phpwiki-talk

[Phpwiki-talk] tab-width and new database API

[Jeff D. <da...@da...>]

On Aug 28, 2001, Steve Wainstead said:
> > (defun my-php-mode-hook-func ()
> >    (c-set-mode "gnu")
> >    (setq tab-width 4
....

I really dislike this tab-width = 4 thing.
It results in the source code looking mangled when viewed
by viewers which think the tabs stops should be every eight
columns (like most of them.)

Could we change our "coding standard" so that it departs from
PEARs in this one respect?

Jeff

BTW, I've been working feverishly on a new database API.
It's changed somewhat compared to what's outlined in the
PhpWikiWiki.

A brief synopsis, to whet your appetite:
  The basic model now uses three basic objects: WikiDatabase,
  WikiPage, and WikiPageRevision.

  A WikiDatabase is an infinite set of WikiPages.  All possible
  pages are always present in a WikiDatabase.  (Pages which have
  not been explicitly initialized have only one revision,
  which contains some default content,.e.g. "Describe 
  [PageName] here.") 
  
  As such WikiDatabases have read-only semantics.  About
  all you can do with them is fetch WikiPages from them.  (Of course,
  there are various search methods which can be used to finds 
  sets of WikiPages.)  You never have to add pages to
  a WikiDatabase, since they're all already there.

  WikiPages have methods to get and set page-level meta-data
  (e.g. is the page locked?).
  Additionally, WikiPages are containers for WikiPageRevisions.
  A WikiPage always contains at least one WikiPageRevision:
  the default revision (e.g.  "Describe [ThisPage] here.")
  There are methods to get WikiPageRevisions from WikiPages,
  as well as methods to create new revisions and to
  delete old ones.

  WikiPageRevisions have methods to access the page content, as well
  as revision-level meta-data (e.g. author, revision time).
  WikiPageRevisions have read-only semantics: you can not modify
  a revision, you may only create new ones (and delete old ones).

The (PEAR::DB-based) MySQL backend is roughly working.
I hope to have a dba based backend working by the end of the week,
at which time I'll check my hacks into the CVS so you can look at it.

As these are major changes, I think I'll check them into a new
branch ('Jeff_Hacks_Again_Branch' ?) to give you all a chance to veto
the changes if you don't like them before they're dropped into
the main branch.

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

That did it. Thanks!


~swain

On Tue, 28 Aug 2001, Jeff Dairiki wrote:

> On Aug 28, 2001, Steve Wainstead said:
> > On Tue, 28 Aug 2001, Jeff Dairiki wrote:
> >
> > > (defun my-php-mode-hook-func ()
> > >    (c-set-mode "gnu")
> > >    (setq tab-width 4
> > >          c-basic-offset 4
> > >          c-hanging-comment-ender-p nil
> > >          indent-tabs-mode nil))
> > >
> > > (add-hook 'php-mode-hook 'my-php-mode-hook-func)
> >
> > Success! Except for (c-set-mode "gnu") which causes the error:
> >
> > File mode specification error: (void-function c-set-mode)
>
> Wups.  Try (c-set-style "gnu").
> Jeff
>
>
>
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
>

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Jeff D. <da...@da...>]

On Aug 28, 2001, Steve Wainstead said:
> On Tue, 28 Aug 2001, Jeff Dairiki wrote:
> 
> > (defun my-php-mode-hook-func ()
> >    (c-set-mode "gnu")
> >    (setq tab-width 4
> >          c-basic-offset 4
> >          c-hanging-comment-ender-p nil
> >          indent-tabs-mode nil))
> >
> > (add-hook 'php-mode-hook 'my-php-mode-hook-func)
> 
> Success! Except for (c-set-mode "gnu") which causes the error:
> 
> File mode specification error: (void-function c-set-mode)

Wups.  Try (c-set-style "gnu").
Jeff

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Reini U. <ru...@x-...>]

