'NONE' end

  def isRemoval ; toVer == 'NONE' end

  #def isModification ; !(isAddition || isRemoval) end
end

$commitinfo_tags = nil

def get_commitinfo_tag(filename)
  if $commitinfo_tags.nil?
    return nil unless FileTest.exists?("#{$datadir}/commitinfo-tags")
    File.open("#{$datadir}/commitinfo-tags") do |file|
      $commitinfo_tags = Hash.new
      file.each_line do |line|
	line =~ /([^\t]+)\t(.+)/
	key = $2
	val = $1
	key.sub!(/^$cvsroot\//, '')
	$commitinfo_tags[key] = val
      end
    end
  end
  return $commitinfo_tags[filename]
end

def process_log(cvs_info)
  cvsroot = $cvsroot

  $datadir = find_data_dir()

  raise "missing data dir (#{$tmpdir}/#{$dirtemplate}-XXXXXX)" if
$datadir==nil

  line = $stdin.gets
  unless line =~ /^Update of (.+)/
    fail "Log preamble looks suspect (doesn't start 'Update of ...')"
  end

  # cvs_info comes from the command line, ultimately as the expansion of
the
  # %{sVv} in $CVSROOT/loginfo.  It isn't possible to parse this value
  # unambiguously, but we make an effort to get it right in as many
cases as
  # possible.

  $path = $1
  unless $path.slice(0,cvsroot.length) == cvsroot
    fail "CVSROOT ('#{cvsroot}') doesn't match log preamble
('#{$path}')"
  end

  $repository_path = $path.slice(cvsroot.length+1,
$path.length-cvsroot.length-1)
  unless cvs_info.slice(0, $repository_path.length+1) ==
"#{$repository_path} "
    fail "calculated repository path ('#{$repository_path}') doesn't
match start of command line arg ('#{cvs_info}')"
  end

  version_info = cvs_info.slice($repository_path.length,
                                cvs_info.length-$repository_path.length)

  changes = Array.new
  # make a list of changed files given on the command line
  while version_info.length>0
    if version_info.sub!(/^ (.+?),(NONE|[.0-9]+),(NONE|[.0-9]+)/, '') ==
nil
      fail "'#{version_info}' doesn't match ' <name>,<ver>,<ver> ...'"
    end
    changes << ChangeInfo.new($1, $2, $3)
  end

  # look for the start of the user's comment
  $stdin.each do |line|
    break if line =~ /^Log Message/
  end

  unless line =~ /^Log Message/
    fail "Input did not contain a 'Log Message:' entry"
  end

  File.open("#{$datadir}/logfile",
File::WRONLY|File::CREAT|File::APPEND) do |file|
    $stdin.each do |line|
      file.puts "#> #{line}"
    end

    changes.each do |change|

      # record version information
      file.puts "#V #{change.fromVer},#{change.toVer}"

      # note if the file is on a branch
      tag = nil
      if change.isRemoval
	tag = get_commitinfo_tag("#{$repository_path}/#{change.file}")
      else
        status = nil
        safer_popen($cvs_prog, "-nq", "status", change.file) do |io|
          status = io.read
        end
        fail "couldn't get cvs status: #{$!}" unless ($?>>8)==0

	if status =~ /^\s*Sticky Tag:\s*(.+) \(branch: +/m
	  tag = $1
	end
      end
      file.puts "#T #{tag}" unless tag.nil?

      diff_cmd = Array.new << $cvs_prog << "-nq" << "diff" << "-Nu"
      diff_cmd << "-kk" if $diff_ignore_keywords

      if change.isAddition
        file.write "#A "
        # cruft up a date in the distant past, when the file would not
have
        # existed, so that the diff will show all lines as added
        diff_cmd << "-D1/26/1977" << "-r#{change.toVer}"
      elsif change.isRemoval
        file.write "#R "
        # just specifying one version, cvs will diff between that
version and
        # the current version (will show all lines removed)
        diff_cmd << "-r#{change.fromVer}"
      else
        file.write "#M "
        diff_cmd << "-r#{change.fromVer}" << "-r#{change.toVer}"
      end
      file.puts "#{$repository_path}/#{change.file}"
      diff_cmd << change.file
      # do a cvs diff and place the output into our temp file
      blah("about to run #{diff_cmd.join(' ')}")
      safer_popen(*diff_cmd) do |pipe|
        # skip over cvs-diff's preamble
        pipe.each do |line|
          break if line =~ /^diff /
        end
        file.puts "#U #{line}"
        pipe.each do |line|
          file.puts "#U #{line}"
        end
      end
      # TODO: don't how to do this reliably on different systems...
      #fail "cvsdiff did not give exit status 1 for invocation:
#{diff_cmd.join(' ')}" unless ($?>>8)==1
    end
  end

end


def choose_operation(op)
  if op =~ / - New directory$/
    blah("No action taken on directory creation")
  elsif op =~ / - Imported sources$/
    blah("Imported not handled")
  else
    process_log(op)
    mailtest
  end
end

def mailtest
  lastdir = nil
  File.open("#{$datadir}/lastdir") do |file|
    lastdir = file.gets
  end
  if $path == lastdir
    blah("sending spam.  (I am #{$0})")
    # REVISIT: $0 will not contain the path to this script on all
systems
    cmd = File.dirname($0) + "/cvsspam.rb"
    unless system(cmd, "#{$datadir}/logfile", *$passthoughArgs)
      fail "problem running '#{cmd}'"
    end
    if $debug
      blah("leaving file #{$datadir}/logfile")
    else 
      File.unlink("#{$datadir}/logfile")
    end
    File.unlink("#{$datadir}/lastdir")
    File.unlink("#{$datadir}/commitinfo-tags") if
FileTest.exists?("#{$datadir}/commitinfo-tags")
    Dir.rmdir($datadir) unless $debug
  else
    blah("not spam time yet, #{$path}!=#{lastdir}")
  end
end

