- postmodern-devel - mailman3.common-lisp.net

[postmodern-devel] Postmodern and Windows
by Daniel Brunner 14 Mar '12

14 Mar '12

Hello, I have flipped through the mailing list's archive but did not find a solution: Is Postmodern known to run under Windows? I have several PCs with Windows (XP and 7) where I use Clozure CL but Postmodern is not working. My setup: - Windows XP or Windows 7 (32 Bit) - The database is local (9.1.3) or via the network (8.4). - I use the newest Postmodern via the March distribution of Quicklisp - I user Clozure CL version 1.7 I can do a CONNECT-TOPLEVEL and with the pgadmin3 tool I can see the connection. But after a simple query like (QUERY "select 10") I get an error: --8<-- Database error: on #<BASIC-TCP-STREAM: CLOSED #xCB1E1BE>: Das Geraet erkennt den Befehl nicht. (error #22) during write [Condition of type CL-POSTGRES:DATABASE-SOCKE-ERROR --8<-- The "Das Geraet erkennt den Befehl nicht" would be something like "The device doesn't recognise the command." in English. I do not have the original error message at hand. In the PostgreSQL-Log I see the message "ungueltige Message-Laenge". That would be something like "invalid message length". Am I doing something wrong? Wrong configuration? Encoding issue? I have no idea how to fix that problem. Kind regards, Daniel.

3 4

[postmodern-devel] Regression in handling integer dates and timestamps
by Lucas Hope 14 Feb '12

14 Feb '12

Hi there, The following code sample used to work on postmodern-20110219-git (using quicklisp from that period). However it now breaks. I can no longer insert integers as dates or timestamps into tables. Updates don't work either, but the example below uses insert: (ql:quickload "postmodern") ;; quicklisp version released yesterday - postmodern-20111203-git (pomo:connect-toplevel "date_test" "snip" "snip" "localhost") (gethash "integer_datetimes" (pomo::connection-parameters pomo:*database*)) => "on" => T (pomo:execute "create table date ( value date );") (pomo:execute "insert into date ( value ) values ( $1 )" "2010-01-01") (setf date (pomo:query "select * from date limit 1" :single)) => 3471292800 (ignore-errors (pomo:execute "insert into date ( value ) values ( $1 )" date)) => NIL => #<CL-POSTGRES-ERROR:INVALID-DATETIME-FORMAT {1004A00391}> (pomo:execute "create table timestamp ( value timestamp );") (pomo:execute "insert into timestamp ( value ) values ( $1 )" "2010-01-01") 1 1 (setf timestamp (pomo:query "select * from timestamp limit 1" :single)) => 3471292800 (ignore-errors (pomo:execute "insert into timestamp ( value ) values ( $1 )" timestamp)) => NIL => #<CL-POSTGRES-ERROR:INVALID-DATETIME-FORMAT {100403F391}> -- --------------------------------------------------- Dr Lucas Hope - lucas.r.hope@skype Machine Learning and Software Engineering Consultant Melbourne, Australia

2 4

[postmodern-devel] "WHERE id IN (1, 2, 3)" conditions
by Nicolas Martyanoff 28 Dec '11

28 Dec '11

Hi, I'm trying to use queries such as: SELECT a, b, c FROM foo WHERE id IN (1, 2, 3) While it's possible to nest queries with IN: (:select 'a 'b 'c :from 'foo :where (:in 'id (:select 'd :from 'bar))) It's not possible to have: (:select 'a 'b 'c :from 'foo :where (:in 'id (list 1 2 3)) => Value (1 2 3) can not be converted to an SQL literal. Would you be ok with lists being converted to (a, b, ...) SQL literals ? Regards, -- Nicolas Martyanoff http://codemore.org khaelin(a)gmail.com

2 2

[postmodern-devel] OSX 10.6.8 Intel - SBCL 1.0.54 The function cl-postgres::write-uint4 is undefined.
by Carr Colin 20 Dec '11

