Hello,

Thanks for the report - however I don't think the tag DBA is appropriate for this report as we are not responsible for that part of the infrastructure.

I would suggest you take a look at this tags:

mediawiki-database (https://phabricator.wikimedia.org/project/view/165/)
mediawiki-api (https://phabricator.wikimedia.org/project/profile/200/)

Or if not completely sure, maybe mediawiki-general-or-unknown (https://phabricator.wikimedia.org/project/view/135/)

Let me know if I can help further
Cheers

Manuel.

Try to use Database::selectSQLText to build SQL

$dbr->selectSQLText(
	[ $qobj->alias => $qobj->joinTable ],
	[
		'id' => $qobj->alias.smw_id,
		't' => $qobj->alias.smw_title,
		'ns' => $qobj->alias.smw_namespace,
		'iw' => $qobj->alias.smw_iw,
		'so' => $qobj->alias.smw_subobject,
		'sortkey' => $qobj->alias.smw_sortkey,
	],
	$qobj->where !== '' ? [ $qobj->where ] : [],
	__METHOD__,
	array_merge( [ 'DISTINCT' ], $sqlOptions )
);

[ Untested, I am not sure what is in $qobj->from ]

Was changed in I855d0e0d6571ea15c03d2a27bf55ad5e14cd15d4 / 1bd86b3490fb8d9b6270dfa031097d78d1e7d185

Try to use Database::selectSQLText to build SQL

While I appreciated the effort, I don't think this is a solution.

[ Untested, I am not sure what is in $qobj->from ]

It contains concatenate (generated) INNER/LEFT JOIN statements hence the use of $db->query( ... ) later on.

The EXPLAIN will be part of an output that looks like http://sandbox.semantic-mediawiki.org/wiki/Issue/1171

The mentioned method was public in previous releases hence I'd expect [0] to apply.

[0] https://www.mediawiki.org/wiki/Deprecation#How_to_deprecate_a_method.2C_function.2C_etc.

We could make it public again?

Caused by rMW1bd86b3490fb8d9b6270dfa031097d78d1e7d185

Adding @aaron to task since he did the commit and not sure why we changed to protected from public.

C'mon, another breaking change disregarding the deprecation policy? If I where working for WMF I'd be embarrassed by the frequency of those. Also: yay for tons of protected methods, makes it clear how good the design is from miles away.

In T147550#2774338, @JeroenDeDauw wrote:

C'mon, another breaking change disregarding the deprecation policy?

What deprecation policy are you talking about?

If I where working for WMF I'd be embarrassed by the frequency of those. Also: yay for tons of protected methods, makes it clear how good the design is from miles away.

Fun fact: MediaWiki is a community supported project too. You've made plenty of contributions to core, I'm sure you could submit a patch to make a single function public again. And I have no idea what your point about protected methods is supposed to mean. Are you saying the design is good? Or is it sarcasm?

Tons of public methods is a bad design if anything, since it leads to a gigantically wide interface.

Tons of public methods is a bad design if anything, since it leads to a gigantically wide interface.

We are all (more or less) proficient developers and no one would dispute this but (as I pointed out above) removing its signature from one day to another, having an extension to fail, spending hours on investing the issue, is not what I would call the most professional approach.

Leaving a bug release [0] to fail on grounds of incompatibility is even less comprehensible.

My comment from T147683#2784789 applies here as well.

[0] https://github.com/SemanticMediaWiki/SemanticMediaWiki/issues/1751#issuecomment-258651052

In T147550#2784833, @mwjames wrote:

Tons of public methods is a bad design if anything, since it leads to a gigantically wide interface.

We are all (more or less) proficient developers and no one would dispute this but (as I pointed out above) removing its signature from one day to another, having an extension to fail, spending hours on investing the issue, is not what I would call the most professional approach.

Leaving a bug release [0] to fail on grounds of incompatibility is even less comprehensible.

My comment from T147683#2784789 applies here as well.

[0] https://github.com/SemanticMediaWiki/SemanticMediaWiki/issues/1751#issuecomment-258651052

Sure, my comment was in response to https://phabricator.wikimedia.org/T147550#2774338