$config = "$cvsroot/CVSROOT/cvsspam.conf"
$cvs_prog = "cvs"
$debug = false
$diff_ignore_keywords = false
$task_keywords = []

unless ENV.has_key?('CVSROOT')
  fail "$CVSROOT not defined.  It should be when I am invoked from
CVSROOT/loginfo"
end


require 'getoptlong'

opts = GetoptLong.new(
  [ "--to",     "-t", GetoptLong::REQUIRED_ARGUMENT ],
  [ "--config", "-c", GetoptLong::REQUIRED_ARGUMENT ],
  [ "--debug",  "-d", GetoptLong::NO_ARGUMENT ],
  [ "--from",   "-u", GetoptLong::REQUIRED_ARGUMENT ]
)

# arguments to pass though to 'cvsspam.rb'
$passthoughArgs = Array.new
opts.each do |opt, arg|
  if ["--to", "--config", "--from"].include?(opt)
    $passthoughArgs << opt << arg
  end
  if ["--debug"].include?(opt)
    $passthoughArgs << opt
  end
  $config = arg if opt=="--config"
  $debug = true if opt == "--debug"
end

blah("CVSROOT is $cvsroot")
blah("ARGV is '#{ARGV.join(', ')}'")

if FileTest.exists?($config)
  def addHeader(name,val)
  end
  def addRecipient(who)
  end
  load $config
else
  blah("Config file '#{$config}' not found, ignoring")
end


if ARGV.length != 1
  $stderr.puts "Expected arguments missing"
  $stderr.puts "* You shouldn't run collect_diffs by hand, but from a
CVSROOT/loginfo entry *"
  $stderr.puts "Usage: collect_diffs.rb [ --to <email> ] [ --config
<file> ] %{sVv}"
  $stderr.puts "       (the sequence '%{sVv}' is expanded by CVS, when
found in CVSROOT/loginfo)"
  exit
end
choose_operation(ARGV[0])






From dave at badgers-in-foil.co.uk  Thu Sep 11 17:36:04 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Thu, 11 Sep 2003 17:36:04 +0000
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <004801c37888$bc360c90$6400a8c0@fatbastard>
References: <004801c37888$bc360c90$6400a8c0@fatbastard>
Message-ID: <20030911173604.GA19151@vhost.badgers-in-foil.co.uk>

On Thu, Sep 11, 2003 at 11:22:57AM -0600, Matt Raible wrote:
> I'm having issues with my cvs spam install because CVSROOT is not an
> environment variable on my server.  So I altered collect_diffs.rb to
> have:
> 
> $cvsroot = ENV["CVSROOT"] || "/usr/bin/cvs"
> 
> And then I reference $cvsroot wherever it's making a call to get the
> environment variable.  Now I'm getting:

Oops, I'd assumed that cvs would always set $CVSROOT to the 'right'
value, so that its children could pick it up.  Maybe this in only the
case when using cvs in pserver mode (which is what I've most
experience with).  Are you using a local repository in this case?


> /usr/bin/ruby: invalid option -
>   (-h will show valid options)

Looks suspiciously like the script has been converted to DOS format.
(That would add a <CR> to the end of the line, which ruby would see as
an extra character tagged onto the '-w' option, hence the odd-looking
error message.)


dave


From matt at raibledesigns.com  Thu Sep 11 17:59:14 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Thu, 11 Sep 2003 11:59:14 -0600
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <20030911173604.GA19151@vhost.badgers-in-foil.co.uk>
Message-ID: <005801c3788e$6c9f6360$6400a8c0@fatbastard>

Local repository: no, it's on a linux box in my office.  It is a
pserver.

DOS: yes, I did copy the file to my XP box to edit the file and do a
search/replace since it wasn't working in vi.  

Thinking that this might be a problem with the DOS format, I copied the
collect_diffs.rb out of the original distro and added the line:

ENV['CVSROOT'] = "/usr/bin/cvs"

Now I'm getting:

/repos/CVSROOT/collect_diffs.rb:101:in `process_log': missing data dir
(/tmp/#cvsspam.2777.508-XXXXXX) (RuntimeError)
	from /repos/CVSROOT/collect_diffs.rb:216:in `choose_operation'
	from /repos/CVSROOT/collect_diffs.rb:300

Does the /tmp directory need to have write permissions for the cvs user?

Thanks,

Matt

-----Original Message-----
From: cvsspam-devel-admin@lists.badgers-in-foil.co.uk
[mailto:cvsspam-devel-admin@lists.badgers-in-foil.co.uk] On Behalf Of
David Holroyd
Sent: Thursday, September 11, 2003 11:36 AM
To: cvsspam-devel@lists.badgers-in-foil.co.uk
Subject: Re: [cvsspam-devel] Error in collect_diffs


On Thu, Sep 11, 2003 at 11:22:57AM -0600, Matt Raible wrote:
> I'm having issues with my cvs spam install because CVSROOT is not an 
> environment variable on my server.  So I altered collect_diffs.rb to
> have:
> 
> $cvsroot = ENV["CVSROOT"] || "/usr/bin/cvs"
> 
> And then I reference $cvsroot wherever it's making a call to get the 
> environment variable.  Now I'm getting:

Oops, I'd assumed that cvs would always set $CVSROOT to the 'right'
value, so that its children could pick it up.  Maybe this in only the
case when using cvs in pserver mode (which is what I've most experience
with).  Are you using a local repository in this case?


