CHI works with memcached - http://cpan.uwinnipeg.ca/htdocs/CHI/CHI/
Driver/Memcached.html :)
I think your proposal sounds very useful!
Jon
On Jan 29, 2008, at 1:52 PM, Justin Ellison wrote:
> Hi list,
>
> First, let me apologize for sending my last email to the list - I
> should
> have chec
John-
Would you mind elaborating on how __PACKAGE__->meta and ->setup work?
I was working on new ways to automagically call/include my rose
objects ( ie: include an entire perl namespace at once, not each
individual file ) and hijack the setup registry to push some
information elsewhere.
i
On Nov 14, 2007, at 3:56 PM, John Siracusa wrote:
> On Nov 14, 2007 2:59 PM, Darren Duncan <[EMAIL PROTECTED]>
> wrote:
>> Or I see something about a company that does online gambling.
>
> That's the one :)
"This position requires does require occasionally lifting up to 20
(twenty) pounds"
> On 11/6/07 5:18 PM, Jonathan Swartz wrote:
>> On Oct 30, 2007, at 7:54 AM, John Siracusa wrote:
>>> Can you edit QueryBuilder.pm to see if you can at least work
>>> around it? For
>>> example, use $column->name ne $method instead of relying on
>&g
On Oct 30, 2007, at 7:54 AM, John Siracusa wrote:
> On 10/30/07, Jonathan Swartz <[EMAIL PROTECTED]> wrote:
>> $column is a Rose::DB::Object::Metadata::Column::Timestamp, according
>> to the error. The errors originate from all sorts of operations, but
>> they al
On Oct 30, 2007, at 7:54 AM, John Siracusa wrote:
> On 10/30/07, Jonathan Swartz <[EMAIL PROTECTED]> wrote:
>> $column is a Rose::DB::Object::Metadata::Column::Timestamp, according
>> to the error. The errors originate from all sorts of operations, but
>> they al
On Oct 29, 2007, at 5:08 PM, John Siracusa wrote:
> On 10/29/07 7:41 PM, Jonathan Swartz wrote:
>> A few days ago we started getting this error intermittently in
>> QueryBuilder::build_select:
>>
>> error_log.2007-10-29:[Mon Oct 29 15:29:37 2007] [error] [client
A few days ago we started getting this error intermittently in
QueryBuilder::build_select:
error_log.2007-10-29:[Mon Oct 29 15:29:37 2007] [error] [client
69.26.166.38] Operation "ne": no method found, left argument has no
overloaded magic, right argument in overloaded package
Rose::D
Anyone else get this when updating?
t/setup...NOK 3/4
# Failed test 'ROSEDBRC 2'
# in t/setup.t at line 22.
# got: 'localhost'
# expected: 'othervalue'
# Looks like you failed 1 test of 4.
t/setup...dubious
Test returned status 1 (wstat 25
Hi,
I find Rose::DB's reference counting of DBI handles confusing, and
would like to effectively neutralize it for my entire project. For
example I want
$dbh = Rose::DB->new()->dbh
to work even though it is incorrect (because it is such a common
mistake to make). I'm happy to let DBI j
s sort of stuff for
me...
On Aug 11, 2007, at 9:08 PM, John Siracusa wrote:
> On 8/11/07, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> so let me rephrase this question:
>> i thought strip() wasn't supposed to remove fkeys/rels ?
>> it seems
>> t
e for
writing with_objects instead of with. it was correct on my code,
just not on the posting.
// Jonathan Vanasco
Founder/President - FindMeOn
Fonder/CEO - Road
using .765
has_loaded_related ( helpers ) keeps returning false, when the
related item is loaded via a manager with_objects call ( and i know
its via a data::dumper profile )
$obj= $class->load( with_objects=>['subclass'] );
print "this always returns false " . $obj->has_loade
Kevin-
I usually get the effect you want in a slightly different way
a) I have
Oasis::Model::Asset::Object::Helpers
b) but I might also have
Oasis::Model::Asset::Object::Helpers_SpecifcNameHere
c) when i need a function in a rose object, i just stuff
Oasis::Model::Asset::Ob
John posted some suggestions on getting it done, you can check the
archives for them.
