Page 8 of 19

Re:Working on the OS FAQ

Posted: Sun Oct 24, 2004 4:12 am
by Candy
I don't see how executing a php server-side script could have any consequences of the browser you use, except if it can't process the browser info (in which case it's probably php's fault).

Re:Working on the OS FAQ

Posted: Sun Oct 24, 2004 4:23 am
by df
neither do i. I tested it and IE6 didnt give me any errors but firefox does.

odd that upgrading php from 4.0.6 to 4.3.9 gave the error....

obviously its something to do with the passed headers... I am waiting for 1.3.11 to come out of phpwiki and I am going to upgrade it.

Re:Working on the OS FAQ

Posted: Sun Oct 24, 2004 12:14 pm
by bubach
if you type in http://www.osdev.org/osfaq2/index.php instead of http://www.osdev.org/osfaq2/ it works..

[EDIT] wierd... it works if you first try one of the addresses and then the second right after..

Re:Working on the OS FAQ

Posted: Sun Oct 24, 2004 2:05 pm
by Candy
df wrote: neither do i. I tested it and IE6 didnt give me any errors but firefox does.
Actually, IE6 on win2k does give the error, but since microsoft decided to switch on strict parsing all of a sudden (no problem with that, but try to follow one course...), it crashes out. The error (in dutch, can't help the language here, all other is english):
Microsoft wrote: De XML-pagina kan niet worden weergegeven
Kan XML-invoer niet lezen met opmaakmodel XSL. Herstel de fout en klik vervolgens op de knop Vernieuwen of probeer het later opnieuw.


--------------------------------------------------------------------------------

DOCTYPE-declaratie mag niet buiten de beginsectie voorkomen. Fout bij het verwerken van bron http://www.osdev.org/osfaq2/. Regel 4, positie 11

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
----------^
In other words, it encountered a doctype after the start, which is invalid. It trips.

It seems there's a url rewriter rewriting it to ./index.php instead of ./, after which it outputs some header. Then, the gz handler can't take the header anymore, because it's been output, and then the page is output after the error. Switching off the display of errors will fix it in a way, although the first page will consume slightly more space.

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 5:08 am
by Pype.Clicker
what did happen to the FAQ?

It looks that it reverted to the OSX style and that plenty of pages are missing (BareBones, BeginnerMistake, ...)

uh ? that's even weirder: the pages are actually there, but the WIKI beleive they're not there and suggest we create them ... and trying to edit a page that was claimed missing edits a blank page (??)

Could that be that we only have the HTML cache and no content any longer ??

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 5:32 am
by Solar
df: I could e-mail a content backup as of 2004-10-19 immediately and probably a more up-to-date one in a couple of hours when I get home. (Though I'm not sure my latest backup isn't already affected by this "page loss". Really strange...)

Edit: If your HTML proxy / cache confuses you, http://www.osdev.org/osfaq2/index.php/AllPages gives you a list of available pages, and indeed lots and lots of 'em are missing.

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 6:00 am
by Pype.Clicker
no, actually, i can get the "pre-html cached" content from the Wiki ... so it seems even for pages we lost, we still have the "rendered" content (try out http://www.osdev.org/osfaq2/index.php/PypeClicker, for instance)

The current wikidump seems only to keep the ancient content (the pages that were in the static version)

i don't have backup more recent than july 19th, unfortunately...

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 6:27 am
by Solar
Ah... I'd say that's damage to one of the PhpWiki MySQL tables, then: The page itself is there (when requested by name directly), but parts of the PhpWiki logic think it isn't (when rendering Wiki links or querying for AllPages).

I'm afraid that means df has to do some SQL patching... (recreating the damaged table from data in other tables). Can't say I envy him. ;)

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 8:11 am
by df
the wiki table 'recent' corrupted itself...

all the pages are there in the page table but whatever was in the 'recent' table is gone.

so the wiki looks kinda funny right now... there is lots of links on the front page that say they are not there but if you click on them, the pages are there (i cleared the html cache as well). somehow I need to figure out how to get the wiki to realise the pages are there and to make the correct links...

the wiki is annoying me bigtime.... :/ mysql is a piece of **** but its all I have. access never corrupted itself as much as my experience with mysql has. (and Ive done some nasty things to access databases.).

aaaanyway. i dont know what I am going to do.

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 9:28 am
by Solar
Harumph. Dumping the Wiki doesn't help, the dump doesn't include the "vanished" pages...

I'll see what I can do with my latest backup at home and a bit of creative scripting. I won't cater for pages' history, but expect me to mail you a snapshot of current contents to re-start the Wiki with.

And next thing I do is making the backups automatic and daily. He who laughs last, has made a backup. ;)

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 10:02 am
by df
I admint I got lazy, I was doing a daily mysqldump on the wiki then things stabalized and I had no problems :( so I didnt dump for a couple of weeks...

Re:Working on the OS FAQ

Posted: Tue Nov 09, 2004 11:06 pm
by Solar

Re:Working on the OS FAQ

Posted: Wed Nov 10, 2004 2:38 am
by df
cheers. i did the sql and it genereated the same table as I had from the original mysql repair from last night.

anyway, the todo page lists all the pages that became orphaned. I think I will just copy the markup, delete page, recreate page. paste markup back...

time consuming but... i will see if we get any more responses on that thread... otherwise tomorow I will do the cut/paste/delete/create on all the orphaned pages.

Re:Working on the OS FAQ

Posted: Wed Nov 10, 2004 3:12 am
by Solar
You can probably get away with just calling the page (shouldn't that be "WantedPages" instead of "OrphanedPages"?) with ...?action=edit, and saving it.

Note that the man over at PhpWiki asked for contacting him by e-mail. I just told him that I wasn't the man with the MySQL access, and forwarded the stuff, so you might want to get in touch with him.

Re:Working on the OS FAQ

Posted: Wed Nov 10, 2004 11:47 am
by df
all fixed now. I posted how I fixed it on the sourceforge forum thread...

eash! from no on, I will do daily backups for a while then go down to once a week. (unfortunatly I dont have cron access and wouldnt want to put my password into a cron job either;))