> /usr/bin/ruby: invalid option -
>   (-h will show valid options)

Looks suspiciously like the script has been converted to DOS format.
(That would add a <CR> to the end of the line, which ruby would see as
an extra character tagged onto the '-w' option, hence the odd-looking
error message.)


dave

_______________________________________________
cvsspam-devel mailing list cvsspam-devel@lists.badgers-in-foil.co.uk
http://lists.badgers-in-foil.co.uk/mailman/listinfo/cvsspam-devel




From dave at badgers-in-foil.co.uk  Thu Sep 11 18:45:39 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Thu, 11 Sep 2003 18:45:39 +0000
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <005801c3788e$6c9f6360$6400a8c0@fatbastard>
References: <20030911173604.GA19151@vhost.badgers-in-foil.co.uk> <005801c3788e$6c9f6360$6400a8c0@fatbastard>
Message-ID: <20030911184539.GB19151@vhost.badgers-in-foil.co.uk>

On Thu, Sep 11, 2003 at 11:59:14AM -0600, Matt Raible wrote:
> Thinking that this might be a problem with the DOS format, I copied the
> collect_diffs.rb out of the original distro and added the line:
> 
> ENV['CVSROOT'] = "/usr/bin/cvs"

Ah, I missed that in your previous message; CVSROOT contains the location
of the repository, not the CVS executable.  Looking at the output
below, I guess you need the value "/repos"?

[In our pserver setup, nothing in the configuration of the server sets
$CVSROOT, it's just available when the script needs it.  Looking at the
latest CVS source, there's

  (void) sprintf (env, "%s=%s", CVSROOT_ENV, current_parsed_root->directory);
  (void) putenv (env);

in server.c, but maybe this isn't being used in the binaries you have.
What version of CVS does the server run?]


> Now I'm getting:
> 
> /repos/CVSROOT/collect_diffs.rb:101:in `process_log': missing data dir
> (/tmp/#cvsspam.2777.508-XXXXXX) (RuntimeError)
> 	from /repos/CVSROOT/collect_diffs.rb:216:in `choose_operation'
> 	from /repos/CVSROOT/collect_diffs.rb:300
> 
> Does the /tmp directory need to have write permissions for the cvs user?

Yes, the tmp dir needs to be writable by the owner of the cvs pserver
process.  Check that record_lastdir.rb is set up ok, as this is the
script that actually creates the directory for each commit.


dave


From matt at raibledesigns.com  Thu Sep 11 20:34:24 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Thu, 11 Sep 2003 14:34:24 -0600
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <20030911184539.GB19151@vhost.badgers-in-foil.co.uk>
Message-ID: <009701c378a4$1729c950$6400a8c0@fatbastard>

I think this was all caused by having settings for a project in loginfo
and then not having that same project configured in commitinfo.  But now
I can't get back to where it was working yesterday.

I'm getting the following errors:

/repos/CVSROOT/cvsspam.rb:916: warning: global variable `$viewcvsURL'
not initialized
/repos/CVSROOT/cvsspam.rb:919: warning: global variable `$choraURL' not
initialized
/repos/CVSROOT/cvsspam.rb:921: warning: global variable `$cvswebURL' not
initialized

Which should be fixed by hardcoding this line in cvsspam.rb:

$config = "/repos/cvsspam.conf"

And putting cvsspam.conf here and then configuring:

$cvs_prog = "/usr/bin/cvs"

But no dice, I'm still getting the above error messages.

Should I un-comment these variables to eliminate this message?

Sorry for all the trouble.

Matt


-----Original Message-----
From: cvsspam-devel-admin@lists.badgers-in-foil.co.uk
[mailto:cvsspam-devel-admin@lists.badgers-in-foil.co.uk] On Behalf Of
David Holroyd
Sent: Thursday, September 11, 2003 12:46 PM
To: cvsspam-devel@lists.badgers-in-foil.co.uk
Subject: Re: [cvsspam-devel] Error in collect_diffs


On Thu, Sep 11, 2003 at 11:59:14AM -0600, Matt Raible wrote:
> Thinking that this might be a problem with the DOS format, I copied 
> the collect_diffs.rb out of the original distro and added the line:
> 
> ENV['CVSROOT'] = "/usr/bin/cvs"

Ah, I missed that in your previous message; CVSROOT contains the
location of the repository, not the CVS executable.  Looking at the
output below, I guess you need the value "/repos"?

[In our pserver setup, nothing in the configuration of the server sets
$CVSROOT, it's just available when the script needs it.  Looking at the
latest CVS source, there's

  (void) sprintf (env, "%s=%s", CVSROOT_ENV,
current_parsed_root->directory);
  (void) putenv (env);

in server.c, but maybe this isn't being used in the binaries you have.
What version of CVS does the server run?]


> Now I'm getting:
> 
> /repos/CVSROOT/collect_diffs.rb:101:in `process_log': missing data dir
> (/tmp/#cvsspam.2777.508-XXXXXX) (RuntimeError)
> 	from /repos/CVSROOT/collect_diffs.rb:216:in `choose_operation'
> 	from /repos/CVSROOT/collect_diffs.rb:300
> 
> Does the /tmp directory need to have write permissions for the cvs 
> user?

Yes, the tmp dir needs to be writable by the owner of the cvs pserver
process.  Check that record_lastdir.rb is set up ok, as this is the
script that actually creates the directory for each commit.


dave

_______________________________________________
cvsspam-devel mailing list cvsspam-devel@lists.badgers-in-foil.co.uk
http://lists.badgers-in-foil.co.uk/mailman/listinfo/cvsspam-devel




From dave at badgers-in-foil.co.uk  Thu Sep 11 20:53:14 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Thu, 11 Sep 2003 20:53:14 +0000
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <009701c378a4$1729c950$6400a8c0@fatbastard>
References: <20030911184539.GB19151@vhost.badgers-in-foil.co.uk> <009701c378a4$1729c950$6400a8c0@fatbastard>
Message-ID: <20030911205314.GA23708@vhost.badgers-in-foil.co.uk>

--C7zPtVaVf+AK4Oqc
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Thu, Sep 11, 2003 at 02:34:24PM -0600, Matt Raible wrote:
> I'm getting the following errors:
> 
> /repos/CVSROOT/cvsspam.rb:916: warning: global variable `$viewcvsURL'
> not initialized
> /repos/CVSROOT/cvsspam.rb:919: warning: global variable `$choraURL' not
> initialized
> /repos/CVSROOT/cvsspam.rb:921: warning: global variable `$cvswebURL' not
> initialized

I think maybe older versions of Ruby (that I've been using) are happy to
treat undefined variables as nil-valued.  Try the attached patch, which
explicitly initializes the problem variables.


