From lispnik at gmail.com  Fri Jan  4 11:40:23 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Fri, 04 Jan 2008 17:40:23 +0600
Subject: [postmodern-devel] Inserting into related tables
Message-ID: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>

I don't quite understand how to do inserts into related table.

For example, we have table 'items' with field 'producer_id' and table
'producer' with 'producer_id' and 'producer_name' fields:

,----
| (deftable item ()
|   ((producer-id :type (or integer db-null)")
|    ... ; Other fields
|   )
|   (:auto-id id))
| 
| (postmodern:deftable producers ()
|   ((producer-name :type (string 50)))
|   (:auto-id producer-id)
|   (:indices producer))
| 
| ;; Add unique index manually
| (postmodern:execute
|      (:create-unique-index 'producers-name-uniq-idx
|                            :on 'producers
|                            :fields 'producer-name))
`----

When we add new item to the database, producer may exist or may be new.
We have to check producer's existense and add it to table if required.
The only safe way to do it seems to be

,----
| ;; We ignore error if name exists.  But what about other errors?
| (ignore-errors
|   (let ((id (postmodern:next-id 'producers)))
|             (postmodern:execute
|              (:insert-into table :set
|                            'producer-id id
|                            'producer-name name))))
| 
| ;; Now get ID to insert into items.producer-id.
| (postmodern:query
|     (:select 'producer-id :from table :where (:= 'producer-name name))
|   :single)
`----

However, on each insertion we increment producers-producer-id-seq even
if producer name already exists in the table.  This is somewhat clumsy.
And faulty query wrapped with ignore-errors doesn't work with transactions.
Is there any other postmodern-way?

-- 
Ivan Boldyrev

                                       XML -- new language of ML family.


From marijnh at gmail.com  Fri Jan  4 14:35:07 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 15:35:07 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801040635r46e9df81w7d963c79483a9572@mail.gmail.com>

Hi Ivan,

Why don't you just start by looking for the record, and create a new
record only if it is not found? Something like (untested):

(defun get-producer-id (name)
  (or (query (:select 'id :from 'producer :where (:= 'name name)))


On Jan 4, 2008 12:40 PM, Ivan Boldyrev <lispnik at gmail.com> wrote:
> I don't quite understand how to do inserts into related table.
>
> For example, we have table 'items' with field 'producer_id' and table
> 'producer' with 'producer_id' and 'producer_name' fields:
>
> ,----
> | (deftable item ()
> |   ((producer-id :type (or integer db-null)")
> |    ... ; Other fields
> |   )
> |   (:auto-id id))
> |
> | (postmodern:deftable producers ()
> |   ((producer-name :type (string 50)))
> |   (:auto-id producer-id)
> |   (:indices producer))
> |
> | ;; Add unique index manually
> | (postmodern:execute
> |      (:create-unique-index 'producers-name-uniq-idx
> |                            :on 'producers
> |                            :fields 'producer-name))
> `----
>
> When we add new item to the database, producer may exist or may be new.
> We have to check producer's existense and add it to table if required.
> The only safe way to do it seems to be
>
> ,----
> | ;; We ignore error if name exists.  But what about other errors?
> | (ignore-errors
> |   (let ((id (postmodern:next-id 'producers)))
> |             (postmodern:execute
> |              (:insert-into table :set
> |                            'producer-id id
> |                            'producer-name name))))
> |
> | ;; Now get ID to insert into items.producer-id.
> | (postmodern:query
> |     (:select 'producer-id :from table :where (:= 'producer-name name))
> |   :single)
> `----
>
> However, on each insertion we increment producers-producer-id-seq even
> if producer name already exists in the table.  This is somewhat clumsy.
> And faulty query wrapped with ignore-errors doesn't work with transactions.
> Is there any other postmodern-way?
>
> --
> Ivan Boldyrev
>
>                                        XML -- new language of ML family.
> _______________________________________________
> postmodern-devel mailing list
> postmodern-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/postmodern-devel
>


From marijnh at gmail.com  Fri Jan  4 14:37:26 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 15:37:26 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>

Ugh, previous message got sent before I finished typing. Here we go again...

(defun get-producer-id (name)
 (or (query (:select 'id :from 'producer :where (:= 'name name)))
       (let ((new-producer (make-instance 'producer :name name))
             (insert-dao new-producer)
             (get-id new-producer))))

Does that work (it is probably full of typos and mistakes, but you get
the idea).

Cheers,
Marijn


From marijnh at gmail.com  Fri Jan  4 15:52:19 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 16:52:19 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
Message-ID: <befaa7aa0801040752h72df7382k4398064f55e7eeb9@mail.gmail.com>

> (defun get-producer-id (name)
>  (or (query (:select 'id :from 'producer :where (:= 'name name)))
>        (let ((new-producer (make-instance 'producer :name name))
>              (insert-dao new-producer)
>              (get-id new-producer))))

(Actually, now that I think about it, this would probably not be
thread-safe. I'll see if I can figure out a better approach.)

Cheers,
Marijn


From lispnik at gmail.com  Fri Jan  4 16:01:46 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Fri, 04 Jan 2008 22:01:46 +0600
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	(Marijn Haverbeke's message of "Fri\, 4 Jan 2008 15\:37\:26 +0100")
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
Message-ID: <jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10074 day of my life Marijn Haverbeke wrote:
> (defun get-producer-id (name)
>  (or (query (:select 'id :from 'producer :where (:= 'name name)))
>        (let ((new-producer (make-instance 'producer :name name))
>              (insert-dao new-producer)
>              (get-id new-producer))))
>
> Does that work (it is probably full of typos and mistakes, but you get
> the idea).