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - -
On Jul 10, 2007, at 9:49 AM, John Siracusa wrote:
> On 7/9/07, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> here's my testcase, reproduced on 3 machines (all running current
>> rose::db::object , not svn )
>
> I see the bug when using 0.764 from CPAN, but I can
ansactions belong in an ORM-- they belong
in your application code.
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
| CEO/Founder SyndiClick Networks
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
saction start . otherwise, i
like tylers interface much more than my own, and am hoping to migrate
to it soon.
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - -
onvenience method, which will
print the current state of the obj to STDERR. comes up as a 0 every
time. looking at the pg table , it inserts a 0 as well, not an undef.
i must be doing something blatantly wrong here -- and someone will
notice it just looking at the class setup. i hope so.
On Jul 9, 2007, at 11:34 AM, John Siracusa wrote:
> I've added is/is_not operators to QueryBuilder in SVN. Give them a
> spin and let me know if they do what you want.
awesome. will test tonight
can't wait to rid the 'clauses' sections from my manag
setting this col to undef, and it is not a 'not_null' col
suggestions?
- Jonathan Vanasco
- findmeon.com / RoadSound.com
Sent from my iPhone
-
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - t
object_class=> object_class(),
seems to work.
thanks!
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - -
o+cto sucks, but I have a new
operations guy starting next week!
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
| CEO/Founder SyndiClick Networks
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
On Jun 20, 2007, at 12:16 AM, David Lloyd wrote:
>
> Is it possible to have separate read and write database connections in
> Rose::DB::Object?
>
> For example, I might have read to:
>
> read.db.com
>
> ...and...write to:
>
> write.db.com
>
> ...is it possible to get any updates/deletes and i
array instead of a normal array.
the blessed array had sort and other methods on it...
so the code to interact would be:
my $names= myapp::names::manager->load_names();
foreach my $name ( @{$names->blessed_class_
{'com.2xlp'}{'F_display_only_b'} ;
$self->{'com.2xlp'}{'F_display_only_b'}= ( $self->id * 3.1417 ) * 2;
return $self->{'com.2xlp'}{'F_display_only_b
or primary column to be silently "inserted" into my
search?
This seems like a side-effect of using the same object to both
initiate a search (with load()) and to insert a new row. I'm sure
there's some history and good justification behind this decision,
though.
Jon
O
Thanks. I had the feeling it would be a short answer. :)
Jon
On Jun 21, 2007, at 11:07 AM, John Siracusa wrote:
> On 6/21/07, Jonathan Swartz <[EMAIL PROTECTED]> wrote:
>> If I create a new Rose::DB::Object without specifying some not-null
>> fields, should the getter
If I create a new Rose::DB::Object without specifying some not-null
fields, should the getters return a defined or undefined value?
The following script provides different answers for sqlite and mysql.
swartz> ./rose.pl sqlite
email is not defined
swartz> ./rose.pl mysql
email is defined
and I
On Jun 16, 2007, at 1:11 AM, Jeffrey Horn wrote:
> Aw shucks! You guys think of everything! :)
>
> Thanks! This is exactly what I was looking for!
>
> -- Jeff Horn
i think thats pretty much straight from the docs
i just converted a micro-app to rose last night, and had that handy ;)
---
On Jun 15, 2007, at 6:40 PM, Jeffrey Horn wrote:
> I'm wondering if there's any code out there that will generate
> Rose::DB::Object's for all tables in a database. I have my own
> makeshift
> version which will do it for a given table, and the loop to simply
> iterate
> through tables isn't
On Jun 5, 2007, at 8:36 AM, John Siracusa wrote:
>
> I made some changes to the error messages in SVN, but I could make
> them better if you could give me a concrete example of a call that
> produces an error for you, and what you'd like the error to be
> instead.