> $config = "/repos/cvsspam.conf"

If you place your config at "/repos/CVSROOT/cvsspam.conf", it should be
picked up automatically.


> Sorry for all the trouble.

No trouble!


dave

--C7zPtVaVf+AK4Oqc
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename="fix_global_var_not_initialized-patch.diff"

Index: cvsspam.rb
===================================================================
RCS file: /var/lib/cvs/cvsspam/cvsspam.rb,v
retrieving revision 1.37
diff -u -r1.37 cvsspam.rb
--- cvsspam.rb	7 Sep 2003 18:10:50 -0000	1.37
+++ cvsspam.rb	11 Sep 2003 20:41:50 -0000
@@ -878,6 +878,9 @@
 $bugzillaURL = nil
 $jiraURL = nil
 $ticketURL = nil
+$viewcvsURL = nil
+$choraURL = nil
+$cvswebURL = nil
 $from_address = nil
 $subjectPrefix = nil
 $files_in_subject = false;

--C7zPtVaVf+AK4Oqc--


From matt at raibledesigns.com  Thu Sep 11 21:04:03 2003
From: matt at raibledesigns.com (Matt Raible)
Date: Thu, 11 Sep 2003 15:04:03 -0600
Subject: [cvsspam-devel] Error in collect_diffs
In-Reply-To: <20030911205314.GA23708@vhost.badgers-in-foil.co.uk>
Message-ID: <00a601c378a8$3ff07d30$6400a8c0@fatbastard>

This worked - thanks!  I'm using the latest version of Ruby, so that
explains the problem.

Matt

-----Original Message-----
From: cvsspam-devel-admin@lists.badgers-in-foil.co.uk
[mailto:cvsspam-devel-admin@lists.badgers-in-foil.co.uk] On Behalf Of
David Holroyd
Sent: Thursday, September 11, 2003 2:53 PM
To: cvsspam-devel@lists.badgers-in-foil.co.uk
Subject: Re: [cvsspam-devel] Error in collect_diffs


On Thu, Sep 11, 2003 at 02:34:24PM -0600, Matt Raible wrote:
> I'm getting the following errors:
> 
> /repos/CVSROOT/cvsspam.rb:916: warning: global variable `$viewcvsURL' 
> not initialized
> /repos/CVSROOT/cvsspam.rb:919: warning: global variable `$choraURL' 
> not initialized
> /repos/CVSROOT/cvsspam.rb:921: warning: global variable `$cvswebURL' 
> not initialized

I think maybe older versions of Ruby (that I've been using) are happy to
treat undefined variables as nil-valued.  Try the attached patch, which
explicitly initializes the problem variables.


> $config = "/repos/cvsspam.conf"

If you place your config at "/repos/CVSROOT/cvsspam.conf", it should be
picked up automatically.


> Sorry for all the trouble.

No trouble!


dave




From dave at badgers-in-foil.co.uk  Mon Sep 22 22:01:36 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Mon, 22 Sep 2003 22:01:36 +0000
Subject: [cvsspam-devel] [ANN] CVSspam 0.2.8
Message-ID: <20030922220135.GA12273@vhost.badgers-in-foil.co.uk>

The latest version is now available at,

  http://www.badgers-in-foil.co.uk/projects/cvsspam/

Mail can now be sent using Net::SNMP, for people with no MTA on their
CVS server.  ViewCVS support has been extended to handle confugurations
using multiple repositories.  The documentation was reoganised a bit,
and some undocumented features were noted.  A few 'uninitialized
variable' warnings have been fixed.


Long ChangeLog:

2003-09-22 22:18  dave

	* cvsspam.rb: When I changed the header name, I forgot to put
	'CVSspam' into the value

2003-09-21 13:47  dave

	* cvsspam.rb: Fix for the name of the ViewCVS parameter, from
	Dmitry.

2003-09-11 22:29  dave

	* CREDITS: Matt pointed out some uninitialised variables were
	causing problems in the latest Ruby (that being 1.8.0).

2003-09-11 22:15  dave

	* cvsspam.rb: Explicitly initialise variables to prevent warning
	messages

2003-09-07 19:10  dave

	* cvsspam.rb: "X-Mailer" is a more 'standard' non-standard
	header.

2003-09-07 19:09  dave

	* cvsspam.rb: Bump version number in preparation for a new
	release.

2003-09-07 19:08  dave

	* cvsspam-doc.xml: Re-orded sections in the documentation to
	follow what I think users will want.  Noted that
	'/etc/cvsspam/cvsspam.conf' will be used if available.

2003-09-07 19:05  dave

	* collect_diffs.rb: Look for cvsspam.conf under /etc/cvsspam/ as
	well as in $CVSROOT/CVSROOT/.  The derived filename is passed to
	cvsspam.rb, to save it duplicating this logic.

2003-09-07 15:31  dave

	* CREDITS, collect_diffs.rb, cvsspam-doc.xml, cvsspam.rb:
	Support	for multiple repositories with ViewCVS, suggested by
	Dmitry V. abanin

2003-09-06 23:53  dave

	* collect_diffs.rb: Debug should go to stderr, not stdout

2003-09-06 23:25  dave

	* cvsspam-doc.xml: Caution users against breaking their remote
	repositories.