There is a race condition:

     Session A                             Session B

(:select ... "AA")) => nil            (:select ... "AA")) => nil      

(sequence-next ...) => 10             (sequence-next ...) => 11

(:insert 'producer :set 'id   10      (:insert producer :set id   11
                        'name "AA")                          name "AA")

So what is ID of "AA": 10 or 11?  And if UNIQUE index is created, one of
sessions will fail.  If no UNIQUE index is created, we have two lines of
"AA" in 'producers'.

Perhaps, I misunderstand something, I'm quite new to SQL, transactons
and so on.

MySQL has REPLACE and INSERT IGNORE extensions.  Perhaps, the only sane
way of doing it in PostgreSQL is using locks, but they are another can
of worms...

-- 
Ivan Boldyrev

                       Perl is a language where 2 x 2 is not equal to 4.


From marijnh at gmail.com  Fri Jan  4 16:15:01 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 17:15:01 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	<jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801040815w35a29a39u1114b321b0c47d96@mail.gmail.com>

> There is a race condition:

You are entirely right, of course. I asked about this on #postgresql
on IRC (freenode), and they didn't give me a better answer than your
first try. Wasting IDs is rather ugly -- though I'm not sure this
would be a concrete problem. Using ignore-errors is probably bad. Each
postmodern exception has a database-error-code accessor, which can be
used to identify a specific type of errors, and let all others
through.

Hope that helps. If you come up with a better way, let me know.

Cheers,
Marijn


From marijnh at gmail.com  Fri Jan  4 16:16:47 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 17:16:47 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	<jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801040816t3292fd25ld22e73e33287ac48@mail.gmail.com>

Oh (and I'm really sending too many separate mails today) -- have you
tried just using the producer names as their IDs? This has worked
quite well for me in the past (of course, if you have to add more
information to producers you will still need the producer table
anyway, but  you do get rid of the useless ID numbers).

Marijn


From lispnik at gmail.com  Fri Jan  4 17:09:23 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Fri, 04 Jan 2008 23:09:23 +0600
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <befaa7aa0801040815w35a29a39u1114b321b0c47d96@mail.gmail.com>
	(Marijn Haverbeke's message of "Fri\, 4 Jan 2008 17\:15\:01 +0100")
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	<jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040815w35a29a39u1114b321b0c47d96@mail.gmail.com>
Message-ID: <8eejcxy0u4.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10074 day of my life Marijn Haverbeke wrote:
> Hope that helps. If you come up with a better way, let me know.

I have googled for "postgres insert ignore" and found solution using
SAVEPOINT and ROLLBACK TO:
http://robbat2.livejournal.com/214267.html

(The example is from PosgreSQL UPDATE documentation).

On 10074 day of my life Marijn Haverbeke wrote:
> Oh (and I'm really sending too many separate mails today) -- have you
> tried just using the producer names as their IDs?

It only solves wasted IDs problem, but no other, doesn't it?  And joins
on strings are slower than on ints (though I cannot say my application
is so much performance-constrained).

-- 
Ivan Boldyrev

              "Assembly of Japanese bicycle require great peace of mind."


From marijnh at gmail.com  Fri Jan  4 17:43:59 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Fri, 4 Jan 2008 18:43:59 +0100
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <8eejcxy0u4.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	<jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040815w35a29a39u1114b321b0c47d96@mail.gmail.com>
	<8eejcxy0u4.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801040943x2a68eb18y6e1889e415799390@mail.gmail.com>

> I have googled for "postgres insert ignore" and found solution using
> SAVEPOINT and ROLLBACK TO:
> http://robbat2.livejournal.com/214267.html

Does that actually solve the wasted id problem?

> > have you tried just using the producer names as their IDs?
>
> It only solves wasted IDs problem, but no other, doesn't it?

The idea was that you wouldn't need the table that maps names to ids
at all, so this whole problem goes away. I'm not sure that joining on
strings is really that much slowing in Postgres, but I haven't got any
benchmarks to really find out.

Anyway, I also came up with this monstrosity, it doesn't solve the
wasted id problem, but take a look:

(let ((name "dummy") (id (generate-next-id)))
  (execute (:insert-into 'producer (:except (:select name id) (:select
'* :from 'producer :where (:= 'name name))))))

It uses EXCEPT to remove the new row from the set of inserted rows if
it already exists. If you want to use it, you'll have to pull in a
patch I just added to Postmodern because my :insert-into operator was
a bit too picky about the input it allowed.

Cheers,
Marijn


From lispnik at gmail.com  Sat Jan  5 06:53:06 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 12:53:06 +0600
Subject: [postmodern-devel] WITH-SAVEPOINT macro
Message-ID: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>

WITH-SAVEPOINT macro is modelled after WITH-TRANSACTION.  I would be
glad to see it in 1.05 :)

-- 
Ivan Boldyrev

       Assembly of a Japanese bicycle requires greatest peace of spirit.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: savepoint.lisp
Type: text/x-common-lisp
Size: 1636 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/postmodern-devel/attachments/20080105/2b804655/attachment.bin>

From lispnik at gmail.com  Fri Jan  4 18:49:26 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 00:49:26 +0600
Subject: [postmodern-devel] Bug report: Incorrect handling of SIMPLE-DATE
	types in S-SQL
Message-ID: <18fxxdfmtl.fsf@ibhome.cgitftp.uiggm.nsc.ru>

This doesn't work:

,----
| (postmodern:query
|  (:select (:- :localtimestamp
|               (simple-date:encode-interval :hour 2 :minute 10))))
| 
| Database error 22007: invalid input syntax for type timestamp: "2 hours 10 minutes "
| Query: (SELECT (localtimestamp - '2 hours 10 minutes '))
`----

This does work:

,----
| (postmodern:query
|  (:select (:- :localtimestamp (:raw "interval '2 hours 10 minutes'"))))
`----

It was tested with PostgreSQL 8.0.13 and 8.2.4.