20 Dec '11

Hi, having some difficulty with OSX 10.6.8 Intel - SBCL 1.0.54 I used quicklisp (most current) to install and my debian (6.0.3) boxes are fine but my Macbook Pro (10.6.8) is having the issue. The same error is happening when I use; postgresql 8.4 or 9.1 locally (running on OSX) or remote to a debian box running 8.4 and even a VM (VirtualBox) running debian (6.0.3) on the Mac with NAT forwarding to postgres 8.4. I can connect with psql in the various configurations described for the remote connection I made the appropriate "all/all/trusted" additions to the "/etc/" files so it is unencumbered by any security. I deleted my FASLs and I tried connecting without a database running and the FASL removal and "re-load" did not solve the issue and the connect-toplevel without a database running error was what I expected. The Postgres is a "clean" install with only the ytestdb added on all systems I tested against. ytestdb is the database postgres is the user post is the password ;;ERROR cl-postgre::write-uint4 undefined (messages.lisp) I hope this information is helpful. Thanks, - C cl-user> (asdf:oos 'asdf:load-op :postmodern) #<asdf:load-op nil {1003053F83}> nil cl-user> (use-package :postmodern) t cl-user> (connect-toplevel "ytestdb" "postgres" "post" "localhost") ; Evaluation aborted on #<cl-postgres:database-socket-error {100874D7A3}>. ;;; THIS IS WITH NO POSTGRES RUNNING cl-user> (connect-toplevel "ytestdb" "postgres" "post" "localhost") The function cl-postgres::write-uint4 is undefined. [Condition of type undefined-function] Restarts: 0: [retry] Retry SLIME REPL evaluation request. 1: [*abort] Return to SLIME's top level. 2: [terminate-thread] Terminate this thread (#<thread "repl-thread" running {1002BE0043}>) Backtrace: 0: ("bogus stack frame") [No Locals] 1: (cl-postgres::startup-message #<unavailable argument> #<unavailable argument> #<unavailable argument>) Locals: sb-debug::arg-0 = :<not-available> sb-debug::arg-1 = :<not-available> sb-debug::arg-2 = :<not-available> 2: (cl-postgres::authenticate #<unavailable argument> "postgres" "post" "ytestdb" :no) Locals: sb-debug::arg-0 = :<not-available> sb-debug::arg-1 = "postgres" sb-debug::arg-2 = "post" sb-debug::arg-3 = "ytestdb" sb-debug::arg-4 = :no 3: (cl-postgres::initiate-connection #<database-connection {10089CD923}>) Locals: sb-debug::arg-0 = #<database-connection {10089CD923}> 4: (cl-postgres:open-database #<unavailable argument> #<unavailable argument> #<unavailable argument> #<unavailable argument> #<unavailable argument> #<unavailable argument>) Locals: cl-postgres::database = :<not-available> cl-postgres::host = :<not-available> cl-postgres::password = :<not-available> cl-postgres::port = :<not-available> cl-postgres::use-ssl = :<not-available> cl-postgres::user = :<not-available> 5: (connect-toplevel "ytestdb" "postgres" "post" "localhost" :port 5432 :use-ssl nil) Locals: postmodern::database = "ytestdb" postmodern::host = "localhost" #:n-supplied-32 = nil postmodern::password = "post" postmodern::user = "postgres" #:|port-DEFAULTING-TEMP| = 5432 #:|use-ssl-DEFAULTING-TEMP| = nil 6: (sb-int:simple-eval-in-lexenv (connect-toplevel "ytestdb" "postgres" "post" "localhost") #<NULL-LEXENV>) Locals: sb-debug::arg-0 = (connect-toplevel "ytestdb" "postgres" "post" "localhost") sb-debug::arg-1 = #<NULL-LEXENV> 7: (eval (connect-toplevel "ytestdb" "postgres" "post" "localhost")) Locals: sb-debug::arg-0 = (connect-toplevel "ytestdb" "postgres" "post" "localhost") 8: (swank::eval-region "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\")\n") Locals: sb-debug::arg-0 = "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\")\n" 9: ((lambda () :in swank::repl-eval)) [No Locals] 10: (swank::track-package #<CLOSURE (lambda # :in swank::repl-eval) {10089CD43B}>) Locals: sb-debug::arg-0 = #<CLOSURE (lambda () :in swank::repl-eval) {10089CD43B}> 11: (swank::call-with-retry-restart "Retry SLIME REPL evaluation request." #<CLOSURE (lambda # :in swank::repl-eval) {10089CD35B}>) Locals: sb-debug::arg-0 = "Retry SLIME REPL evaluation request." sb-debug::arg-1 = #<CLOSURE (lambda () :in swank::repl-eval) {10089CD35B}> 12: (swank::call-with-buffer-syntax nil #<CLOSURE (lambda # :in swank::repl-eval) {10089CD32B}>) Locals: sb-debug::arg-0 = nil sb-debug::arg-1 = #<CLOSURE (lambda () :in swank::repl-eval) {10089CD32B}> 13: (swank::repl-eval "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\")\n") Locals: sb-debug::arg-0 = "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\")\n" 14: (sb-int:simple-eval-in-lexenv (swank:listener-eval "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\") ..) Locals: sb-debug::arg-0 = (swank:listener-eval ..) sb-debug::arg-1 = #<NULL-LEXENV> 15: (eval (swank:listener-eval "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\") ..) Locals: sb-debug::arg-0 = (swank:listener-eval ..) 16: (swank:eval-for-emacs (swank:listener-eval "(connect-toplevel \"ytestdb\" \"postgres\" \"post\" \"localhost\")\n") "COMMON-LISP-USER" 40) Locals: sb-debug::arg-0 = (swank:listener-eval ..) sb-debug::arg-1 = "COMMON-LISP-USER" sb-debug::arg-2 = 40 17: (swank::process-requests nil) Locals: sb-debug::arg-0 = nil 18: ((lambda () :in swank::handle-requests)) 19: ((lambda () :in swank::handle-requests)) 20: (swank-backend::call-with-break-hook #<FUNCTION swank:swank-debugger-hook> #<CLOSURE (lambda # :in swank::handle-requests) {1002BF011B}>) 21: ((flet swank-backend:call-with-debugger-hook :in "/Users/cpc26/quicklisp/dists/quicklisp/software/slime-20111105-cvs/swank-sbcl.lisp") ..) 22: (swank::call-with-bindings ..) 23: (swank::handle-requests #<swank::connection {1003DA2BF3}> nil) 24: ((flet #:without-interrupts-body-243284 :in sb-thread:make-thread)) 25: ((flet sb-thread::with-mutex-thunk :in sb-thread:make-thread)) 26: ((flet #:without-interrupts-body-88835 :in sb-thread::call-with-mutex)) 27: (sb-thread::call-with-mutex ..) 28: (sb-thread::initial-thread-function) 29: ("foreign function: call_into_lisp") 30: ("foreign function: new_thread_trampoline") 31: ("foreign function: _pthread_start") 32: ("foreign function: thread_start")