2003-09-06 22:27  dave

	* cvsspam-doc.xml: Make this an SGML, rather than an XML
	document. The docbook toolchain I have installed (Debian
	unstable) seems to be broken for XML documents.

2003-09-01 00:41  dave

	* cvsspam-doc.xml: Cleanup, and document the '--from' option

2003-07-26 16:09  dave

	* cvsspam.conf, cvsspam.rb: Support for SMTP, instead of
	$sendmail

2003-07-15 23:33  dave

	* cvsspam.rb: Start factoring out HTML generating code from the
	rest of the file


From daniel.pfuhl at imk.fraunhofer.de  Tue Sep 23 13:10:02 2003
From: daniel.pfuhl at imk.fraunhofer.de (Daniel Pfuhl)
Date: Tue, 23 Sep 2003 15:10:02 +0200
Subject: [cvsspam-devel] missing data dir
Message-ID: <3F70462A.2050304@imk.fraunhofer.de>

hi

after installing and configuring cvsspam like described in the docs
i got the following error message after committing a change to a
file.

does anyone has an idea about how to avoid this ?

---------

cvs -z5 commit -m "no message" collect_diffs.rb loginfo (in directory 
C:\data\cvs\CAT1\CVSROOT\)
Checking in collect_diffs.rb;
/var/cvs/daniel/CVSROOT/collect_diffs.rb,v  <--  collect_diffs.rb
new revision: 1.3; previous revision: 1.2
done
Checking in loginfo;
/var/cvs/daniel/CVSROOT/loginfo,v  <--  loginfo
new revision: 1.4; previous revision: 1.3
done
/var/cvs/daniel/CVSROOT/collect_diffs.rb:100:in `process_log': missing 
data dir (/tmp/#cvsspam.24380.527-XXXXXX) (RuntimeError)
	from /var/cvs/daniel/CVSROOT/collect_diffs.rb:215:in `choose_operation'
	from /var/cvs/daniel/CVSROOT/collect_diffs.rb:314
cvs server: Rebuilding administrative file database

***** CVS exited normally with code 0 *****

----------------

thanks in advance !!


-- 
daniel




From dave at badgers-in-foil.co.uk  Tue Sep 23 13:39:12 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Tue, 23 Sep 2003 13:39:12 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <3F70462A.2050304@imk.fraunhofer.de>
References: <3F70462A.2050304@imk.fraunhofer.de>
Message-ID: <20030923133912.GA25055@vhost.badgers-in-foil.co.uk>

Hello,

On Tue, Sep 23, 2003 at 03:10:02PM +0200, Daniel Pfuhl wrote:
> cvs -z5 commit -m "no message" collect_diffs.rb loginfo (in directory 
> C:\data\cvs\CAT1\CVSROOT\)
> Checking in collect_diffs.rb;
> /var/cvs/daniel/CVSROOT/collect_diffs.rb,v  <--  collect_diffs.rb
> new revision: 1.3; previous revision: 1.2
> done
> Checking in loginfo;
> /var/cvs/daniel/CVSROOT/loginfo,v  <--  loginfo
> new revision: 1.4; previous revision: 1.3
> done
> /var/cvs/daniel/CVSROOT/collect_diffs.rb:100:in `process_log': missing 
> data dir (/tmp/#cvsspam.24380.527-XXXXXX) (RuntimeError)
> 	from /var/cvs/daniel/CVSROOT/collect_diffs.rb:215:in 
> 	`choose_operation'
> 	from /var/cvs/daniel/CVSROOT/collect_diffs.rb:314
> cvs server: Rebuilding administrative file database


Have you got the 'record_lastdir.rb' script configured properly in
'CVSROOT/commitinfo'?  (This is the script that creates the temporary
dir.)

If you do have an entry for 'record_lastdir.rb', ensure that the
expression on left matches the one you specified for the module in
'loginfo'.


dave


From dave at badgers-in-foil.co.uk  Tue Sep 23 14:18:35 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Tue, 23 Sep 2003 14:18:35 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <3F705425.5030501@imk.fraunhofer.de>
References: <3F70462A.2050304@imk.fraunhofer.de> <20030923133912.GA25055@vhost.badgers-in-foil.co.uk> <3F705425.5030501@imk.fraunhofer.de>
Message-ID: <20030923141835.GB25055@vhost.badgers-in-foil.co.uk>

On Tue, Sep 23, 2003 at 04:09:41PM +0200, Daniel Pfuhl wrote:
> hi dave,
> 
> how do i configure all the scripts in the right way ?
> at this time i had only the collect_diffs.rb configured
> as follows:
> 
> ALL /var/cvs/daniel/CVSROOT/collect_diffs.rb --to 
> daniel.pfuhl@imk.fraunhofer.de %{sVv}
> 
> do i have to add the same for all the thre scripts ?
> 
> btw: i have configured this in loginfo - this shouldn't
> be a problem, is it ?

You have to do all the configuration before commiting.  Given your
example values, I'd suggest you use something like this:


== in commitinfo ==

ALL /var/cvs/daniel/CVSROOT/record_lastdir.rb



== in loginfo ==

ALL /var/cvs/daniel/CVSROOT/collect_diffs.rb --to daniel.pfuhl@imk.fraunhofer.de %{sVv}


You also need to add all three of the scripts to CVSROOT and then commit
everything once all is in place.


[Note that if you have no other way of accessing /var/cvs/daniel/ on the
server than by using CVS itself, using the 'ALL' match is dangerous,
because if there is a problem with the script, you will no longer ba
able to commit to teh repository.]



Hope that helps!

dave


From richard.bos at xs4all.nl  Tue Sep 23 19:25:50 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Tue, 23 Sep 2003 21:25:50 +0200
Subject: [cvsspam-devel] missing data dir
Message-ID: <200309232125.50064.radoeka@xs4all.nl>

Hello,

I get the same error.  Some information about my setup:

I have several lines in commit- and loginfo, both ending in ALL.
cvsspam works fine for commits that are being performed via ALL.  
For the other entries the commit fails (is it possible that the regular 
expression fails in collect_diffs.rb?).