SQL-IZE methods for SIMPLE-DATE objects should return escaped strings
like "interval '2 hours 10 minutes'", not just "2 hours 10 minutes", and
second value of SQL-IZE must be NIL:

b
,----
|   (:method ((arg simple-date:date))
|       (multiple-value-bind (year month day) (simple-date:decode-date arg)
|         (format nil "date ~A"
|                 (sql-escape-string (format nil "~4,'0d-~2,'0d-~2,'0d" year month day)))))
`----

(Untested!)

Perhaps, sql-escape-string can be omitted: the output has no dangerous
characters anyway.

,----
|  (:method ((arg simple-date:date))
|     (multiple-value-bind (year month day) (simple-date:decode-date arg)
|       (format nil "date '~4,'0d-~2,'0d-~2,'0d'" year month day)))
`----

-- 
Ivan Boldyrev

Violets are red, Roses are blue. //
I'm schizophrenic, And so am I.


From marijnh at gmail.com  Sat Jan  5 09:40:30 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sat, 5 Jan 2008 10:40:30 +0100
Subject: [postmodern-devel] WITH-SAVEPOINT macro
In-Reply-To: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>

> WITH-SAVEPOINT macro is modelled after WITH-TRANSACTION.  I would be
> glad to see it in 1.05 :)

Looks good. Is it intentional that the savepoint is rolled back after
its body finishes, unless the body explicitly released/committed the
savepoint? Also, inheriting from transaction-handle means that
abort-transaction and commit-transaction are also valid on savepoints.
Does this make sense, or should we just create a whole new class for
savepoints?

Cheers,
Marijn


From marijnh at gmail.com  Sat Jan  5 09:44:28 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sat, 5 Jan 2008 10:44:28 +0100
Subject: [postmodern-devel] Bug report: Incorrect handling of SIMPLE-DATE
	types in S-SQL
In-Reply-To: <18fxxdfmtl.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <18fxxdfmtl.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801050144l3ae4fddbm955ea03df97777e0@mail.gmail.com>

So far I always used these types like (:select (:type my-date-value
date)), but explicitly putting 'date' in front of the value string
seems to work (I had never seen it before). Only, there is a reason
that sql-ize allows one to take the un-escaped version of a value --
when passing arguments to queries with $1-like 'holes' in them, you
need to pass them as un-escaped strings, and something like "date
'2000-01-01'" produces an error. I've modified sql-escape-string to
take an optional second parameter, a prefix, and when sql-ize returns
a string as its second value (rather than T or NIL), this now means
that the value should be escaped, and that string can be prefixed as
type info. This seems to work fine (and has been applied to the
repository).

Cheers,
Marijn


From lispnik at gmail.com  Sat Jan  5 10:01:19 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 16:01:19 +0600
Subject: [postmodern-devel] Inserting into related tables
In-Reply-To: <befaa7aa0801040943x2a68eb18y6e1889e415799390@mail.gmail.com>
	(Marijn Haverbeke's message of "Fri\, 4 Jan 2008 18\:43\:59 +0100")
References: <o0tzltyg2g.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040637x3e3378c1v986dafd8e76f86a2@mail.gmail.com>
	<jplk75y3yt.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040815w35a29a39u1114b321b0c47d96@mail.gmail.com>
	<8eejcxy0u4.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801040943x2a68eb18y6e1889e415799390@mail.gmail.com>
Message-ID: <wv7iioa8wg.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10074 day of my life Marijn Haverbeke wrote:
>> SAVEPOINT and ROLLBACK TO:
>> http://robbat2.livejournal.com/214267.html
> Does that actually solve the wasted id problem?

No, "a nextval operation is never rolled back".

>> > have you tried just using the producer names as their IDs?
>>
>> It only solves wasted IDs problem, but no other, doesn't it?
>
> The idea was that you wouldn't need the table that maps names to ids
> at all, so this whole problem goes away.

It was original design (created by some student).  However, I need to do
queries like "get all producers like 'abc%'" and "get all producers" for
autocompletion and input validation.  SELECT DISTINCT works, but look at
the timings:

CL-USER> (time (dotimes (i 100) (postmodern:query (:select (:distinct 'producer) :from 'catalog))))
Evaluation took:
  3.353 seconds of real time
  0.056004 seconds of user run time
  0.016001 seconds of system run time
  0 calls to %EVAL
  0 page faults and
  417,472 bytes consed.
NIL
CL-USER> (time (dotimes (i 100) (postmodern:query (:select 'producer :from 'producers))))
Evaluation took:
  0.142 seconds of real time
  0.032002 seconds of user run time
  0.008001 seconds of system run time
  0 calls to %EVAL
  0 page faults and
  273,816 bytes consed.
NIL

(Indicies are present).  Well, for my application such difference is not
really important, but what if it will in next project?

