phpwiki-talk

Re: [Phpwiki-talk] 1.3.x series

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

On Fri, 2 Feb 2001, Aredridel wrote:

> > I think there are some obvious things to drop, like:
> > 
> > references (that is, [1] [2] links)
> 
> By all means. Let's drop it. Perhaps put footnotes code in instead of
> references (ala NBTSCWiki)?

I like the footnotes feature; it makes sense vs. the references. Let's add
it.

> > Third, I still long for a better search interface than what we have now
> > (which doesn't work too bad, but could be better).
> 
> ooh, yeah... what can we work up, here? I'd like to use regexps, but I'm a
> regexp nut. There's also SQL-like wildcards (% and _) and globs (*, ?, {x,y}
> and [a-q]) that'd be nice as syntax, as well as altavista like boolean ops--
> +foo -bar or standard boolean (and/or/xor/not)

I have to revisit Neilson's "Designing Web Usability." He has a whole
chapter on search interfaces and it has a lot of good practical ideas. In
general, there should be a simple search and an advanced search, so we can
go crazy with the advanced features.

> > Any navigational aids we can contrive would help, like SiteMap, or
> > HotLinks or whatever.
> 
> Mmm, and I'd love to see 'em as modules, enablable by adding the block to a
> template or uncommenting something or setting a config variable. I still
> don't like templates that much (which is why I like PHP so much), but it
> makes a lot of sense in many ways.

Excellent idea... I think this is now the case anyways with Arno's rewrite
of transform.php.

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] "dangling references"

[<ph...@de...>]

On Tue, 06 Feb 2001 09:15:02 -0800, you wrote:
=> 	How about a patch/instructions for those of us with 1.2
=> who would like to modify the current version for this kind of
=> table name separation? 

	I'm sorry, I really was reading the code, but I'm just
blind or stupid.  Obviously the answer is already right there in
the table name variable assignments in config.php (right above
where I filled in my MySQL stuff).

	I thought I had looked everywhere (after I did the prelim
stuff in config). I feel like such a dunce.

	Nonetheless, *thanks* for phpWiki to all involved, from
the guy now sitting facing the corner, with the pointed hat on
and being *very* quiet for awhile.

	<sigh>

	- Don

Re: [Phpwiki-talk] "dangling references"

[<ph...@de...>]

On Tue, 6 Feb 2001 10:29:35 -0500 (EST), you wrote:
=> OK. I've been meaning to mention we can again make whatever radical schema
=> changes we want... we should start with renaming the tables so they don't
=> cause conflicts (wiki_hotlinks, wiki_archive etc).

	Great idea!

	[Just starting up with phpWiki and I *love* it - thanks
to everyone involved in it's birth and development.]

	How about a patch/instructions for those of us with 1.2
who would like to modify the current version for this kind of
table name separation?  Assign a $wiki_prefix variable in config
and append it to all table use?  Sure would be appreciated here.

	Thanks again to everyone. Wow!

	- Don

Re: [Phpwiki-talk] phpwiki - The page cannot be displayed

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

Hi frusti,

> I have downloaded the phpwiki and installed.
> After this i have started the index.php and phpwiki has insert some tables
> and so on. Thats allway OK!
> 
> Now then i click on anyone Link i become an error: "The page cannot be
> displayed", but the URL is ok.

This is an error message from IE, isn't it? It sounds like you are getting
a server error. Can you check the log for your web server and see what the
error message is?

Also, can you tell us what operating system, what web server and what
version of PHP you are using?

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

[Phpwiki-talk] phpwiki - The page cannot be displayed

[<fr...@gm...>]

Hello!

I have downloaded the phpwiki and installed.
After this i have started the index.php and phpwiki has insert some tables
and so on. Thats allway OK!

Now then i click on anyone Link i become an error: "The page cannot be
displayed", but the URL is ok.