Steve Wainstead schrieb:
> (setq load-path (append load-path (list "~swain/.elisp")))
> (load-library "php-mode-099")
> 
> (defun php-mode-hook ()
>   (setq tab-width 4
>         c-basic-offset 4
>         c-hanging-comment-ender-p nil
>         indent-tabs-mode
>         (not
>          (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
>               (string-match "\.php$" (buffer-file-name))
>               (string-match "\.phtml$" (buffer-file-name))))))
> ;; end file
> 
> I'm using php-mode-099 from Sourceforge (http://php-mode.sourceforge.net).
> I was previously using a php-mode from a guy at the ACM. There is a
> php-mode-pear-hook in php-mode-099 but my Emacs knowledge is not good
> enough to figure out how to run it...

since php mode simply inherits all these settings from c-mode it must 
be something wrong there. 

aah, my php-mode.el is the old one from Fred Yankowski <fc...@ac...>
$Id: php-mode.el,v 1.24 2000/12/08 17:44:17 fred Exp $
this works fine with mmm-mode. the new one crashes on my old beta xemacs....

  I stopped fixing things with GNU emacs. 
  stallman made me sick, in xemacs everything is fixed immediately. 
  though speedbar never really worked fine on xemacs.

re hooks:
I load a number of hooks like this:
;;; fontify these automatically
(mapcar '(lambda (hook) (add-hook hook 'turn-on-font-lock))
	'(emacs-lisp-mode-hook dired-mode-hook lisp-mode-hook
	  html-mode-hook hm--ht--html-mode-hook cperl-mode-hook 
	  perl-mode-hook autolisp-mode-hook c-mode-hook
	  php-mode-hook sql-mode-hook))

for a single mode, like this:
(add-hook 'text-mode-hook 'flyspell-mode)

and 
  (add-hook 'php-mode-hook 'my-php-mode-hook-func) ; wrong
is wrong. it should be:
  (add-hook 'php-mode-hook 'php-mode)

but if it has the same name (like php-mode-hook for php-mode) it 
is loaded automatically. it only should be defined, before 
you load the library.
in the meantime I have to wait until my xemacs bug is fixed.
-- 
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

On Tue, 28 Aug 2001, Jeff Dairiki wrote:

> (defun my-php-mode-hook-func ()
>    (c-set-mode "gnu")
>    (setq tab-width 4
>          c-basic-offset 4
>          c-hanging-comment-ender-p nil
>          indent-tabs-mode nil))
>
> (add-hook 'php-mode-hook 'my-php-mode-hook-func)

Success! Except for (c-set-mode "gnu") which causes the error:

File mode specification error: (void-function c-set-mode)

~swain


---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Jeff D. <da...@da...>]

On Aug 28, 2001, Steve Wainstead sai> On Tue, 28 Aug 2001, Reini Urban 
wrote:
> 
> I tried this .emacs file on netbsd and red hat linux, and it flat out does
> not work. So something is missing either from php-mode-099 or the .emacs
> file.

You're defining the php-mode-hook as a function.  That looks wrong.
php-mode-hook is a variable which contains a list of functions to run.
(Note also that in your code, indent-tabs-mode always gets set true,
since (buffer-file-name) can't match both "\.php$" and "\.phtml$".)
(I've also added a call to (c-set-mode) to ensure the correct defaults.)

Try this (untested, of course):

...
(defun my-php-mode-hook-func ()
   (c-set-mode "gnu")
   (setq tab-width 4
         c-basic-offset 4
         c-hanging-comment-ender-p nil
         indent-tabs-mode nil))

(add-hook 'php-mode-hook 'my-php-mode-hook-func)
...

Jeff.

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

On Tue, 28 Aug 2001, Reini Urban wrote:

> XEmacs 21.2-b37 "Pan" configured for `i586-pc-win32'.