> Anyway, I also came up with this monstrosity, it doesn't solve the
> wasted id problem, but take a look:

Thanks, I will note.  But so far SAVEPOINT solution looks best for me.

-- 
Ivan Boldyrev

                                   Violets are red
                                   Roses are blue
                                   It's amazing what DNA splicing can do.


From lispnik at gmail.com  Sat Jan  5 10:09:01 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 16:09:01 +0600
Subject: [postmodern-devel] Bug report: Incorrect handling of SIMPLE-DATE
	types in S-SQL
In-Reply-To: <befaa7aa0801050144l3ae4fddbm955ea03df97777e0@mail.gmail.com>
	(Marijn Haverbeke's message of "Sat\, 5 Jan 2008 10\:44\:28 +0100")
References: <18fxxdfmtl.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050144l3ae4fddbm955ea03df97777e0@mail.gmail.com>
Message-ID: <jny7b48tz6.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10075 day of my life Marijn Haverbeke wrote:
> So far I always used these types like (:select (:type my-date-value
> date))...

Ouch, I didn't new of such way.

-- 
Ivan Boldyrev

                                                  Your bytes are bitten.


From lispnik at gmail.com  Sat Jan  5 10:07:12 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 16:07:12 +0600
Subject: [postmodern-devel] WITH-SAVEPOINT macro
In-Reply-To: <befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>
	(Marijn Haverbeke's message of "Sat\, 5 Jan 2008 10\:40\:30 +0100")
References: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>
Message-ID: <j13atca8mn.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10075 day of my life Marijn Haverbeke wrote:
> Looks good. Is it intentional that the savepoint is rolled back after
> its body finishes, unless the body explicitly released/committed the
> savepoint?

Sorry, this is a last-minute-introduced bug.  Swap release-savepoint and
rollback-savepoint in with-savepoint definition.

> Also, inheriting from transaction-handle means that abort-transaction
> and commit-transaction are also valid on savepoints.  Does this make
> sense, ...?

No.  I subclassed transaction-handle just for code writing speed.  If
you think it is important, make savepoint-handle an independent class.

-- 
Ivan Boldyrev

       Assembly of a Japanese bicycle requires greatest peace of spirit.


From marijnh at gmail.com  Sat Jan  5 11:04:05 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sat, 5 Jan 2008 12:04:05 +0100
Subject: [postmodern-devel] WITH-SAVEPOINT macro
In-Reply-To: <j13atca8mn.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>
	<j13atca8mn.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801050304g6fbdf098ib2b9b66e15d4761a@mail.gmail.com>

Okay, I think I fixed the code, and I added documentation. Could you
see whether the current version works for you? If so, I will release
it as a new version.

Cheers,
Marijn


From lispnik at gmail.com  Sat Jan  5 13:03:44 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sat, 05 Jan 2008 19:03:44 +0600
Subject: [postmodern-devel] String escaping (was: WITH-SAVEPOINT macro)
In-Reply-To: <befaa7aa0801050304g6fbdf098ib2b9b66e15d4761a@mail.gmail.com>
	(Marijn Haverbeke's message of "Sat\, 5 Jan 2008 12\:04\:05 +0100")
References: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>
	<j13atca8mn.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050304g6fbdf098ib2b9b66e15d4761a@mail.gmail.com>
Message-ID: <a1lk748lvz.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10075 day of my life Marijn Haverbeke wrote:
> Okay, I think I fixed the code, and I added documentation. Could you
> see whether the current version works for you? If so, I will release
> it as a new version.

My code seems to work fine with new version.

But wait a minute. :)

CL-USER> (postmodern:query (:select "\\'"))
WARNING: Postgres warning: nonstandard use of \\ in a string literal
Use the escape string syntax for backslashes, e.g., E'\\'.
(("\\'"))

SQL-IZE escapes #\' and #\\, however PosgreSQL (unlike MySQL) doesn't
like \ in ordinary string, but have special E'...' syntax.  Of course,
warnings can be just ignored, and I can also supress this warning with
setting standard_conforming_strings off, but isn't it better to output
all string literals with E'...' syntax?  Just add one letter :)

-- 
Ivan Boldyrev

                       Perl is a language where 2 x 2 is not equal to 4.


From lispnik at gmail.com  Sun Jan  6 17:03:00 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Sun, 06 Jan 2008 23:03:00 +0600
Subject: [postmodern-devel] Binary and unary operators
Message-ID: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>

1.  You are missing such operators like || and && (former is string
    concatenation, latter is tests overlapping geometric objects).

    There are also
      a. math: !, !!, #, >>, <<.

      b. geometric: ##, @@, @-@, ~=, <->, &<, &>, <<|, |>>, &<|, |&>,
         <^, >^, ?#, ?|, ?-|, ?||, @>, <@.

      c. network: <<=, >>=.

    ?| is both binary and unary operator, ! is postfix (but is
    equivalent to prefix !!, so you may define :! that expands to prefix
    !! or just :!! that expands to !!).

    There are also another operators like @, but they have
    functions-substitutions (@ x <=> abs(x)).

    psql's command \do (describe operators) also lists other operators.

2.  As PSQL allows defining own operators, I think it worth exporting
    s-sql::def-infix-ops or creating some more flexible user macro (one
    might prefer to use :concat instead of :\|\|).

    With such facility you don't need to define such rare-used operators
    like @-@ and ?-|, allowing user to define them.

-- 
Ivan Boldyrev

                                   Violets are red
                                   Roses are blue
                                   It's amazing what DNA splicing can do.