> Again, I think I need to see
value )
b- note the manager function. in order to trace, i find myself
opening 1 file to see which function i'm calling, then i get to open
the manager file
it would be even neater if rose could just infer/guess something or
bypass this -- since rose uses bind variables, there shouldn&#
On Jun 4, 2007, at 12:35 PM, John Siracusa wrote:
> Yes, and it works the same way as for columns. Like columns,
> relationships and foreign keys have N named method types created on
> their behalf. Examples of method type names are get, get_set,
> get_set_on_save, find, count, etc. There are
$b= b->new();
$a->b( $b );
right now i'm using the relationship name, because it works. but
thats not the best way.
do column_accessor_value_pairs/column_mutator_value_pairs work for
relation names too?
can we have a seperate relation function ?
sorry this took so long to respond -- been too busy on the business
side of work, little time for code.
On May 18, 2007, at 2:21 PM, John Siracusa wrote:
> I think it's pretty clear given that the table definition is right
> above it, and contains:
>
> UNIQUE(product_id, region)
ok. th
ed the rose code and pglog -- and it gave every
indication rose wasn't. Then other sections started failing with
random code and modules. I have no idea WTF happened.
//
please disregard the last question.
additional troubleshooting showed that , somehow , perl on that dev
box is completely corrupted.
reinstalling *everything* seems to have fixed it after we tested the
same app on another server.
-
SELECT t1.available_api_application_keys AS
t1_available_api_application_keys
and using a lookup hash to map t1_available_api_application_keys to
the t1 object and column when creating objects
// Jonathan Vanasco
| - - - - - - - - - - - - - - - -
where
appropriate, but it will do a one-to-many at the least.
On a side note-- if you're migrating to innodb and using foreign
keys, i'd suggest taking out any mysqlisms while your at it
( converting set/enum cols to lookup tables , etc ). if you ever
want to switch dbs from
On May 23, 2007, at 9:36 AM, John Siracusa wrote:
> Does anyone have any good ideas for a namespace for module that
> augment or
> extend RDBO, but that are not part of the "official" RDBO
> distribution? The
> first thing that springs to my mind is:
>
> Rose::DBx::*
At first I really l
que keys don't have
names, they have columns. foreign keys and relationships have user
defined names.
thats my documentation gripe of the week :)
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - -
On May 17, 2007, at 3:12 PM, Michael Reece wrote:
> here is a bit of ugliness that hints at what i am after:
I think i do some similar stuff, so I'll give my advice -- which you
may be better off to not take ;)
I never use 'make manager methods'. i started it as a memory saver
because i us
hen called-- but the
only way to get at that elsewhere in your code is by calling !
not_found -- which is like an evil tongue twister.
I'm using my own found method in a base class to keep my sanity
(
ow there is no where clause, if 'use_key' is removed, we get
SELECT id , name FROM item WHERE id = ? - bind params: 10
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
| SyndiClick.com
| - - - - - - -
;
have you looked at the manager class ?
http://search.cpan.org/~jsiracusa/Rose-DB-Object-0.764/lib/Rose/DB/
Object/Manager.pm
i think that should handle everything you need.
/
I'm trying to consolidate + standardize some code.
I realized several manager methods were identical, so I merged them
into a shared manager class.
Can anyone suggest a better way than what follows ?
Basically, I just have these objects inherit from another manager
subclass, which calls $cla
I'm making quite a bit of custom rose functions for my projects , and
looking for feedback on some that i think would be useful if rolled
back into Rose::DB::Helpers
the newest one is related_index
related_index takes a rose object and creates an internal hash to
quickly access specific LO
ng code, bug fixes, or just suggestions.
its even longer than you think:
at the least, you missed documenting the strip() feature in the
changelog!