Heathen!! ;-)

I tried this .emacs file on netbsd and red hat linux, and it flat out does
not work. So something is missing either from php-mode-099 or the .emacs
file.

;; snip....

(setq load-path (append load-path (list "~swain/.elisp")))
(load-library "php-mode-099")

(defun php-mode-hook ()
  (setq tab-width 4
        c-basic-offset 4
        c-hanging-comment-ender-p nil
        indent-tabs-mode
        (not
         (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
              (string-match "\.php$" (buffer-file-name))
              (string-match "\.phtml$" (buffer-file-name))))))
;; end file


I'm using php-mode-099 from Sourceforge (http://php-mode.sourceforge.net).
I was previously using a php-mode from a guy at the ACM. There is a
php-mode-pear-hook in php-mode-099 but my Emacs knowledge is not good
enough to figure out how to run it...

~swain

>
> php-mode: C-h a c-basic-offset
>
> `c-basic-offset' (buffer: *Hyper Apropos*, mode: Hyper-Apropos)
> Customizable user variable:
>   value: 2
>   *Amount of basic offset used by + and - symbols in `c-offsets-alist'.
>
> Property-list:
>   custom-requests: 		nil
>   custom-type: 			integer
>   standard-value: 		(4)
>   saved-variable-comment: 	nil
>   theme-value: 			((user set 4))
>   saved-value: 			(4)
>
>
> Steve Wainstead schrieb:
> > Right, this is on the PEAR page, but I put it in my .emacs and it didn't
> > work. Which is to say, I open a php file and c-basic-offset is 3 locally
> > and 2 globally, indent-tabs-mode is t, and so on. I'll try it again later
> > tonight.
> >
> > ~swain
> >
> > On Mon, 27 Aug 2001, Reini Urban wrote:
> >
> > > (defun php-mode-hook ()
> > >   (setq tab-width 4
> > >       c-basic-offset 4
> > >       c-hanging-comment-ender-p nil
> > >       indent-tabs-mode
> > >       (not
> > >        (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
> > >             (string-match "\.php$" (buffer-file-name))
> > >             (string-match "\.phtml$" (buffer-file-name))))))
>
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
>

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

RE: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

I have, but my choice is to go with something that will ship with PHP in
the future, as PEAR and DB do already (though it always seems to be broken
in the latest release ;-)

DB will give us everything we need and as coders we will be using the
latest tools that come with PHP; there has to be something in this for us
too, after all :-)

Also see
http://phpwiki.sourceforge.net/phpwiki/index.php?PhpDatabaseAccessLibraries

~swain


On Tue, 28 Aug 2001, Pablo Roca wrote:

> > 1. Write the abstract base class (the interface definition)
> > 2. Write a class that uses DB and extends the base class
> > 3. Integrate that into the code base (basically, search and
> > replace all database calls) 4. Test, test, test.
>
> Hum ..... Have you considered in using ADODB for database
> Abstraction? I think is a cool work.
>
> See at:
>
> http://php.weblogs.com/adodb
>
> Regards,
>
> Pablo Roca
> Sysop de PortalFox (http://www.portalfox.com)
>
>
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
>

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Reini U. <ru...@x-...>]

XEmacs 21.2-b37 "Pan" configured for `i586-pc-win32'.

php-mode: C-h a c-basic-offset

`c-basic-offset' (buffer: *Hyper Apropos*, mode: Hyper-Apropos)
Customizable user variable:
  value: 2
  *Amount of basic offset used by + and - symbols in `c-offsets-alist'.