From marijnh at gmail.com  Sun Jan  6 20:55:48 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sun, 6 Jan 2008 21:55:48 +0100
Subject: [postmodern-devel] String escaping (was: WITH-SAVEPOINT macro)
In-Reply-To: <a1lk748lvz.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <bpbq80ahm5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050140q43c62e39vc74600fa1480e721@mail.gmail.com>
	<j13atca8mn.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801050304g6fbdf098ib2b9b66e15d4761a@mail.gmail.com>
	<a1lk748lvz.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801061255i46eb928fyac710dc3268d379e@mail.gmail.com>

> My code seems to work fine with new version.

Great, I will consider it functional then.

> SQL-IZE escapes #\' and #\\, however PosgreSQL (unlike MySQL) doesn't
> like \ in ordinary string, but have special E'...' syntax.  Of course,
> warnings can be just ignored, and I can also supress this warning with
> setting standard_conforming_strings off, but isn't it better to output
> all string literals with E'...' syntax?  Just add one letter :)

Done. I also added a *standard-sql-strings* variable which can be used
to configure S-SQL to just use standard strings. It would have been
nicer if this could just be derived directly from the connection, but
at the moment S-SQL does not have a concept of connections, and that
is kind of convenient, so if you really want to use standard strings
you'll just have to change this parameter.

Cheers,
Marijn


From marijnh at gmail.com  Sun Jan  6 20:57:30 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sun, 6 Jan 2008 21:57:30 +0100
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>

Hey Ivan,

> 1.  You are missing such operators like || and && (former is string
>     concatenation, latter is tests overlapping geometric objects).

You've probably noticed how easy it is to add operators to S-SQL.
Patches defining additional operators that are commonly used are
welcome -- || would be a good candidate.

> 2.  As PSQL allows defining own operators, I think it worth exporting
>     s-sql::def-infix-ops or creating some more flexible user macro (one
>     might prefer to use :concat instead of :\|\|).

I very much agree. Wanna write a patch? ;)

Cheers,
Marijn


From lispnik at gmail.com  Mon Jan  7 14:50:38 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Mon, 07 Jan 2008 20:50:38 +0600
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
	(Marijn Haverbeke's message of "Sun\, 6 Jan 2008 21\:57\:30 +0100")
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
Message-ID: <xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10077 day of my life Marijn Haverbeke wrote:
> ... a patch?

Comment string may require full rewriting.

diff -rN -u old-postmodern/postmodern/package.lisp new-postmodern/postmodern/package.lisp
--- old-postmodern/postmodern/package.lisp      2008-01-07 20:38:27.000000000 +0600
+++ new-postmodern/postmodern/package.lisp      2008-01-07 20:38:27.000000000 +0600
@@ -25,6 +25,7 @@
    #:smallint #:bigint #:numeric #:real #:double-precision
    #:bytea #:text #:varchar
    #:*escape-sql-names-p* #:sql-escape-string
+   #:def-infix-ops
 
    ;; Condition type from cl-postgres
    #:database-error #:database-error-message #:database-error-code
diff -rN -u old-postmodern/s-sql/s-sql.lisp new-postmodern/s-sql/s-sql.lisp
--- old-postmodern/s-sql/s-sql.lisp     2008-01-07 20:38:27.000000000 +0600
+++ new-postmodern/s-sql/s-sql.lisp     2008-01-07 20:38:27.000000000 +0600
@@ -16,6 +16,7 @@
            #:to-sql-name
            #:sql-ize
            #:*escape-sql-names-p*
+           #:def-infix-ops
            #:sql
            #:sql-compile
            #:enable-s-sql-syntax))
@@ -373,21 +374,58 @@
       (destructuring-bind ,arglist ,args-name
         , at body))))
 
-(defun expand-infix-op (operator allow-unary args)
-  (if (cdr args)
-      `("(" ,@(sql-expand-list args (strcat " " operator " ")) ")")
-      (if allow-unary
-          (sql-expand (first args))
-          (error "SQL operator ~A takes at least two arguments." operator))))
-
-(defmacro def-infix-ops (allow-unary &rest ops)
+(defun expand-infix-op (operator class args)
+  (declare (type (member t :both nil) class))
+  (cond
+    ((cdr args)
+     `("(" ,@(sql-expand-list args (strcat " " operator " ")) ")"))
+    ((eq class t)
+     (sql-expand (first args)))
+    ((eq class :both)
+     `(,operator "(" ,@(sql-expand (first args)) ")"))
+    (t
+       (error "SQL operator ~A takes at least two arguments." operator))))
+
+(defmacro def-infix-ops (class &rest ops)
+  (declare (type (member t :both nil) class))
+  "Define infix operators.
+CLASS may be either T, :BOTH or NIL.
+
+  1. T.  S-SQL operators may be both binary and unary, but unary form
+     is equivalent to the only argument itself (e.g.
+     (:AND condition) => \"condition\").
+
+  2. :BOTH.  S-SQL operators may be both binary and unary, but unlike
+     T, operator is kept before the argument in unary form.  Example
+     is PosgreSQL's ?| operator:
+     (def-infix-ops :both :?\\|)
+     (:?\\| a)    => \"?|(a)\"
+     (:?\\| a b)  => \"a ?| b\".
+
+  3. NIL.  Operator is binary only.
+
+OPS is a list of operator designators.  There are two kinds of
+operator designators:
+
+  1. A keyword.  Downcased symbol value of the keword is used as SQL
+     name of operator.
+
+  2. Two-element list: (KEYWORD STRING).  String is used as SQL name
+     of operator, and KEYWORD is S-SQL name of operator.
+     Example: (:concat \"||\")."
   `(progn
     ,@(mapcar (lambda (op)
-                `(defmethod expand-sql-op ((op (eql ,op)) args)
-                  (expand-infix-op ,(string-downcase (symbol-name op)) ,allow-unary args)))
+                (when (keywordp op)
+                  (setf op (list op (string-downcase (symbol-name op)))))
+                (let ((kwd (first op))
+                      (txt (second op)))
+                  `(defmethod expand-sql-op ((op (eql ,kwd)) args)
+                     (expand-infix-op ,txt ,class args))))
               ops)))
-(def-infix-ops t :+ :* :& :|\|| :and :or :union)
-(def-infix-ops nil := :/ :!= :< :> :<= :>= :^ :intersect :except :~* :!~ :!~* :like :ilike)
+
+(def-infix-ops t :+ :* :& :|\|| :and :or :union :\|\| (:concat "||"))
+(def-infix-ops nil := :/ :!= :< :> :<= :>= :^ :intersect :except :~* :!~ :!~*
+                   :like :ilike :&& :% :\# :<< :>>)
 