My commitinfo:
sqr     /home/cvs/bin/record_lastdir.rb
sucont  /home/cvs/bin/record_lastdir.rb
suverif /home/cvs/bin/record_lastdir.rb

ALL     /home/cvs/bin/record_lastdir.rb


My loginfo file:
sqr     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ron@blah.net   %{sVv}

sucont  /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ruin@blah.net %{sVv}

suverif /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ron@blah.net --to heid@blah.net %{sVv}

ALL     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to bos@blah.net %{sVv}


The modules directory looks as follows:
web/sqr
tools/sucont
tools/suverif

The regular expression I used does not start at CVSROOT.  Is that required?
The *info entries are tab seperated.

Did I make a configure mistake or is something else wrong?

>On Tue, Sep 23, 2003 at 04:09:41PM +0200, Daniel Pfuhl wrote:
>> hi dave,
>>
>> how do i configure all the scripts in the right way ?
>> at this time i had only the collect_diffs.rb configured
>> as follows:
>>
>> ALL /var/cvs/daniel/CVSROOT/collect_diffs.rb --to
>> daniel.pfuhl@imk.fraunhofer.de %{sVv}
>>
>> do i have to add the same for all the thre scripts ?
>>
>> btw: i have configured this in loginfo - this shouldn't
>> be a problem, is it ?
>
>You have to do all the configuration before commiting.  Given your
>example values, I'd suggest you use something like this:
>
>
>== in commitinfo ==
>
>ALL /var/cvs/daniel/CVSROOT/record_lastdir.rb
>
>
>
>== in loginfo ==
>
>ALL /var/cvs/daniel/CVSROOT/collect_diffs.rb --to 
daniel.pfuhl@imk.fraunhofer.de %{sVv}
>
>
>You also need to add all three of the scripts to CVSROOT and then commit
>everything once all is in place.
>
>
>[Note that if you have no other way of accessing /var/cvs/daniel/ on the
>server than by using CVS itself, using the 'ALL' match is dangerous,
>because if there is a problem with the script, you will no longer ba
>able to commit to teh repository.]

-- 
Richard Bos
Without a home the journey is endless



From dave at badgers-in-foil.co.uk  Wed Sep 24 10:00:21 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Wed, 24 Sep 2003 10:00:21 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <200309232125.50064.radoeka@xs4all.nl>
References: <200309232125.50064.radoeka@xs4all.nl>
Message-ID: <20030924100021.GA8894@vhost.badgers-in-foil.co.uk>

Hi,

On Tue, Sep 23, 2003 at 09:25:50PM +0200, Richard Bos wrote:
> I have several lines in commit- and loginfo, both ending in ALL.
> cvsspam works fine for commits that are being performed via ALL.  
> For the other entries the commit fails (is it possible that the regular 
> expression fails in collect_diffs.rb?).
> 
> My commitinfo:
> sqr     /home/cvs/bin/record_lastdir.rb
> sucont  /home/cvs/bin/record_lastdir.rb
> suverif /home/cvs/bin/record_lastdir.rb
> 
> ALL     /home/cvs/bin/record_lastdir.rb
> 
> 
> My loginfo file:
> sqr     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ron@blah.net   %{sVv}
> 
> sucont  /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ruin@blah.net %{sVv}
> 
> suverif /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to ron@blah.net --to heid@blah.net %{sVv}
> 
> ALL     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to bos@blah.net %{sVv}

I the problem may be that you're using 'ALL', but you need to use
'DEFAULT'.  I think that if you specify hooks for the 'ALL' match, these
will be run *in addition* to any other scripts specified.

Does email get sent, even when you see the error message?


If that's not the issue, try changing the commitinfo entries so that
they just run 'echo Hello World', and see if that appears on your
console when you do a commit.  (I'd expect record_lastdir.rb to give you
an error if it couldn't create the data-dir in the first place, so maybe
it isn't even getting executed?)


> The modules directory looks as follows:
> web/sqr
> tools/sucont
> tools/suverif
> 
> The regular expression I used does not start at CVSROOT.  Is that required?
> The *info entries are tab seperated.

CVS doesn't require it, but in my experience the most restrictive
possible expression should be used.  Consider that this text may appear
as a substring of a file's path in some other module:  If you create
a file called 'sucont.txt' in 'tools/suverif', you might get unexpected
results.


I'd use expressions like this in the config files,

  ^web/sqr
  ^tools/sucont
  ^tools/suverif

(However, I don't think CVSspam will mind the setup you've already got.)


> Did I make a configure mistake or is something else wrong?

Well, let us know if changing ALL to DEFAULT helps.


dave


From dave at badgers-in-foil.co.uk  Wed Sep 24 10:07:19 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Wed, 24 Sep 2003 10:07:19 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <200309232125.50064.radoeka@xs4all.nl>
References: <200309232125.50064.radoeka@xs4all.nl>
Message-ID: <20030924100718.GB8894@vhost.badgers-in-foil.co.uk>

On Tue, Sep 23, 2003 at 09:25:50PM +0200, Richard Bos wrote:
> For the other entries the commit fails (is it possible that the regular 
> expression fails in collect_diffs.rb?).

Oh, I forgot to say; could you post the text of the error message too?


thanks,
dave


From richard.bos at xs4all.nl  Wed Sep 24 20:51:16 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Wed, 24 Sep 2003 22:51:16 +0200
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <20030924100718.GB8894@vhost.badgers-in-foil.co.uk>
References: <200309232125.50064.radoeka@xs4all.nl> <20030924100718.GB8894@vhost.badgers-in-foil.co.uk>
Message-ID: <200309242251.16298.radoeka@xs4all.nl>

Op woensdag 24 september 2003 12:07, schreef David Holroyd:
> On Tue, Sep 23, 2003 at 09:25:50PM +0200, Richard Bos wrote:
> > For the other entries the commit fails (is it possible that the regular
> > expression fails in collect_diffs.rb?).
>
> Oh, I forgot to say; could you post the text of the error message too?