Property-list:
  custom-requests: 		nil
  custom-type: 			integer
  standard-value: 		(4)
  saved-variable-comment: 	nil
  theme-value: 			((user set 4))
  saved-value: 			(4)


Steve Wainstead schrieb:
> Right, this is on the PEAR page, but I put it in my .emacs and it didn't
> work. Which is to say, I open a php file and c-basic-offset is 3 locally
> and 2 globally, indent-tabs-mode is t, and so on. I'll try it again later
> tonight.
> 
> ~swain
> 
> On Mon, 27 Aug 2001, Reini Urban wrote:
> 
> > (defun php-mode-hook ()
> >   (setq tab-width 4
> >       c-basic-offset 4
> >       c-hanging-comment-ender-p nil
> >       indent-tabs-mode
> >       (not
> >        (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
> >             (string-match "\.php$" (buffer-file-name))
> >             (string-match "\.phtml$" (buffer-file-name))))))

RE: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Pablo R. <pr...@cl...>]

> 1. Write the abstract base class (the interface definition)
> 2. Write a class that uses DB and extends the base class
> 3. Integrate that into the code base (basically, search and 
> replace all database calls) 4. Test, test, test.

Hum ..... Have you considered in using ADODB for database
Abstraction? I think is a cool work.

See at:

http://php.weblogs.com/adodb

Regards,

Pablo Roca
Sysop de PortalFox (http://www.portalfox.com)

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

Right, this is on the PEAR page, but I put it in my .emacs and it didn't
work. Which is to say, I open a php file and c-basic-offset is 3 locally
and 2 globally, indent-tabs-mode is t, and so on. I'll try it again later
tonight.

~swain

On Mon, 27 Aug 2001, Reini Urban wrote:

> (defun php-mode-hook ()
>   (setq tab-width 4
> 	c-basic-offset 4
> 	c-hanging-comment-ender-p nil
> 	indent-tabs-mode
> 	(not
> 	 (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
> 	      (string-match "\.php$" (buffer-file-name))
> 	      (string-match "\.phtml$" (buffer-file-name))))))
>
>
> Jeff Dairiki schrieb:
> >
> > > Will do... Postgresql is good; I created "frontpage" and it's different
> > > from FrontPage in the pagename column. I'll check mSQL next week.
> >
> > Thanks!
> >
> > > 1. Write the abstract base class (the interface definition)
> > > 2. Write a class that uses DB and extends the base class
> > > 3. Integrate that into the code base (basically, search and replace all
> > > database calls)
> > > 4. Test, test, test.
> >
> > Okay, I'll work on it now.  (I promise.)  I hope to have something
> > basically working by the end of the week.
> >
> > > The reason I stopped working on the abstract base class (aside from being
> > > in Florida ;-) was I couldn't get Emacs to format the code according to
> > > PEAR. I posted a question to gnu.emacs.help and got no answer, which is
> > > odd.
> >
> > I've recently been using:
> >  <?php
> >  // Local Variables:
> >  // mode: php
> >  // eval: (c-set-style "cc-mode")
> >  // End:
> >  ?>
> > Which seems to work.  It's a bit of a pain, since emacs now asks if you
> > really want to eval the elisp code each time you load the file.
>
> use the hook instead. eval in user docs is evil!
> --
> Reini Urban
> http://xarch.tu-graz.ac.at/home/rurban/
>
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
>

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Reini U. <ru...@x-...>]

(defun php-mode-hook ()
  (setq tab-width 4
	c-basic-offset 4
	c-hanging-comment-ender-p nil
	indent-tabs-mode
	(not
	 (and (string-match "/\\(PEAR\\|pear\\)/" (buffer-file-name))
	      (string-match "\.php$" (buffer-file-name))
	      (string-match "\.phtml$" (buffer-file-name))))))


Jeff Dairiki schrieb:
> 
> > Will do... Postgresql is good; I created "frontpage" and it's different
> > from FrontPage in the pagename column. I'll check mSQL next week.
> 
> Thanks!
> 
> > 1. Write the abstract base class (the interface definition)
> > 2. Write a class that uses DB and extends the base class
> > 3. Integrate that into the code base (basically, search and replace all
> > database calls)
> > 4. Test, test, test.
> 
> Okay, I'll work on it now.  (I promise.)  I hope to have something
> basically working by the end of the week.
> 
> > The reason I stopped working on the abstract base class (aside from being
> > in Florida ;-) was I couldn't get Emacs to format the code according to
> > PEAR. I posted a question to gnu.emacs.help and got no answer, which is
> > odd.
> 
> I've recently been using:
>  <?php
>  // Local Variables:
>  // mode: php
>  // eval: (c-set-style "cc-mode")
>  // End:
>  ?>
> Which seems to work.  It's a bit of a pain, since emacs now asks if you
> really want to eval the elisp code each time you load the file.