 (def-sql-op :- (first &rest rest)
   (if rest


-- 
Ivan Boldyrev

                                      Life!  Don't talk to me about life.


From lispnik at gmail.com  Mon Jan  7 15:44:11 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Mon, 07 Jan 2008 21:44:11 +0600
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru> (Ivan Boldyrev's
	message of "Mon\, 07 Jan 2008 20\:50\:38 +0600")
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
	<xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <rcir25myic.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10077 day of my life Ivan Boldyrev wrote:
> On 10077 day of my life Marijn Haverbeke wrote:
>> ... a patch?
> +(defmacro def-infix-ops (class &rest ops)

BTW: names of definition macros with minus in their names are
traditionally started with define, not def (define-modify-macro,
define-method-combination, but defvar, defun, defmacro).  So proper name
for public macro should be define-infix-ops.

-- 
Ivan Boldyrev

Violets are red, Roses are blue. //
I'm schizophrenic, And so am I.


From lispnik at gmail.com  Mon Jan  7 17:36:01 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Mon, 07 Jan 2008 23:36:01 +0600
Subject: [postmodern-devel] Performance patch
Message-ID: <7xejctmtby.fsf@ibhome.cgitftp.uiggm.nsc.ru>

My application uses quite long and complex query, and profiling showed
that slowest function is SQL-COMPILE.  First I reimplemented it with
S-SQL::STRCAT, but then I thought: what if it will hit
CALL-ARGUMENTS-LIMIT (it is only 4096 on GNU Clisp!)?  So I changed
implementation of S-SQL:STRCAT to get list as the only argument, not as
&rest.

SQL-COMPILE is much faster and consumes on my query 10 times less
memory.  As SQL-COMPILE is usually called on run time, not compilation
time, I think it is very important.

--- old/s-sql/s-sql.lisp~	2008-01-07 20:05:17.000000000 +0600
+++ new/s-sql/s-sql.lisp	2008-01-07 23:26:03.000000000 +0600
@@ -25,7 +25,7 @@
 
 ;; Utils
 
-(defun strcat (&rest args)
+(defun strcat (args)
   "Concatenate a list of strings into a single one."
   (let ((result (make-string (reduce #'+ args :initial-value 0 :key 'length))))
     (loop :for pos = 0 :then (+ pos (length arg))
@@ -36,11 +36,11 @@
 (defun implode (sep list)
   "Reduce a list of strings to a single string, inserting a separator
 between them."
-  (apply 'strcat
-         (loop :for element :on list
-               :collect (car element)
-               :if (cdr element)
-               :collect sep)))
+  (strcat
+   (loop :for element :on list
+      :collect (car element)
+      :if (cdr element)
+      :collect sep)))
 
 (defun split-on-keywords% (shape list)
   "Helper function for split-on-keywords. Extracts the values
@@ -338,11 +338,11 @@
   (let ((list (reduce-strings (sql-expand form))))
     (if (= 1 (length list))
         (car list)
-        `(strcat , at list))))
+        `(strcat (list , at list)))))
   
 (defun sql-compile (form)
   (let ((*expand-runtime* t))
-    (car (reduce-strings (sql-expand form)))))
+    (strcat (sql-expand form))))
 
 ;; The reader syntax.
 
@@ -378,7 +378,7 @@
   (declare (type (member t :both nil) class))
   (cond
     ((cdr args)
-     `("(" ,@(sql-expand-list args (strcat " " operator " ")) ")"))
+     `("(" ,@(sql-expand-list args (strcat (list " " operator " "))) ")"))
     ((eq class t)
      (sql-expand (first args)))
     ((eq class :both)


-- 
Ivan Boldyrev

                                      Life!  Don't talk to me about life.


From marijnh at gmail.com  Mon Jan  7 17:53:56 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Mon, 7 Jan 2008 18:53:56 +0100
Subject: [postmodern-devel] Performance patch
In-Reply-To: <7xejctmtby.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <7xejctmtby.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801070953l12d07ffej46aa5ef17c7052df@mail.gmail.com>

> SQL-COMPILE is much faster and consumes on my query 10 times less
> memory.  As SQL-COMPILE is usually called on run time, not compilation
> time, I think it is very important.

