Package home | Report new bug | New search | Development Roadmap Status: Open | Feedback | All | Closed Since Version 1.32.0

Bug #16559 Unmigrated data vs UTF8 output
Submitted: 2009-09-01 13:22 UTC
From: doconnor Assigned: doconnor
Status: Closed Package: pearweb (version CVS)
PHP Version: 5.2.5 OS:
Roadmaps: (Not assigned)    
Subscription  


 [2009-09-01 13:22 UTC] doconnor (Daniel O'Connor)
Description: ------------ http://validator.w3.org/check?uri=http%3A%2F%2Ftest.pear.php.net%2Fbugs%2Fbug.php%3Fid%3D12916&charset=%28detect+automatically%29&doctype=Inline&group=0&user-agent=W3C_Validator%2F1.654 We need to: 1. SELECT * FROM pearweb 2. If !utf8; convert to UTF8 from ISO-8859-2 or whatever it is encoded as 3. REPLACE :D

Comments

 [2009-09-01 13:22 UTC] doconnor (Daniel O'Connor)
 [2009-09-01 14:00 UTC] doconnor (Daniel O'Connor)
(4:56:03 PM) cweiske: do we use iconv? (4:56:03 PM) cweiske: no (4:56:48 PM) cweiske: it's an apache vhost estting (5:23:05 PM) CloCkWeRX: alright; i think i have this straight in my head (5:23:21 PM) cweiske: CloCkWeRX, you do have access to sg1? (5:23:26 PM) CloCkWeRX: not right now (5:23:32 PM) cweiske: ohk (5:24:26 PM) CloCkWeRX: 1) pearweb spat out ISO whatever for quite a while, so bug reports form data was ISO when posted back, and stuck into an ISO database (5:24:26 PM) CloCkWeRX: 2) We swapped the table storage to UTF-8; using an ALTER TABLE CONVERT TO CHARACTER SET; and migrated all of the data then (5:25:01 PM) cweiske: iso-8850-15 (5:25:04 PM) cweiske: iso-8859-15 (5:26:14 PM) CloCkWeRX: 3) time went by and we kept INSERTing ISO data which was fine-ish (ie, pearweb still said it was ISO); even though mysql allocated enough space to store it as (mangled) UTF8 (5:26:27 PM) cweiske: 3 sound possible (5:26:52 PM) CloCkWeRX: 4) ; and now I've made pearweb say it's producing UTF-8, which is actually ISO in the database / internal handling; it gets rendered in its mangled UTF8 form to the browser; and everyone has noticed :( (5:27:26 PM) yunosh [~jan@p5486B29D.dip.t-dialin.net] entered the room. (5:27:32 PM) CloCkWeRX: can you check the charset of the bugdb table to see if #3 is likely? (5:27:44 PM) cweiske: not now, having a meeting now
 [2009-09-02 13:48 UTC] cweiske (Christian Weiske)
It could be we just need to "SET NAMES 'utf8'" and are done.
 [2009-09-04 14:07 UTC] cweiske (Christian Weiske)
 [2009-09-04 14:08 UTC] cweiske (Christian Weiske)
 [2009-09-04 19:51 UTC] doconnor (Daniel O'Connor)
[288045] does SET NAMES, whcih resolves the display of umlauts. However, locally, the format-bug-strip-only-some-html stuff is broken for Bug #12916
 [2009-09-08 11:46 UTC] cweiske (Christian Weiske)
-Status: Open +Status: Closed -Assigned To: +Assigned To: doconnor
Thank you for your bug report. This issue has been fixed in the latest released version of the package, which you can download at http://pear.php.net/get/
 [2009-09-08 11:47 UTC] cweiske (Christian Weiske)
Added patch test.
 [2009-09-08 11:47 UTC] cweiske (Christian Weiske)
Added patch test.