use the hook instead. eval in user docs is evil!
-- 
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/

Re: [Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Jeff D. <da...@da...>]

> Will do... Postgresql is good; I created "frontpage" and it's different
> from FrontPage in the pagename column. I'll check mSQL next week.

Thanks!

> 1. Write the abstract base class (the interface definition)
> 2. Write a class that uses DB and extends the base class
> 3. Integrate that into the code base (basically, search and replace all
> database calls)
> 4. Test, test, test.

Okay, I'll work on it now.  (I promise.)  I hope to have something
basically working by the end of the week.

> The reason I stopped working on the abstract base class (aside from being
> in Florida ;-) was I couldn't get Emacs to format the code according to
> PEAR. I posted a question to gnu.emacs.help and got no answer, which is
> odd.

I've recently been using:
 <?php
 // Local Variables:
 // mode: php
 // eval: (c-set-style "cc-mode")
 // End:
 ?>
Which seems to work.  It's a bit of a pain, since emacs now asks if you 
really
want to eval the elisp code each time you load the file.

Jeff

[Phpwiki-talk] databases, DB, classs, Emacs, oh my

[Steve W. <sw...@pa...>]

On Mon, 27 Aug 2001, Jeff Dairiki wrote:

> From what I can tell by reading postgres and miniSQL docs, my guess
> is that neither of those backends have the case-insensitivy problem
> (i.e. 'SandBox' and 'SanDbox' are truly different pages.)  Can you
> verify that that is the case?

Will do... Postgresql is good; I created "frontpage" and it's different
from FrontPage in the pagename column. I'll check mSQL next week.

> > What I'd like to set up is an automated test that I wrote on my last
[...]

> Yes, that would be good.  Also would be good to run an HTML validator
> automatically on all pages (or is that done by your test suite too?).

http-unit chokes on nonvalid HTML, and it's pretty strict.

> > Also, I still want to move to DB... perhaps I should just take a short cut
> > and write a db.php replica of mysql.php as a start. Writing the class and
> > changing all database calls is a lot of work...
>
> Should I work on "new" DB API (like I've been promising)?  Or are we
> happy with the hacks to the old API that Joel Uckelman has started?

We most definitely are moving to DB (and possible dbx for kicks, but I
doubt it... since we'll still have to support flat files and dbm). I've
committed the barest of class files to the tree already, so you may want
to start with that. I want to move one step at a time: devise the database
class, and integrate that into the code base. At that point we should do
an alpha release of 1.3. I would think:

1. Write the abstract base class (the interface definition)
2. Write a class that uses DB and extends the base class
3. Integrate that into the code base (basically, search and replace all
database calls)
4. Test, test, test.

The reason I stopped working on the abstract base class (aside from being
in Florida ;-) was I couldn't get Emacs to format the code according to
PEAR. I posted a question to gnu.emacs.help and got no answer, which is
odd.
<URL:http://groups.google.com/groups?q=group:gnu.emacs.help+author:swain%40panix.com&hl=en&safe=off&rnum=1&selm=9edb24c1.0106262044.62d055ce%40posting.google.com>

I've made three attempts to get Emacs to set c-basic-indent to four spaces
to no avail. Most irritating. I can do ESC-: and enter the setq by hand,
and it works; but I can't do it from .emacs or from the variables list
you inserted in the files.