Great find, I will try to apply it... but could you *please* use darcs
to create patch files? I always have strange problems with raw diff
files, and as an added advantage patches you record with darcs will
automatically have your name associated with them. Just (assuming you
have a darcs checkout) record your patches (with 'darcs record') and
then type 'darcs send -o somefile.patch' and it will ask you which
changes to include in the patch file. Then mail me the file, and
applying your patches will be a breeze.

Cheers,
Marijn


From marijnh at gmail.com  Tue Jan  8 20:02:22 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Tue, 8 Jan 2008 21:02:22 +0100
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <rcir25myic.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
	<xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<rcir25myic.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801081202k1cc5ab2va0c5cbf0993b316d@mail.gmail.com>

Ivan,

I pushed patches making a register-sql-operators macro available and
changing the way strings are passed to strcat. Take a look.

Cheers,
Marijn


From lispnik at gmail.com  Wed Jan  9 16:57:25 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Wed, 09 Jan 2008 22:57:25 +0600
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <befaa7aa0801081202k1cc5ab2va0c5cbf0993b316d@mail.gmail.com>
	(Marijn Haverbeke's message of "Tue\, 8 Jan 2008 21\:02\:22 +0100")
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
	<xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<rcir25myic.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801081202k1cc5ab2va0c5cbf0993b316d@mail.gmail.com>
Message-ID: <k9ejcr2ayy.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10079 day of my life Marijn Haverbeke wrote:
> I pushed patches making a register-sql-operators macro available and
> changing the way strings are passed to strcat. Take a look.

Excellent, but you forgot to insert comma at symbol NAME in a :unary
branch of MAKE-EXPANDER.

======================================================================
I tried to send appropriate one-line patch with darcs mail, but darcs
doesn't allow sending only this particular patch.  This is not a first
time I stuck into such darcs' behaviour, and this is one of reasons I
dislike this VCS.


$ darcs send -u -i -o comma.patch ../postmodern
Creating patch to "../postmodern"...

Tue Jan  8 15:21:32 NOVT 2008  Ivan Boldyrev <lispnik at gmail.com>
  * Enchance infix operators definition macro.
  def-infix-ops is renamed to define-infix-ops, and its first argument now
  accepts :both.  expand-infix-op is changed accordingly.
  Some new infix operators are added.
Shall I send this patch? (1/4)  [ynWvpxqadjk], or ? for help: n

Tue Jan  8 15:22:01 NOVT 2008  Ivan Boldyrev <lispnik at gmail.com>
  * Make SQL generation faster.
Shall I send this patch? (2/4)  [ynWvpxqadjk], or ? for help: n
You don't want to send any patches, and that's fine with me!

======================================================================

-- 
Ivan Boldyrev

                        Today is the first day of the rest of your life.


From marijnh at gmail.com  Wed Jan  9 17:54:12 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Wed, 9 Jan 2008 18:54:12 +0100
Subject: [postmodern-devel] Binary and unary operators
In-Reply-To: <k9ejcr2ayy.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <x2fxxa6g57.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801061257j2009f116x5913eb11257705f6@mail.gmail.com>
	<xlzlvhy9j5.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<rcir25myic.fsf@ibhome.cgitftp.uiggm.nsc.ru>
	<befaa7aa0801081202k1cc5ab2va0c5cbf0993b316d@mail.gmail.com>
	<k9ejcr2ayy.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801090954g47e80718ta6dbee7ab0d27a72@mail.gmail.com>

Hey Ivan,

That's what I get for making 'trivial' changes just before committing
something. I pushed a fix. The reason darcs doesn't prompt you to send
that patch is (I think) that you have a bunch of other patches in your
repository that are not in mine, and this change depends on one of
those.

Cheers,
Marijn


From marijnh at gmail.com  Fri Jan 11 23:02:56 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Sat, 12 Jan 2008 00:02:56 +0100
Subject: [postmodern-devel] Release 1.05
Message-ID: <befaa7aa0801111502y6ec797e1qf838eb595a3a1638@mail.gmail.com>

Hello list,

Thanks to Ivan's relentless bug finding and feature requesting,
version 1.05 of Postmodern is now a reality. See project page
(http://common-lisp.net/project/postmodern/) or darcs log
(http://common-lisp.net/cgi-bin/darcsweb/darcsweb.cgi?r=postmodern-postmodern;a=summary)
for a change list.

Enjoy,
Marijn


From airbaggins at gmail.com  Mon Jan 14 08:19:30 2008
From: airbaggins at gmail.com (William Robinson)
Date: Mon, 14 Jan 2008 08:19:30 +0000
Subject: [postmodern-devel] deftable class-name and unique indices problems
Message-ID: <478B1B12.70400@gmail.com>

Hi.

Thanks for the library. I was struggling a bit with pomo's DAO support. 
A few points:

I think it is worth reinforcing in the deftable documentation that a 
class-name NEEDS to be specified for the data-access methods to function.

Can we have/is there a facility for creating unique multi-column indices 
in deftable?

Some more examples for deftable would certainly not go amiss, 
particularly with repect to usage of :auto-id, :indices and :defer-id.

Thanks
Bill


From marijnh at gmail.com  Mon Jan 14 15:06:18 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Mon, 14 Jan 2008 16:06:18 +0100
Subject: [postmodern-devel] deftable class-name and unique indices problems
In-Reply-To: <478B1B12.70400@gmail.com>
References: <478B1B12.70400@gmail.com>
Message-ID: <befaa7aa0801140706q221fc066n577d4748532b86c3@mail.gmail.com>

Hey William,

I retouched the deftable documentation a bit, hopefully making it
clearer. There was no facility for directly cleating unique indices
from deftable, but I have added a patch to the repository that makes
this possible -- using "(:unique field-1 field-1)" instead of just
"(field-1 field-2)" after the :indices keyword.