2 1

[postmodern-devel] Patch: Make :|| (symbol with an empty name) equivalent to :\|\|
by Stas Boukarev 21 Nov '11

21 Nov '11

:|| is a valid symbol in CL, its name is an empty string, "", and it would be much more pleasant to use it for concatenation instead of :\|\|. The attached patch adds :|| operator. -- With best regards, Stas.

3 5

[postmodern-devel] <> operator
by Zach Beane 18 Nov '11

18 Nov '11

Out of SQL habit I used <> for inequality yesterday and got this error: function __(integer, integer) does not exist Should I just stick with !=? Zach

3 2

[postmodern-devel] cl-postgres fails to read an empty array
by Andy Chambers 03 Nov '11

03 Nov '11

Hi, I'm not sure if this is the case in the latest version of postgresql but in my version, it fails to read an empty integer correctly from the DB. The following patch fixes for me... diff --git Postmodern/cl-postgres/interpret.lisp Postmodern/cl-postgres/\ interpret.lisp index 7d30f75..a6ce46f 100644 --- Postmodern/cl-postgres/interpret.lisp +++ Postmodern/cl-postgres/interpret.lisp @@ -196,8 +196,13 @@ used. Correct for sign bit when using integer format." (unless (< (incf pos) (length value)) (return ws))) 'simple-vector))))) +(defun maybe-parse-integer (s) + (if (string= "" s) + :null + (parse-integer s))) + ;; Integral array types -(let ((read-integral (read-array-value #'parse-integer))) +(let ((read-integral (read-array-value #'maybe-parse-integer))) (dolist (oid '(1561 1005 1007 1016 1028)) (set-sql-reader oid read-integral))) Cheers, Andy

2 1

[postmodern-devel] Release 1.18
by Marijn Haverbeke 19 Oct '11

19 Oct '11

I just tagged the current version as release 1.18. See the project page [1] for a summary of changes. [1]: http://marijnhaverbeke.nl/postmodern

1 0

[postmodern-devel] setting default value for serial column in prepared insert statement execution
by Antony 05 Sep '11

05 Sep '11

Hi I was trying to find how to use the 'serial' (auto increment) type through cl-postgres prepared insert statement execution. According to http://www.postgresql.org/docs/9.0/interactive/datatype-numeric.html#DATATY… I basically have two choices for defaulting a serial column value 1. pass the DEFAULT keyword for the column value 2. do not include the serial column in the insert statement The way my code is setup, it's more convoluted to do #2 I wanted to check if it's possible to do #1. That would really make a big difference to me. -Antony

4 5

[postmodern-devel] cl-postgress build warning due to multiple loading
by Antony 13 Aug '11

13 Aug '11

Hi I am beginning to use cl-postgress part of postmodern It is working well, Thank you. I have an issue that would be nice to fix. When I do (asdf:oos 'asdf:load-op :cl-postgres) I see couple of warnings like ;Compiler warnings for "/media/sf_git/thirdparty/postmodern/simple-date/cl-postgres-glue.lisp" : ; In an anonymous lambda form at position 857: Duplicate definitions of (:METHOD CL-POSTGRES:TO-SQL-STRING (DATE)), in this file and in "/media/sf_git/thirdparty/postmodern/simple-date/cl-postgres-glue.lisp" I think they are due to in file cl-postgress.asd at the end you have (asdf:oos 'asdf:load-op :simple-date-postgres-glue) and again in simple-date.asd you have the same line (asdf:oos 'asdf:load-op :simple-date-postgres-glue) Can this be fixed. I am using CCL Clozure Common Lisp Version 1.7-dev-r14812M-trunk (LinuxX8664)! My ASDF is a bit old (2.011.9), asdf is one hting I don't like upgrading, causes me a lot of pain. -Antony

3 6