~swain

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

[Phpwiki-talk] PhpWiki on OS X

[Steve W. <sw...@pa...>]

Today I installed PhpWiki 1.2 on OS X release 10.0.0.4. PhpWiki will run
"out of the box" if the installation of PHP is like so:

./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var
 --mandir=/usr/share/man --with-apxs --with-zlib --disable-pear
 --with-pgsql=/usr/local/pgsql --with-ndbm

I was trying to use --enable-dba=shared but that wouldn't work; PHP
claimed there were no dba functions.

I followed all the instructions on:

http://techdocs.postgresql.org/installguides.php#macosx

http://www.oreillynet.com/pub/a/mac/2001/07/20/php_macosx.html

OS X ships with the SDBM library, which is compatible with the NDBM API.
It doesn't seem to suffer from NDBM's limit of 1000 bytes per value (which
limits page length to 1000 chars).

I haven't switched it to Postgresql but that should be trivial.

~swain

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] 1.2.1 task list

[Steve W. <sw...@pa...>]

On Fri, 17 Aug 2001, Jeff Dairiki wrote:

> > How drastic?
>
> I've really only thought about the MySQL backend so far.
> There, columns which hold pagenames need to have the 'binary'
> attribute set (their type needs to be changed from
> 'VARCHAR(100) NOT NULL' to 'VARCHAR(100) BINARY NOT NULL'.)
> I think this can be done without necessitating a full
> dump/restore of the database.  If someone "forgot" to do
> this when they upgraded to 1.2.1, I think the worst that would
> happen is that his/her wiki would continue to have case
> insensitivity bugs.

Sounds fine... as long as it doesn't require hours of work on your part,
I'd say go ahead.

> The flat-file back-end has problems on Win32 systems, because
> the stupid filesystem is not case sensitive.  The best fix
> I can think of for that is to urlencode() the capital letters
> in the page names when forming file names.  (This would
> probably require a full dump/restore of the database when
> upgrading.)

It's so painful to support Windows systems regardless of type. I'm
inclined to say "You're on your own" if you run it on Windows... sigh. Of
course it's not fair to them. A lot of people have no choice in the
matter.

> I'm no expert on pgsql and msql --- they may not even be broken.
> If they are, I suspect they could be fixed in a manner similar
> to the mysql back-end.

I haven't set up 1.2 on pgsql in a while and I'm having problems vis. the
pgsql setup. Hopefully I'll work it out tomorrow or Sunday. Gotta go scuba
diving early tomorrow morning.

~swain

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] 1.2.1 task list

[Jeff D. <da...@da...>]

On Aug 17, 2001, Steve Wainstead said:
> 
> In two weeks I will be back in Manhattan, where my other Linux box is with
> msql... I will test it then. (Don't feel like installing it yet again and
> recompiling && etc.)
> 

That's fine.

> > I would still like to fix the page name case-sensitivity issues, but
> > would first like to hear from Steve and others how kosher you think
> > schema changes are within 1.2.x.
> 
> How drastic?

I've really only thought about the MySQL backend so far.
There, columns which hold pagenames need to have the 'binary'
attribute set (their type needs to be changed from
'VARCHAR(100) NOT NULL' to 'VARCHAR(100) BINARY NOT NULL'.)
I think this can be done without necessitating a full
dump/restore of the database.  If someone "forgot" to do
this when they upgraded to 1.2.1, I think the worst that would
happen is that his/her wiki would continue to have case
insensitivity bugs.

 
I don't think the dbm/dba backend is broken, so no problems there.

The flat-file back-end has problems on Win32 systems, because
the stupid filesystem is not case sensitive.  The best fix
I can think of for that is to urlencode() the capital letters
in the page names when forming file names.  (This would
probably require a full dump/restore of the database when
upgrading.)

I'm no expert on pgsql and msql --- they may not even be broken.
If they are, I suspect they could be fixed in a manner similar
to the mysql back-end.