Can someone give me a info? To information, i don't use the standard-port
80, i use the port 1028, but in the config.php its set from the SERVER_PORT
variable and that i have also checked, it is OK!

thanks, 
frusti

-- 
Sent through GMX FreeMail - http://www.gmx.net

Re: [Phpwiki-talk] wish for 1.3 version

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

Oh no! Not tables again! We've gotten two patches already! :-)

Well, it would be an interesting challenge for Arno ;-)

~swain

On Tue, 6 Feb 2001, Didier Bretin wrote:

> Hello,
> 
>   Do you plan to code the use of table ?
> 
> Regards.
> -- 
>             .------------------------------------------------.
>     .^.     | Didier Bretin, France | db...@in...    |
>     /V\     |-----------------------| www.informactis.com    |
>    // \\    |                       `------------------------|
>   /(   )\   | Visit: http://www.multimania.com/cieexcalibur/ |
>    ^^-^^    `------------------------------------------------'
> 
> 
> _______________________________________________
> Phpwiki-talk mailing list
> Php...@li...
> http://lists.sourceforge.net/lists/listinfo/phpwiki-talk
> 

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] Migrating to 1.2

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

On 6 Feb 2001, Barnett, Paul G wrote:

> >Did you try http://yourhost.yourdomain/phpwiki/admin.php  ?
> 
> 
> This gives me acces to lock, unlock and remove, but
> what is the new way to run "Dump serialized copies of pages",
> "Dump the pages as HTML","Load serialized pages from files",
> "Rebuild the dbm files so they are smaller" etc ?

And dump zip copies... Arno pulled the admin stuff out a while back and
moved it into the Wiki itself; in turn,

* "dump serialized copies" was apparently replaced with dump as zip files.
I recall Jeff feeling very paranoid about letting people write to the file
system, and it would be safer to let them get a zip file from the browser.

* "dump as HTML" was never implemented, my fault.

* "Load from files" was redone to load from zip files, as noted above.

* "Rebuild the DBM" was replaced with a Perl script, included in the
distribution.

~swain


...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] "dangling references"

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

On Tue, 6 Feb 2001, Arno Hollosi wrote:

> > rename links
> 
> I assume you mean rename pages? Yes, that would be made easier.

Well, in the current implementation we could rename a page (update wiki
set pagename='foo' where pagename='bar') but all the links to 'bar'
remain, now undefined. Storing the links in a table and not in the pages,
so to speak, would solve this problem.

> Also, I suggest that we split the wiki table into two: one table as it is 
> now, but instead of pagename with a pageid. Another table which does 
> pagename->pageid. That way we could have aliases of pagenames, e.g. (yet 
> another simplistic example) "WikiName" and "WikiNames" could both point to 
> the same pageid.

OK. I've been meaning to mention we can again make whatever radical schema
changes we want... we should start with renaming the tables so they don't
cause conflicts (wiki_hotlinks, wiki_archive etc).

> > delete links
> > When a page is deleted we can then delete all references to it as well.
> 
> Why would I like to do this? Even if I delete pages I'd rather have a 
> dangling link like "Which explains why StevenWainsteadEatsWorms?" then 
> something like "Which explains why !PageRemoved!"

Then the page might be redefined... let's say I'm tired about all the
emails I get from all over the world asking me about my interesting diet.
Just deleting the page leaves "dangling references" all over, so the
problem never goes away. But the issue of what do we put in place of the
link is a problem.

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] Modularization of transform.php

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

On Tue, 6 Feb 2001, Arno Hollosi wrote:

> 
> > > - redo %%ADMIN-%% tokens -- see my previous email "system page" (Jan 25)
> 
> Is that change ok? (I've not yet written it). Short: mixture mode which 
> allows html mixed with template placeholders and wiki markup.

I haven't looked into this yet but I still have your email. I'll check it
later...

> Right now we don't have nested markup. A simple (senseless) example is
> bold&italic markup inside bracketlinks, e.g. [__bold text for__|BoldPage]
> I guess nested markup is only interesting if phpwiki allows
> multi-line markup. Then we have one markup (multi-line) which may contain 
> other markup (bold, lists, links, ...) So my suggestion was: the multiline 
> markup should then just register another transformer object and do 
> do_transform() with its contained wiki data.

That sounds fine... I wonder, what if we stopped doing line by line
processing? Wouldn't we gain something from that? Or would it be too hard?
For example, tokenize the entire page and process it (via recursive
descent parsing or something similar).

~swain


...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

[Phpwiki-talk] wish for 1.3 version

[Didier B. <db...@in...>]

Hello,

  Do you plan to code the use of table ?