Ah, that is the easy part:

/home/cvs/CVS/tools/suverif/tessum,v  <--  tessum
new revision: 1.6; previous revision: 1.5
done
/home/cvs/bin/collect_diffs.rb:100:in `process_log': missing data dir 
(/tmp/#cvsspam.4129.503-XXXXXX) (RuntimeError)
        from /home/cvs/bin/collect_diffs.rb:215:in `choose_operation'
        from /home/cvs/bin/collect_diffs.rb:314


-- 
Richard Bos
Without a home the journey is endless



From richard.bos at xs4all.nl  Wed Sep 24 21:21:38 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Wed, 24 Sep 2003 23:21:38 +0200
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <20030924100021.GA8894@vhost.badgers-in-foil.co.uk>
References: <200309232125.50064.radoeka@xs4all.nl> <20030924100021.GA8894@vhost.badgers-in-foil.co.uk>
Message-ID: <200309242321.38911.radoeka@xs4all.nl>

Hello David,

Op woensdag 24 september 2003 12:00, schreef David Holroyd:
> Hi,
>
> On Tue, Sep 23, 2003 at 09:25:50PM +0200, Richard Bos wrote:
> > I have several lines in commit- and loginfo, both ending in ALL.
> > cvsspam works fine for commits that are being performed via ALL.
> > For the other entries the commit fails (is it possible that the regular
> > expression fails in collect_diffs.rb?).
> >
> > My commitinfo:
> > sqr     /home/cvs/bin/record_lastdir.rb
> > sucont  /home/cvs/bin/record_lastdir.rb
> > suverif /home/cvs/bin/record_lastdir.rb
> >
> > ALL     /home/cvs/bin/record_lastdir.rb
> >
> >
> > My loginfo file:
> > sqr     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to
> > ron@blah.net   %{sVv}
> >
> > sucont  /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to
> > ruin@blah.net %{sVv}
> >
> > suverif /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to
> > ron@blah.net --to heid@blah.net %{sVv}
> >
> > ALL     /home/cvs/bin/collect_diffs.rb --from $USER@blah.net --to
> > bos@blah.net %{sVv}
>
> I the problem may be that you're using 'ALL', but you need to use
> 'DEFAULT'.  I think that if you specify hooks for the 'ALL' match, these
> will be run *in addition* to any other scripts specified.

I really want it like this...  In this case I don't need to list myself in 
every line.  That is what I should do in case I would use DEFAULT isn't it?

> Does email get sent, even when you see the error message?

Yes and no.  Email is sent via the ALL line, but not via the regular 
expressions.  The commit succeeds successfully as well.

> If that's not the issue, try changing the commitinfo entries so that
> they just run 'echo Hello World', and see if that appears on your
> console when you do a commit.  (I'd expect record_lastdir.rb to give you
> an error if it couldn't create the data-dir in the first place, so maybe
> it isn't even getting executed?)
>
> > The modules directory looks as follows:
> > web/sqr
> > tools/sucont
> > tools/suverif
> >
> > The regular expression I used does not start at CVSROOT.  Is that
> > required? The *info entries are tab seperated.
>
> CVS doesn't require it, but in my experience the most restrictive
> possible expression should be used.  Consider that this text may appear
> as a substring of a file's path in some other module:  If you create
> a file called 'sucont.txt' in 'tools/suverif', you might get unexpected
> results.
>
>
> I'd use expressions like this in the config files,
>
>   ^web/sqr
>   ^tools/sucont
>   ^tools/suverif
>
> (However, I don't think CVSspam will mind the setup you've already got.)

I will try this out.

> > Did I make a configure mistake or is something else wrong?
>
> Well, let us know if changing ALL to DEFAULT helps.

See my comment above.  Should I really need to change to DEFAULT, it should be 
really necessary as cvs allows it.....

-- 
Richard Bos
Without a home the journey is endless



From dave at badgers-in-foil.co.uk  Wed Sep 24 22:06:17 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Wed, 24 Sep 2003 22:06:17 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <200309242321.38911.radoeka@xs4all.nl>
References: <200309232125.50064.radoeka@xs4all.nl> <20030924100021.GA8894@vhost.badgers-in-foil.co.uk> <200309242321.38911.radoeka@xs4all.nl>
Message-ID: <20030924220617.GA18583@vhost.badgers-in-foil.co.uk>

On Wed, Sep 24, 2003 at 11:21:38PM +0200, Richard Bos wrote:
> > I the problem may be that you're using 'ALL', but you need to use
> > 'DEFAULT'.  I think that if you specify hooks for the 'ALL' match, these
> > will be run *in addition* to any other scripts specified.
> 
> I really want it like this...  In this case I don't need to list myself in 
> every line.  That is what I should do in case I would use DEFAULT isn't it?
> 
> > Does email get sent, even when you see the error message?
> 
> Yes and no.  Email is sent via the ALL line, but not via the regular 
> expressions.  The commit succeeds successfully as well.

Ah, I see now.

Well, this presents some implementation difficulties:  The
first script (run from commitinfo) creates some files, and the second
script (run from loginfo) removes them when we're done.  Since
the scripts are listed twice, the 'second' invocation of collect_diffs.rb
will happen after this cleanup has happened, hence the error.

We could implement an extra commandline switch to say "don't delete
the data files just yet".  That's not so elegant, but then having to
hook into both commitinfo and loginfo is a (necessary) hack, in any case.

If you simply want the convenience of not having to specify your own
address for everything (and aren't bothered exactly how it's
implemented) try the addRecipient[1] option in your cvsspam.conf.