Jeff

Re: [Phpwiki-talk] 1.2.1 task list

[Steve W. <sw...@pa...>]

On Fri, 17 Aug 2001, Jeff Dairiki wrote:

> The msql and pgsql back-ends are, at this point, completely untested.
> (And therefore, likely to be broken.)  If any of you are able to test
> either, please do!

In two weeks I will be back in Manhattan, where my other Linux box is with
msql... I will test it then. (Don't feel like installing it yet again and
recompiling && etc.)

> I would still like to fix the page name case-sensitivity issues, but
> would first like to hear from Steve and others how kosher you think
> schema changes are within 1.2.x.

How drastic?

cheers
~swain


---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] 1.2.1 task list

[Jeff D. <da...@da...>]

On Aug 17, 2001, Jeff  said:
> On Aug 17, 2001, Reini Urban said:
> > 
> > Task "Reimplement Backlinks page"
> 
> I've been planning on doing this ...

Okay, well, I just did it.

There is now a real back-link page in the _stable_ branch 
('release-1_2-branch').

It "should" be working with all the database back-ends.  I've 
more-or-less tested the mysql, dba, dbm, and flat-file back-ends.

The msql and pgsql back-ends are, at this point, completely untested.  
(And therefore, likely to be broken.)  If any of you are able to test 
either, please do!


I would still like to fix the page name case-sensitivity issues, but 
would first like to hear from Steve and others how kosher you think 
schema changes are within 1.2.x.

Jeff

Re: [Phpwiki-talk] 1.2.1 task list

[Jeff D. <da...@da...>]

On Aug 17, 2001, Reini Urban said:
> isn't this already fixed in 1.3.x ?
> At least I have a trivial fix for this in main.php.
> 
> // enforce anonical "case insensitive bRaCkeT links". 
> // For DOS flatfiles and mysql and "bad queries"
> if ( IsWikiPage($dbi, $pagename) and function_exists('WikiPageName') ) 
>    $pagename = WikiPageName($dbi, $pagename);
> 
> // Enable the output of most of the warning messages.
> // The warnings will screw up zip files and setpref though. 
> ...
> 

That's one way to fix the problem.  I don't really like it though.
(I don't think it's made it's way into the CVS yet either.)

I'd rather fix it so that pages with different capitalization are truly
different pages.  ('PageOne' and 'PagEone', I think should be different 
pages,
with different content.)  (I think we've had this discussion before:
 http://phpwiki.sf.net/phpwiki/index.php?case%20insensitive%20bRaCkeT%20
links
).

This has been fixed for the mysql backend by David Marsh --- notes about
what he did are on the SF bug report page:
 http://sf.net/tracker/index.php?func=detail&aid=233898&group_id=6121&at
id=106121
His fixes haven't made it into the CVS (either branch) yet though...
 ... and they do require a schema change.

Cheers,
Jeff

Re: [Phpwiki-talk] 1.2.1 task list

[Jeff D. <da...@da...>]

On Aug 17, 2001, Reini Urban said:
> I have a working version for 
> 
> Task "Reimplement Backlinks page"

This appears be in the 1.3.x (devel branch) task list, not the 1.2.x 
(stable)
task list.  (BTW Reini: is there a reason it's marked 'deleted' now?  
Should
I undelete it?)

I've been planning on doing this (though you're welcome to, Reini) in 
the
devel branch, but have been waiting until the database API stabilizes.

On the other hand, I would like to fix this (now) in the stable branch.
(I know there's a "feature freeze", but I consider this more of a bug 
fix,
than a new feature.)   If you only use WikiWords for page names
the current scheme works okay, but currently the "backlinks" search for
[names like this] will hit every page containing the words "names",
"like" and "this" anywhere within the entire page text.

Fixing this will require non-trivial changes to each of the database
backends, so I can also understand arguments that it should be left as
it is, for now.

If there are no objections, however, I will work on this over the next
few days.