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
| Sy
On May 1, 2007, at 3:25 PM, John Siracusa wrote:
> As for sharing the code, didn't Jonathan post it earlier?
The SVN is now here:
http://dev.2xlp.com/svn/mod_perl/cpan_patches/Rose/Rose-DB-Object/
JsonYamlObjectTree/trunk/
It still doesn't support TimeStamp or any
On May 2, 2007, at 1:54 PM, John Siracusa wrote:
> You can do this already, more or less, by making your own
> My::EmailColumn class based on Rose::DB::Object::Metadata::Column,
> then mapping that class to a type name using
> meta->column_type_class(email => 'My::EmailColumn'). Then, in your
>
On May 2, 2007, at 10:13 AM, John Siracusa wrote:
>> I can write this:
>>
>> reason => SCALAR( 255, DEFAULT '', NOT_NULL, LAZY ),
>
> Cute :)
Nice.
This reminds me of an idea I had a while ago...
I wrote my own form class a few years back, because none of the CPAN
modules could meet my fun
On May 1, 2007, at 3:08 PM, Derek Watson wrote:
> I feel like I'm repeating work here -- I've got a similar set of
> functions that serialize and restore trees of objects, only I'm doing
> this in a very specific way, keeping only what I want from specific
> classes (not reusable).
>
> What are t
On May 1, 2007, at 3:25 PM, John Siracusa wrote:
> It won't be in 0.764. It's too young and I want to wring it out a bit
> first. For 0.765, if it has tests and docs and we're all happy with
> the API and behavior, sure. As for sharing the code, didn't Jonathan
&
On May 1, 2007, at 10:32 AM, John Siracusa wrote:
> That's not going to make it into 0.764, but a feature that is in 0.764
> might help you:
>
> query =>
> [
> \q(is_selected is not false), # literal SQL
> ...
> ],
I didn't expect .764
I've got it covered now using t
i'd like to see an "is not" operator
something like:
'!NAME' => undef# COLUMN IS NOT NULL
under postgres, booleans behave as such:
example table
create table test ( id integer not null , is_selected boolean );
insert into test ( id , is_selected ) values ( 1 , NU
On Apr 30, 2007, at 4:37 PM, John Siracusa wrote:
> Even though there are a bunch of things on my short-term list that I
> haven't gotten to, the change list is already quite large and I want
> to cut a release soon (it will be 0.764). Is there any feature that I
> previously mentioned I'd get t
On Apr 27, 2007, at 10:25 AM, John Siracusa wrote:
> Sorry for any delays in responding to RDBO issues or email. We just
> had a baby (our second) and things will likely be a bit hectic here
> for a while... :)
Congrats!
// Jonatha
On Apr 26, 2007, at 6:18 PM, Jonathan Vanasco wrote:
>
> e) when restoring an object, I'm trying to figure out how stuff the
> foreign keys and relationships. i can bring them up as new objects
> just fine, but there aren't really any rose methods for stuffin
i wrote the following functions , which are pretty much nice and fine :
object_tree_to_json
object_tree_to_yaml
object_tree_to_hash
i'm now working on
object_tree_from_json
object_tree_from_yaml
object_tree_from_hash
and running into a few issues
> My only concern was how portable the solution would be.
> Theoretically,
> Rose could add enough value to make it work regardless of the
> database.
> For example, right now, our version of MySQL silently truncates
> values
> not in an enum list and add the empty string into enum fields.
;m struggling with the sf svn server. it keeps dropping out.
hopefully i can pull a complete copy to test.
Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - -
uery
the table structure itself is the same, but using explain you can
understand and influence how the planner plans
on a small dataset, the difference is negligble. but on a set of
40million rows, the difference is betwe
On Apr 20, 2007, at 11:06 AM, John Siracusa wrote:
> On 4/19/07, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> Would it be terribly difficult to port:
>>
>> Rose::DB::Object::MixIn
>>
>> into
>>
>> Rose::Object::MixIn
>
my guy feeling is that maybe you lost some indexes on db tables
during the migration
i'd do a raw test in a sql client directly, bench query on each and
use 'explain' to see how the planner is attempting the call
-
This
calls to
Rose::Class::MakeMethods::Set , which should work on the Rose::Object
base class ( at least with my limited understanding of that stuff )
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - -
a few days to deal with regex hell
for now
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
|
On Apr 18, 2007, at 10:22 AM, John Siracusa wrote:
> On 4/17/07, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> Though that can be handled using a strip helper alone too -- we'd
>> just need an unstrip() function that builds back in a db object and
>> meta
e layer would have its own information to
strip )
Though that can be handled using a strip helper alone too -- we'd
just need an unstrip() function that builds back in a db object and
metadata, and we'd need to loop each object for the right
information to strip or not.
the one PITA all t
of all this is to create cachable objects that are as-
small-as-possible when they go into memory, yet have almost no
overhead to revive.
its rather a large request for restructuring rose for
I'm going to need to cache my objects into memcached , and looking
for options.
the quick&dirty way would be to just use storable. i think
cache::memcached should be able to bring back the entire object
structure -- including nested objects
has anyone done this? pros / cons ?
my other op
1 },
color_id => { type => 'bigint',
not_null => 1 },
foreign_keys => [
product=> {
class => 'MyApp::Product',
key_columns => { product_id => 'id' },
urns the same cols
as the real item
b- subclass the RODB obj and just change the table name in the
package meta.
Which would give you a RODB obj that will always fail on DB writes.
Granted, that only
I have a 4 key map table , and its confusing me how to do this in rose.
Rose Tutorial:
Products
id , name
Colors
id , name
Product2Color
product_id , color_id
My Setup
Adds in a 'user' fkey + time to track the assignment of
On Apr 15, 2007, at 3:50 PM, John Siracusa wrote:
> On 4/14/07 1:16 AM, Jonathan Vanasco wrote:
>> is there a way to get Rose to search based on functions?
>>
>> ie:
>> bad
>> SELECT * FROM useraccount WHERE email = ?
>> good
>> SELECT * FROM user
is there a way to get Rose to search based on functions?
ie:
bad
SELECT * FROM useraccount WHERE email = ?
good
SELECT * FROM useraccount WHERE lower(email) = lower(?)
the second query will be optimized by the postgres planner to search
on this in
ting with the factory --
telling it to begin/rollback/commit. I use my own version of DB
handles that are similar to DBIx-Transaction -- the big difference is
that i overload connect and disconnect to become their own nested
transaction blocks.
you're in Mysql5 -- can't you use a function or stored
procedure to emulate that better?
also, if you're worried about concurrency, why not just stuff your
rose objects with a dbh that is in a transaction and handle s
r have new object
that is makes use of existing helper methods like clone or
init_with_xxx , in the context of a clone style call or instantiation
hook
basically, i want to have my cake and eat i
On Apr 6, 2007, at 10:01 AM, John Siracusa wrote:
> On 4/5/07, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> Any chance of these features being pushed back into rose?
>
> The first thing you're likely to see is support for triggers on save,
> insert, update, and d
_with_column_value_pairs( $writer->column_value_pairs );
or
my $reader= $writer->clone();
bless( $reader , MyApp::RoseDb::Object::ReadonlyFoo );
// Jonathan Vanasc
call me lazy--
I just use a bad db handle for all things rose, then have a custom
method that pulls a read-only or read/write handle from a db factory
as needed. ie, they'll always fail on save, insert, update because
they don't have the db permissions
---
ng the shell manager classes created for all my modules, but i
don't use the default manager methods . i fill that class with
custom calls as needed.
I just regex it out in 5 seconds, but I think thats marginally useful
to others. perhaps not.
