convert dbm to mySql

[ThoughtWorker]

I want to upgrade phpWiki to 1.19 and use mySql to store the data.  I am currently using the build dbm.  How can I populate the data from dbm to mySql

[Steve Wainstead]

The admin tools (try loading admin.php) offered a way of doing this, by dumping the Wiki out as a zip file, and then reloading the zip file. It hasn't been tested recently and you have to have zip support compiled into PHP.

~swain

[Reini Urban]

>dumping the Wiki out as a zip file, and
>then reloading the zip file.

I have now the zip as mail archive and the serialized dump in a seperate directory.

(just added these lines to browse.html)

###IF:ADMIN###
<form action="###SCRIPTURL###" method=POST>
<input type=text name=dumpserial><input type=submit value=Dump></form>
<form action="###SCRIPTURL###" method=POST>
<input type=text name=loadserial><input type=submit value=Load></form>
[<a href="###SCRIPTURL###?zip=all">Zip as mailarchive</A>]
###ENDIF:ADMIN###

Q1: how to import the zip? I see no mail reader.

Q2: how to effectively load a dir of flatfiles into my db (mysql), as current user? Those are much easier to maintain and write in the first instance.

To import it as phpwiki admin I simply put them
into pgsrc and drop the tables by
  $mysql <schemas/schema.mysql
but I want them to appear under my name :)
Some kind of LoadFlatDir as combination of
lib/setupwiki.php and admin/loadserial.php

[Reini Urban]

Sorry, just found the answers to both questions:
Load setupwiki.php with WIKI_PGSRC set to the directory.

this is quite dirty btw.
one should be able to load setupwiki.php to get the functions without loading pgsrc
I just have to find out now, how to use
LoadZipOrDir($dbi, $mydir) as admin author.
I know, it's easy, but this will need a bigger patch to setupwiki.php and all its loaders.

I've made this minor convenience patch:
--- phpwiki/lib/setupwiki.php~  Sun Oct 22 19:33:35 2000
+++ phpwiki/lib/setupwiki.php   Fri Feb  9 23:51:05 2001
@@ -87,10 +87,12 @@
     {
       $handle = opendir($dir = $zip_or_dir);

-      // load default pages
+      // load default pages, ignore backup's
       while ($fn = readdir($handle))
        {
          if (filetype("$dir/$fn") != 'file')
+             continue;
+         if (substr("$fn",-1) == "~")
              continue;
          $stat = stat("$dir/$fn");
          $mtime = $stat[9];

[Geoffrey T. Dairiki]

Yes this is a little clumsy and non-obvious.
I'm currently working on making it better (at least
in the development branch.)

In the mean time, here's how I suggest you

Get either a zip dump, or a serialized dump from your
old wiki.

(If your using a zip dump, it must (for now) be
on the server machine.  Copy it there if it's not
there yet.)

Edit lib/config.php in the new wiki and set WIKI_PGSRC
to point either to your zip file, or the the serialized
dump directory.

Blow away the database of your new wiki.
(Delete the db files for a db base wiki,
or do something like:
  mysql -u <user> -p <database> < schemas/schema.mysql
to blow away a mysql based wiki.

Now with your web browser, go browse to
the FrontPage of your new wiki.  It will detect
that it's working with a virgin database and
will pull in the pages from the source
specified in WIKI_PGSRC.

Jeff

[Arno Hollosi]

Alternatively to Steve's suggestion, you can
use the dump & load serialized pages functions.

Furthermore, I suggest using the latest nightly tarball instead of 1.1.9. There were some important bug fixes since then.

/Arno