Another issue which is probably worth dealing with is the case
insensitivity in page names bug.
 http://sf.net/tracker/index.php?func=detail&aid=233898&group_id=6121&at
id=106121

Fixing this properly, however, would require schema changes.  Is this
acceptable?

Jeff

[Phpwiki-talk] Dutch translations

[Jeff D. <da...@da...>]

On Aug 17, 2001, Steve Wainstead said:> 
> 
> Jeff, some months ago didn't we receive a patch for Dutch and you
> mentioned on the list that it should be applied to 1.2?
> 

Jan Nieuwenhuizen submitted a bunch of patches (on the development
branch) at the end of May.

I applied the patches to the devel branch, but did not back-port
them to the stable version.

At the time, we had this discussion:

On 31 May, 2001 Jan Nieuwenhuizen wrote:
>Jeff Dairiki <da...@da...> writes:
>> (I have not back-ported any of your translation fixes to the stable
>> branch of the CVS --- should they be?)
>
>Nah, not worth the effort imo.

Given that at present, a real beta release of the devel branch still
seems a ways away, maybe it is worth reconsidering...

Jan, are you interested in taking a look at it?  (If not I can try,
I suppose --- but I don't speak Dutch.)

Jeff

Re: [Phpwiki-talk] 1.2.1 task list

[Reini U. <ru...@x-...>]

I have a working version for 

Task "Reimplement Backlinks page"
https://sourceforge.net/pm/task.php?func=detailtask&project_task_id=25971&group_id=6121&group_project_id=7691

and some more, but my current problems are:

1) no time until early september.
2) interwiki links are falsely stored in the backlink table.
3) I have too much additional features in my search.php to extract this one.
   esp. I extracted the output code for generated lists of all kinds 
   (search, index, wanted, orphan, ...) into a seperate list.php to support
rows, limit, 
   offset, accelerators, date and sorted by.
   however the advanced search features (regex, and/or, ...) would be nice to
have. 
   I'll do that in september.

Steve Wainstead schrieb:
> I've started a new task list on SF for a 1.2.1 release. There are a
> handful of patches that need to be applied.
-- 
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/

[Phpwiki-talk] 1.2.1 task list

[Steve W. <sw...@pa...>]

I've started a new task list on SF for a 1.2.1 release. There are a
handful of patches that need to be applied.

Jeff, some months ago didn't we receive a patch for Dutch and you
mentioned on the list that it should be applied to 1.2?

~swain

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC

Re: [Phpwiki-talk] dumping

[Steve W. <sw...@pa...>]

There's a PhpWikiAdministration page... it has a menu for the various
admin functions. For some reason I can't access it on my home box right
now (using dbalib, I'm getting: "Warning: driver initialization failed in
/home/swain/public_html/1.2/phpwiki/lib/dbalib.php on line 39" using both
Apache/PHP setups I have! Curious.

~swain

On Fri, 17 Aug 2001, Didier Bretin wrote:

> Hello,
>
>    I would like to test performance of phpwiki with differents database.
> My current wiki is working with postgresql-7.0. I would like to dump the
> data and upload them in a wiki configured with flat file.
>
>    So I have try to access to the admin.php page and then I can only
> lock or remove a page.
>
>    Arno tells me that the dump is possible. Can you tell me how I do it ?
>
> Thanks.
> --
>              .------------------------------------------------.
>      .^.     | Didier Bretin, France | db...@in...    |
>      /V\     |-----------------------| www.informactis.com    |
>     // \\    |                       `------------------------|
>    /(   )\   | Visit: http://jeuxlibres.free.fr/              |
>     ^^-^^    `------------------------------------------------'
>
>
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
>

---
                  http://www.panix.com/~swain/
"Without music to decorate it, time is just a bunch of boring
production deadlines or dates by which bills must be paid."
                     -- Frank Zappa
http://pgp.document_type.org:11371/pks/lookup?op=get&search=0xF7323BAC