// Jonat
I have a request for the docs:
I'm adding a new table with 2 unique indexes / keys on it , one of
which is a multiple column key. the docs are really unclear on how
thats done in the setup.
example:
unique keys are:
1- single col - 'hex_id'
2- mult co
doc change:
with_managers seems to default to true , it should be listed as such
feature request:
it would be nice if a variation of with_managers could be called ,
that did not create:
__PACKAGE__->make_manager_methods('useraccount');
---
On Apr 2, 2007, at 1:31 AM, Jonathan Vanasco wrote:
>
> Lots of messages like this:
>
> Could not auto-retrieve unique keys for class
> X::RoseDB::Object::X - No columns found for relation id
> 174592, column numbers 0 at /Library/Perl/5.8.6/Rose/DB/Object/
> Metadat
Lots of messages like this:
Could not auto-retrieve unique keys for class
X::RoseDB::Object::X - No columns found for relation id
174592, column numbers 0 at /Library/Perl/5.8.6/Rose/DB/Object/
Metadata/Auto/Pg.pm line 111.
at /Library/Perl/5.8.6/Rose/DB/Object/Loader.pm line 889
wh
On Mar 29, 2007, at 4:07 PM, James Masters wrote:
> So when you say "startup", does this mean upon Apache startup or upon
> your script startup? I can see that if it's in Apache's memory before
> a script even runs, that would be hugely performance enhancing.
apache startup.
there's no point o
On Mar 29, 2007, at 3:23 AM, James Masters wrote:
> Thanks for your very helpful reply, Jonathan. It answered my general
> question but I'm left with a couple of more specific ones.
>
> Pl. see in line.
>
>> I do that with ~20 tables that hold data I consider '
ase itself , and thats greatly alleviated by persistant
connections which most environments now support. make 2 queries or
40 queries probably isn't going to make any significant difference in
your usability unless you're getting into the hundreds of requests
per second situation or
On Mar 20, 2007, at 7:02 PM, Jud wrote:
> On 20/03/07 18:57 -0400, Jonathan Vanasco <[EMAIL PROTECTED]> wrote:
>> ok. its on my list right after the debugger module :)
>
> If you're actually putting together a debug module, I'd put in a
> request
> for a s
27;d just have to add
> more such methods. The Rose::DB::Sequence class wouldn't have any
> db-specific code in it at all.
well yes, but there would need to be a lot of conditionals in that
module / the db specific drivers to implement the
avoid blocking of concurrent transactions that
obtain
numbers from the same sequence, a nextval operation is never rolled
back; that is, once a value has been fetched it is considered used,
even if the transaction that did the nextval later aborts. T
e your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?
> page=join.php&p=sourceforge&CID=DEVDEV
> ___
On Mar 16, 2007, at 1:12 PM, Randal L. Schwartz wrote:
> package My::RDBO;
> use base Rose::DB::Object;
> use strict;
>
> sub init_db {
> our $cached_db ||= do {
> require My::RDB; # my Rose::DB subclass
> My::RDB->new;
> }
> }
>
> 1;
mine rel
ing a read-
only dbh from a factory class that my entire application uses.
As needed, I have a function that rose reset the db with a write dbh
from the factory.
it works exceedingly well for a clustered db environment.
//
t; at best.
no, that makes perfect sense: transactions are a part of application
logic , not orms logic
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - -
n your perl code
a) -00-00 is a mysqlism for putting null dates in a not-null
field -- most other dbs will not accept that as a date.
b) many apps / modules across languages are designed to throw errors
on invalid dates -- especially ones like th
ips => [ vendor => 'many to one' ],
(Again, this all depends on naming the tables, classes, and
columns
in accordance with the expectations of the convention manager.) The
resulting vendor() and delete_vendor() methods behave exactly the
same
is there any way i can coerce this into rose queries?
if not, any chance of including these as flags in the future? i'm
having a db locking / blocking issue on a certain feature right now ,
and that would likely solve it.
// Jonathan Va
1 - 100 of 263 matches
Mail list logo