If you have any specific suggestions for the docs, let me know, and I
can incorporate them.

Cheers,
Marijn


From lispnik at gmail.com  Wed Jan 16 17:06:09 2008
From: lispnik at gmail.com (Ivan Boldyrev)
Date: Wed, 16 Jan 2008 23:06:09 +0600
Subject: [postmodern-devel] Release 1.05
In-Reply-To: <befaa7aa0801111502y6ec797e1qf838eb595a3a1638@mail.gmail.com>
	(Marijn Haverbeke's message of "Sat\, 12 Jan 2008 00\:02\:56 +0100")
References: <befaa7aa0801111502y6ec797e1qf838eb595a3a1638@mail.gmail.com>
Message-ID: <xiir1tbszi.fsf@ibhome.cgitftp.uiggm.nsc.ru>

On 10082 day of my life Marijn Haverbeke wrote:
> version 1.05 of Postmodern is now a reality.

Very nice :)

Btw, I would use different specifiers for char and varchar:

(char n) => char(n)
(string n)  => varchar(n)

To me, STRING is closer to variable-length entities than to fixed-length
entities.  Especially if you compare char[n] and string in C or Pascal :)

-- 
Ivan Boldyrev

Violets are red, Roses are blue. //
I'm schizophrenic, And so am I.


From marijnh at gmail.com  Wed Jan 16 17:51:39 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Wed, 16 Jan 2008 18:51:39 +0100
Subject: [postmodern-devel] Release 1.05
In-Reply-To: <xiir1tbszi.fsf@ibhome.cgitftp.uiggm.nsc.ru>
References: <befaa7aa0801111502y6ec797e1qf838eb595a3a1638@mail.gmail.com>
	<xiir1tbszi.fsf@ibhome.cgitftp.uiggm.nsc.ru>
Message-ID: <befaa7aa0801160951g74b04711l9d806d2a201c13c8@mail.gmail.com>

On Jan 16, 2008 6:06 PM, Ivan Boldyrev <lispnik at gmail.com> wrote:
> Btw, I would use different specifiers for char and varchar:
> (char n) => char(n)
> (string n)  => varchar(n)

That was also my first idea, but there are two objections: Firstly,
Lisp values of type (string n) have to have *exactly* n characters,
and are thus more like char(n) than varchar(n). Secondly, char is an
exported symbol in the package :common-lisp, and thus not a very good
candidate for use in a library (you aren't allowed to define a type by
that name without first shadowing the symbol).

Marijn


From rm at seid-online.de  Wed Jan 23 20:05:59 2008
From: rm at seid-online.de (Ralf Mattes)
Date: Wed, 23 Jan 2008 21:05:59 +0100
Subject: [postmodern-devel] dynamically built queries
Message-ID: <1201118759.6570.21.camel@localhost.localdomain>

Hello list,

I need to create queries dynamically and seem to have run into a silly
showstopper. I want to create queries like ...

 SELECT tmp1.name, tmp1.value  as "{foo}bar" from baz  as tmp1;

... where both the table name alias ("tmp1") as well as the field name
alias ("{foo}bar") are _Strings_. Now, if i do:

 (sql-compile '(:dot "tmp1" value))
 --> "E'tmp1'.value"

which is definitely _not_ what i want. The same for:

 (sql-compile '(:as "tmp1" baz))
 --> "E'tmp1' AS baz"

If possible at all I'd like to avoid the (make-symbol "tmp1") and
similar workarrounds. Any ideas?

 TIA Ralf Mattes
 



From marijnh at gmail.com  Wed Jan 23 20:16:57 2008
From: marijnh at gmail.com (Marijn Haverbeke)
Date: Wed, 23 Jan 2008 21:16:57 +0100
Subject: [postmodern-devel] dynamically built queries
In-Reply-To: <1201118759.6570.21.camel@localhost.localdomain>
References: <1201118759.6570.21.camel@localhost.localdomain>
Message-ID: <befaa7aa0801231216m62cf7f02r2fd344febaa68334@mail.gmail.com>

Hi Ralf,

I think that the :raw keyword can help you out here -- you give it a
string, which is then inserted straight into the query, as in (sql
(:select (:raw "tmp1.name") :from (:as 'baz (:raw "tmp1")))).

Cheers,
Marijn



On Jan 23, 2008 9:05 PM, Ralf Mattes <rm at seid-online.de> wrote:
> Hello list,
>
> I need to create queries dynamically and seem to have run into a silly
> showstopper. I want to create queries like ...
>
>  SELECT tmp1.name, tmp1.value  as "{foo}bar" from baz  as tmp1;
>
> ... where both the table name alias ("tmp1") as well as the field name
> alias ("{foo}bar") are _Strings_. Now, if i do:
>
>  (sql-compile '(:dot "tmp1" value))
>  --> "E'tmp1'.value"
>
> which is definitely _not_ what i want. The same for:
>
>  (sql-compile '(:as "tmp1" baz))
>  --> "E'tmp1' AS baz"
>
> If possible at all I'd like to avoid the (make-symbol "tmp1") and
> similar workarrounds. Any ideas?
>
>  TIA Ralf Mattes
>
>
> _______________________________________________
> postmodern-devel mailing list
> postmodern-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/postmodern-devel
>