Regards.
-- 
            .------------------------------------------------.
    .^.     | Didier Bretin, France | db...@in...    |
    /V\     |-----------------------| www.informactis.com    |
   // \\    |                       `------------------------|
  /(   )\   | Visit: http://www.multimania.com/cieexcalibur/ |
   ^^-^^    `------------------------------------------------'

Re: [Phpwiki-talk] Migrating to 1.2

[Arno H. <aho...@xm...>]

> This gives me acces to lock, unlock and remove, but
> what is the new way to run "Dump serialized copies of pages",
> "Dump the pages as HTML","Load serialized pages from files",
> "Rebuild the dbm files so they are smaller" etc ?

There's a page in pgsrc/ called PhpWikiAdministration

Just put that into your wiki and view it through admin.php.
Shrinking DBM files has not been done in that interface yet.
Send us a patch :o)

/Arno

Re: [Phpwiki-talk] dawn! ISP php safe mode

[Arno H. <aho...@xm...>]

> Hardcoded translations, putting inside the php a special mark
> and them with a program that preprocess all this marks and translating
> all before uploading. After that no need to call funcs neither gettext
> or what others have: a SQL text table, tons of variables...

Hm - the gnu textutils save us a lot of headaches. It's easy to keep the=20
files up to date, ....

If you have problems with php's safemode then I suggest the following:
replace all occurneces of gettext() with mygettext() and modify the gette=
xt=20
check in config.php, so that your mygettext() is used.

I've attached a perl-script for doing replaces on a file list.
Quite handy tool -- you have to know perl regexps -- just be careful and=20
don't forget to escape regexp chars like '('. I suggest making a backup o=
f=20
the files first.

/Arno

=00

Re: [Phpwiki-talk] Migrating to 1.2

[Barnett, P. G <Pau...@Br...>]

Hi Steve - thanks for the reply.


>You can use dbmlib.php if you've compiled support for the dbm* functions
>into PHP. In fact, dbalib.php is just a copy of dbmlib.php with the
>function names and signatures changed. I tried this for all the functions,
>but so far all I get are blank lines for the three link sets at the bottom
>of the pages.
>


It seems that the guy who installed php4 on our server noticed that phpwiki
needed the --with-db option and did it for me. I also got the blank lines, 
but they started to get filled-in once we began to update the pages in 1.2


>> And is there still an admin functions index.php somewhere? - I cannot
>> find it.
>
>Did you try http://yourhost.yourdomain/phpwiki/admin.php  ?


This gives me acces to lock, unlock and remove, but
what is the new way to run "Dump serialized copies of pages",
"Dump the pages as HTML","Load serialized pages from files",
"Rebuild the dbm files so they are smaller" etc ?




-------------------------------------------------------------------------------------------------
21st century air travel     http://www.britishairways.com

Re: [Phpwiki-talk] "dangling references"

[Arno H. <aho...@xm...>]

> We've talked before about storing links in a separate table, which woul=
d
> make it easy to do a few things:
>
> list undefined links

That's easy (and already possible) -- I've already included it in gowiki,=
=20
see http://senseis.xmp.net/?UndefinedPages

> rename links

I assume you mean rename pages? Yes, that would be made easier.

Also, I suggest that we split the wiki table into two: one table as it is=
=20
now, but instead of pagename with a pageid. Another table which does=20
pagename->pageid. That way we could have aliases of pagenames, e.g. (yet=20
another simplistic example) "WikiName" and "WikiNames" could both point t=
o=20
the same pageid.

I have to look at Jeff's db schema again - there was some useful stuff on=
=20
how he used a common structure for wiki and archive.


> delete links
> When a page is deleted we can then delete all references to it as well.

Why would I like to do this? Even if I delete pages I'd rather have a=20
dangling link like "Which explains why StevenWainsteadEatsWorms?" then=20
something like "Which explains why !PageRemoved!"


> [...] These things are fairly easy, and we have most of the code.

Right.

/Arno

Re: [Phpwiki-talk] Modularization of transform.php

[Arno H. <aho...@xm...>]

> > - redo %%ADMIN-%% tokens -- see my previous email "system page" (Jan =
25)

Is that change ok? (I've not yet written it). Short: mixture mode which=20
allows html mixed with template placeholders and wiki markup.


> > Note that recursive calls (nesting markup) are not possible.
> Hmm. I don't quite understand. Can you give an example of nesting marku=
p?
> The TestPage looks fine, so we must not have an example there?

