ahinea

DB Error: mismatch (in a PHP script)

Ivan — Mon, 27 Aug 2007 22:02:53 +0000

If you get something like the “DB Error: mismatch” message in a PHP script, this is probably caused by a bug in the script. If I get it right, the message means: there is a mismatch between the number of placeholders (e.g. “?” signs) in your database query and the number of parameters given to the query() method.

For instance, if your query is “SELECT * FROM people WHERE honesty>3″, but you mistakenly call it like this:

$dbh-> query( "SELECT * FROM people WHERE honesty>3", $threshold );

Where you probably meant:

$dbh-> query( "SELECT * FROM people WHERE honesty>?", $threshold );

I think this is reported by the DB_Common class (since a certain version).

On the topic:

Subversion / Apache griefs: can’t commit, MERGE failed

Ivan — Sat, 03 Mar 2007 12:05:54 +0000

Installed Subversion. Configured Apache to run as a Subversion server. Seems to work. I even tried some WebDAV applications (davfs) on it. Work. Not perfectly (some strange errors), but it does work.

Until I tried to checkout, modify and commit with SVN itself:

svn: Commit failed (details follow):
svn: MERGE request failed on '/'
svn: Can't open directory '/home/ivan/bsite/svn-repo/db/transactions/0-1.txn/transactions': No such file or directory

No matter what I tried SVN can’t commit into a repository without an error.

Turns out this problem is caused by Apache and Subversion being compiled against different versions of the APR libraries. I got it fixed by rebuilding apache and subversion from their fresh sources.

Thank you, Christopher Ness! Your post helped me a lot.

Low-profile AGP video card with 3D acceleration

Ivan — Tue, 23 Jan 2007 06:54:32 +0000

I was looking for a good low-profile AGP video card. nVidia GeForce- or ATI Radeon-based. With DVI output. Preferably silent (with no cooler). Here is what I’ve found so far.

nVidia GeForce:

Gigabyte GV-N62128DS or GV-N55128DS
Asus N6200/TD/128M
XFX Geforce 6200 AGP: PVT44APANG, PV-T44A-WANG.
MSI Video 8998010 GeForce FX5200 or MSI GeForce 6200 (NX6200AX-TD128LF).

ATI:

SAPPHIRE 100576L Radeon 9550SE 128MB DDR AGP 4X/8X Low Profile
PowerColor RADEON 9200SE AGP.
Gigabyte GV-R925128DE.
Asus A9250/TD/128M.

And I got myself Asus N6200/TD/128.

The problem with nVidia cards is that their Linux driver is closed source. That wouldn’t be a problem at all if the driver in it’s current state didn’t break one important thing in my PC usage: hibernate function (swsusp, to be precise). With an old Radeon I could send my PC to sleep and be 100% sure that it would wake back up. Not anymore with GeForge 6200 and the latest nVidia driver.

Proper XPath/XSLT error reporting in XML::LibXSLT. The patch

Ivan — Fri, 14 Jul 2006 03:44:19 +0000

UPDATE: This post and this patch are no longer needed. A new version of XML::LibXSLT fixes the error reporting and some other issues. Below is the original post for purely historical reasons.

XML::LibXSLT is a perl module, written and maintaned by Matt Sergeant. The module provides access to the fine Gnome project’s XSLT library libxslt. XML::LibXSLT versions up to and including 1.58 have one serious deficiency though: they provide rudimentary reporting of XPath and XSLT processing errors.

This patch tries to fix this deficiency and makes XML::LibXSLT die in case of error with a full error message, often multiple lines. The generated message will be similar to the one xsltproc tool would produce.

Get the patch

Download: xml-libxslterror-reporting-patch.txt

A summary of the patch changes follows.

I figured some important errors, like invalid XPath expression errors were generated in libxml2 library and handled by XML::LibXML module. XML::LibXML module installs its own error handler for libxml2 errors and this handler croak()s immediatly in case of error. But that handling isn’t appropriate for LibXSLT, because we can’t really trap it. Or we can trap it, but in that case it is the libxslt library that should trap it and continue it’s processing (to construct a full error message), which it surely wouldn’t do.

So I suggest to temporarily set global xmlGenericErrorFunc variable to our own handler (but then reset it back after each transform or parsestyle operation).

Also in this patch I put LibXSLT_error_handler() function to use. The function was there already before me, but wasn’t used for some reason.

And, finally, I slightly change some wording in the error message phrases of XML::LibXSLT itself. (Remove that depressing Error applying stylesheet: (get error out of libxslt) message…)

* * *

Work on this patch was sponsored by the BuildSite company.