What do you think?


dave


[1] http://www.badgers-in-foil.co.uk/projects/cvsspam/cvsspam-doc.html#AEN107


From richard.bos at xs4all.nl  Wed Sep 24 22:08:34 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Thu, 25 Sep 2003 00:08:34 +0200
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <20030924220617.GA18583@vhost.badgers-in-foil.co.uk>
References: <200309232125.50064.radoeka@xs4all.nl> <200309242321.38911.radoeka@xs4all.nl> <20030924220617.GA18583@vhost.badgers-in-foil.co.uk>
Message-ID: <200309250008.34816.radoeka@xs4all.nl>

Op donderdag 25 september 2003 00:06, schreef David Holroyd:
> On Wed, Sep 24, 2003 at 11:21:38PM +0200, Richard Bos wrote:
> > > I the problem may be that you're using 'ALL', but you need to use
> > > 'DEFAULT'.  I think that if you specify hooks for the 'ALL' match,
> > > these will be run *in addition* to any other scripts specified.
> >
> > I really want it like this...  In this case I don't need to list myself
> > in every line.  That is what I should do in case I would use DEFAULT
> > isn't it?
> >
> > > Does email get sent, even when you see the error message?
> >
> > Yes and no.  Email is sent via the ALL line, but not via the regular
> > expressions.  The commit succeeds successfully as well.
>
> Ah, I see now.
>
> Well, this presents some implementation difficulties:  The
> first script (run from commitinfo) creates some files, and the second
> script (run from loginfo) removes them when we're done.  Since
> the scripts are listed twice, the 'second' invocation of collect_diffs.rb
> will happen after this cleanup has happened, hence the error.
>
> We could implement an extra commandline switch to say "don't delete
> the data files just yet".  That's not so elegant, but then having to
> hook into both commitinfo and loginfo is a (necessary) hack, in any case.
>
> If you simply want the convenience of not having to specify your own
> address for everything (and aren't bothered exactly how it's
> implemented) try the addRecipient[1] option in your cvsspam.conf.

If possible try to prevent the error or get an info message stating what might 
be the error (ALL in this case).  I'll see if I can implement the DEFAULT 
configuration and adding myself to the other lines in commit and loginfo.

-- 
Richard Bos
Without a home the journey is endless



From richard.bos at xs4all.nl  Thu Sep 25 20:20:46 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Thu, 25 Sep 2003 22:20:46 +0200
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <20030924100021.GA8894@vhost.badgers-in-foil.co.uk>
References: <200309232125.50064.radoeka@xs4all.nl> <20030924100021.GA8894@vhost.badgers-in-foil.co.uk>
Message-ID: <200309252220.47005.radoeka@xs4all.nl>

Op woensdag 24 september 2003 12:00, schreef David Holroyd:
> I'd use expressions like this in the config files,
>
>   ^web/sqr
>   ^tools/sucont
>   ^tools/suverif
>
> (However, I don't think CVSspam will mind the setup you've already got.)

I changed to the schema above and it still works (of course :)

> > Did I make a configure mistake or is something else wrong?
>
> Well, let us know if changing ALL to DEFAULT helps.

Yes, no error messages, the right recipients are addressed and the change 
seems to be committed.

1 more question, why is in the underneath example the "file#1" string not a 
url to the file it self?  It now seems to point to nowhere.

Commit in *5ess/updodd* on MAIN
ToDo <#file1>   +1              1.2
<http://.....


-- 
Richard Bos
Without a home the journey is endless



From dave at badgers-in-foil.co.uk  Fri Sep 26 13:31:54 2003
From: dave at badgers-in-foil.co.uk (David Holroyd)
Date: Fri, 26 Sep 2003 13:31:54 +0000
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <200309252220.47005.radoeka@xs4all.nl>
References: <200309232125.50064.radoeka@xs4all.nl> <20030924100021.GA8894@vhost.badgers-in-foil.co.uk> <200309252220.47005.radoeka@xs4all.nl>
Message-ID: <20030926133153.GA18305@vhost.badgers-in-foil.co.uk>

On Thu, Sep 25, 2003 at 10:20:46PM +0200, Richard Bos wrote:
> 1 more question, why is in the underneath example the "file#1" string not a 
> url to the file it self?  It now seems to point to nowhere.
> 
> Commit in *5ess/updodd* on MAIN
> ToDo <#file1>   +1              1.2
> <http://.....

The filenames in the top-of-message index are internal links to the point
in the email where that file's diff is.  Internal links work in some mail
clients, but not in others.


dave


From richard.bos at xs4all.nl  Fri Sep 26 19:25:45 2003
From: richard.bos at xs4all.nl (Richard Bos)
Date: Fri, 26 Sep 2003 21:25:45 +0200
Subject: [cvsspam-devel] missing data dir
In-Reply-To: <20030926133153.GA18305@vhost.badgers-in-foil.co.uk>
References: <200309232125.50064.radoeka@xs4all.nl> <200309252220.47005.radoeka@xs4all.nl> <20030926133153.GA18305@vhost.badgers-in-foil.co.uk>
Message-ID: <200309262125.45498.radoeka@xs4all.nl>

Op vrijdag 26 september 2003 15:31, schreef David Holroyd:
> On Thu, Sep 25, 2003 at 10:20:46PM +0200, Richard Bos wrote:
> > 1 more question, why is in the underneath example the "file#1" string not
> > a url to the file it self?  It now seems to point to nowhere.
> >
> > Commit in *5ess/updodd* on MAIN
> > ToDo <#file1>   +1              1.2
> > <http://.....
>
> The filenames in the top-of-message index are internal links to the point
> in the email where that file's diff is.  Internal links work in some mail
> clients, but not in others.

I've seen it in action now, it's nice :)

-- 
Richard Bos
Without a home the journey is endless