Right now we don't have nested markup. A simple (senseless) example is
bold&italic markup inside bracketlinks, e.g. [__bold text for__|BoldPage]
I guess nested markup is only interesting if phpwiki allows
multi-line markup. Then we have one markup (multi-line) which may contain=
=20
other markup (bold, lists, links, ...) So my suggestion was: the multilin=
e=20
markup should then just register another transformer object and do=20
do_transform() with its contained wiki data.

/Arno

[Phpwiki-talk] "dangling references"

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

I've been reading parts of "Multimedia and Hypertext: the Internet and
Beyond," by Jakob Neilson
(<URL:http://www.amazon.com/exec/obidos/ASIN/0125184085/qid=981432634/sr=1-6/ref=sc_b_6/107-7065042-5444529>.
There's an interesting chapter on hypertext authoring (like writing Wiki
pages) and how little research has been done on what makes authoring
hypertext easier. (We just use a TEXTAREA like anyone else).

We've talked before about storing links in a separate table, which would
make it easy to do a few things:

list undefined links
rename links
delete links

Neilson calls undefined links "dangling references" (to "dangle" means to
hang from something, like an apple dangling from a tree branch).
Apparently in the hypertext systems the book surveyed, and projects he was
involved with, listing dangling refs was very useful to hypertext authors.

The challenge here is we have to preprocess Wiki pages to parse out the
links, update the table of links, and store the page. When the page is
rendered, all links are matched from the page to the table. These things
are fairly easy, and we have most of the code. It's a matter of
reorganization.

When a page is deleted we can then delete all references to it as well.

Any thoughts?

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] Modularization of transform.php

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

On Sun, 4 Feb 2001, Arno Hollosi wrote:

> I had an afternoon at my hands, so I redid transform.php.
> 
> Transform functions are now modularized and registered to a central
> WikiTransform object.
> 
> I think the current design is straight forward, easy to comprehend, and 
> thus easy to extend.
> 
> Things to be done:
> - re-add html mode "|"
> - provide functionality for multi-line markup (i.e. markup spanning several
>   lines)
> - redo %%ADMIN-%% tokens -- see my previous email "system page" (Jan 25)
> 
> I'd like to hear some feedback -- is the current form ok?

I spent some time reading this, and all I can say is wow! Great work! It
was immediately understandable, incredibly clearly written, and very
clean.

> Note that recursive calls (nesting markup) are not possible. It would 
> unnecessarily complicate the issue. Instead, I suggest that if you'd like 
> to do markup-transformation inside another markup you just register another 
> WikiTransform object and call that one with your markup data.

Hmm. I don't quite understand. Can you give an example of nesting markup?
The TestPage looks fine, so we must not have an example there?

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

[Phpwiki-talk] dawn! ISP php safe mode

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

Hi,

I have a problem with two of my ISP's, they don't allow me to change
to LANG setting and caused by this the traslation system don't work.

You builded phpwiki for working with Po files and so.

My point is another:

Hardcoded translations, putting inside the php a special mark
and them with a program that preprocess all this marks and translating
all before uploading. After that no need to call funcs neither gettext
or what others have: a SQL text table, tons of variables...

Here is an approach by Carlos Viana that I like:

http://groups.yahoo.com/group/phpnuke/files/Addons/utt_110.zip

I know this should be changed a bit, but Carlos is very active and
accepts all suggests, so one of my first would be that the special
chars can be %%%text%%%

What do you think?

Pablo Roca (pr...@cl...)
Clavo Congelados, S.A. (http://www.clavo.net)
La Coruna - Espana 

[Phpwiki-talk] Modularization of transform.php

[Arno H. <aho...@xm...>]

I had an afternoon at my hands, so I redid transform.php.

Transform functions are now modularized and registered to a central
WikiTransform object.

I think the current design is straight forward, easy to comprehend, and=20
thus easy to extend.

Things to be done:
- re-add html mode "|"
- provide functionality for multi-line markup (i.e. markup spanning sever=
al
  lines)
- redo %%ADMIN-%% tokens -- see my previous email "system page" (Jan 25)

I'd like to hear some feedback -- is the current form ok?

Note that recursive calls (nesting markup) are not possible. It would=20
unnecessarily complicate the issue. Instead, I suggest that if you'd like=
=20
to do markup-transformation inside another markup you just register anoth=
er=20
WikiTransform object and call that one with your markup data.

/Arno

Re: [Phpwiki-talk] IsWIkiPage() is expensive

[Ori F. <or...@co...>]

2/3/2001 21:17:13, Arno Hollosi <aho...@xm...> wrote:

>So, whenever the page got displayed IsWikiPage() got called 300 times, 
>which translates to 300 SQL queries for one page. The web server almost 
>came to a halt due to mysql eating up all its resources.

For my own wiki (rashreshet.org, it's in Hebrew) IsWikiPage() is caching results from previous calls, and can also check for multiple pages with one call. 
So I call it once with all the linked pages in the displayed page, which uses one SQL query, and it's done.

[Phpwiki-talk] IsWIkiPage() is expensive

[Arno H. <aho...@xm...>]

Hi all,

I nearly killed my wiki by adding a page which ranks pages by=20
click-distance (see: http://senseis.xmp.net/?PagesByDistance). In my firs=
t=20
try, the page contained all other (300) pages with real wikilinks.

So, whenever the page got displayed IsWikiPage() got called 300 times,=20
which translates to 300 SQL queries for one page. The web server almost=20
came to a halt due to mysql eating up all its resources.

A painful lesson.
We need to improve this behaviour during 1.3.x

/Arno

Re: [Phpwiki-talk] 1.3.x series

[Arno H. <aho...@xm...>]

> I think there are some obvious things to drop, like:
> references (that is, [1] [2] links)
> support for 3.x PHP
> dbmlib.php

Some people will mourn the loss of PHP3, but I guess that's the open sour=
ce=20
way: always use the latest toys :o)

> I'd like to see some cruft cleaned out as well. Perhaps a rewrite of
> transform.php (we've talked about modularizing it for a long time), and
> drop all the old tab markup.

I'd like to give this one a shot. I guess I can reuse some code from Jeff=
=2E

> Second, a more permanent archive. Pages should live longer than one
> version. This can be configured to zero, ten, one hundred previous page=
s.

I have modified phpwiki slightly for Sensei's Library (senseis.xmp.net) t=
o=20
allow unlimited versions. It is always storing the whole page. As of toda=
y,=20
there are 300 pages in the wiki, and close to 600 pages in the archive,=20
created by roughly 2000 edits (sum of version numbers).

Btw, I discovered a bug in the diff code -- I don't know how to fix thoug=
h.
If I find the time I'll make an bug entry.

> Any navigational aids we can contrive would help, like SiteMap, or
> HotLinks or whatever.

I've attached my local sitemap code. Just a proof of concept - runs only=20
with mySQL!! Explanation: red line: page in center of map. Middle column=20
above: pages referencing center page. Left/right: pages referencing/being=
=20
referenced by the page in the middle column. Middle column below: pages=20
referenced by center page. Left/right: pages referencing/being referenced=
=20
by the page in the middle column. Not really easy to read - so maybe the=20
layout should be redone and the number of pages should be limited.

p.s. to run it you need to set the variable $related - modify index.php t=
o=20
look for that variable and include navigate.php. Then call like=20
http://site.url/phpwiki/index.php?related=3DAddingPages

/Arno
=00

Re: [Phpwiki-talk] 1.3.x series

[Aredridel <are...@nb...>]

> 
> As 1.2 is shaken down, and I don't expect a lot of bugs at this point,
> it's not too soon to think about the 1.3.x series.
> 
> I think there are some obvious things to drop, like:
> 
> references (that is, [1] [2] links)
> support for 3.x PHP
> dbmlib.php

By all means. Let's drop it. Perhaps put footnotes code in instead of
references (ala NBTSCWiki)?

> I'd like to see some cruft cleaned out as well. Perhaps a rewrite of
> transform.php (we've talked about modularizing it for a long time), and
> drop all the old tab markup.

Mmm, yes. I've been thinking of a way to add NBTSCWiki-style markup (which I
prefer, since it's almost as readable in text as rendered) to PHPWiki
without disrupting things, but I'm not able to think of a way yet.

> Perhaps the most requested feature, which I would like to see, is
> authentication. I would like a model that follows the mailing list: the
> user must have a valid email address. When they register, a link is
> emailed to them that they must follow to validate the account, or perhaps
> we can just mail out the password. Their account name will already be used
> in RecentChanges when they edit a page. This can be configured so it
> doesn't work at all (like PhpWiki is now).

Authentication can be made modular rather easily. I've already done a
passable implementation using HTTP-auth on the NBTSC Wiki. I've -almost- got
a web-form-and-cookie based implementation sitting around too, that could be
extended with a little scripting to do mailout-based account creation.

> Second, a more permanent archive. Pages should live longer than one
> version. This can be configured to zero, ten, one hundred previous pages.

Oh, -yes!-.

I've been pondering using a CVS-like branching system for storing page
changes -- totally eliminating need for page locking for edits, as well as,
if the implementation had a simple interface, would allow a Wiki to have
clean pages, without censoring people who wish to be offensive -- just
branch, and make the clean branch the "main" one.

Really, I want to be able to see diffs of fifteen versions, since every now
and again, I travel and can't keep up on Wiki.

> Third, I still long for a better search interface than what we have now
> (which doesn't work too bad, but could be better).

ooh, yeah... what can we work up, here? I'd like to use regexps, but I'm a
regexp nut. There's also SQL-like wildcards (% and _) and globs (*, ?, {x,y}
and [a-q]) that'd be nice as syntax, as well as altavista like boolean ops--
+foo -bar or standard boolean (and/or/xor/not)

> Any navigational aids we can contrive would help, like SiteMap, or
> HotLinks or whatever.

Mmm, and I'd love to see 'em as modules, enablable by adding the block to a
template or uncommenting something or setting a config variable. I still
don't like templates that much (which is why I like PHP so much), but it
makes a lot of sense in many ways.

> Just some food for thought...

Mmm! Brain food! Need fish crackers...

Ari

[Phpwiki-talk] 1.3.x series

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

As 1.2 is shaken down, and I don't expect a lot of bugs at this point,
it's not too soon to think about the 1.3.x series.

I think there are some obvious things to drop, like:

references (that is, [1] [2] links)
support for 3.x PHP
dbmlib.php

I'd like to see some cruft cleaned out as well. Perhaps a rewrite of
transform.php (we've talked about modularizing it for a long time), and
drop all the old tab markup.

Perhaps the most requested feature, which I would like to see, is
authentication. I would like a model that follows the mailing list: the
user must have a valid email address. When they register, a link is
emailed to them that they must follow to validate the account, or perhaps
we can just mail out the password. Their account name will already be used
in RecentChanges when they edit a page. This can be configured so it
doesn't work at all (like PhpWiki is now).

Second, a more permanent archive. Pages should live longer than one
version. This can be configured to zero, ten, one hundred previous pages.

Third, I still long for a better search interface than what we have now
(which doesn't work too bad, but could be better).

Any navigational aids we can contrive would help, like SiteMap, or
HotLinks or whatever.

Just some food for thought...
~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain 

Re: [Phpwiki-talk] Migrating to 1.2

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

Hi Paul,

On 2 Feb 2001, Barnett, Paul G wrote:

> In config.php it says that for php4, I must use 'dba' (or default , which does
> the same) where I was using 'dbm' for php3.
> 
> However, dbalib.php does not have the new functions RemovePage,
> GetWikiPageLinks or SetWikiPageLinks, which means that some of the new
> functionality is not available.(5 best...., remove page, etc)  They are
> in dbmlib.php, and if I force use of dbmlib in php4, it seems to work. 

You can use dbmlib.php if you've compiled support for the dbm* functions
into PHP. In fact, dbalib.php is just a copy of dbmlib.php with the
function names and signatures changed. I tried this for all the functions,
but so far all I get are blank lines for the three link sets at the bottom
of the pages.


> Must I absolutely use dba in php4, and therefore loose the new functions?


No, just compile PHP to support the old dbm library.

> Am I risking anythink by forcing 'dbm' ?

Again, no. Both dbmlib.php and dbalib.php use the same set of DBM files.
You can stay with dbmlib.php.

> 
> What is the recommended way to migrate the database - If I just keep the
> dbm files, I dont see the recent changes to things like ReleaseNotes - I
> could just cut and paste from a pristine 1.2 but this seems crude - Is
> there a better way? 

Err. Well, that's one problem we've never really solved. The default pages
don't change a whole lot, so not having ReleaseNotes is not a problem. I
think the user's pages count for more than what we provide by default.

> And is there still an admin functions index.php somewhere? - I cannot
> find it. 

Did you try http://yourhost.yourdomain/phpwiki/admin.php  ?

~swain

...............................ooo0000ooo.................................
   Hear FM quality freeform radio through the Internet: http://wcsb.org/
                     home page: www.wcsb.org/~swain