<div>Ok. Yes, changing the existing behaviour could lead to back compatibility issues.</div><div> </div><div>But have you found the functionality to "make an option unrecognizable again" useful? In theory, would it be appropriate to have eg. "UNSET" to perform that?</div><div><br /></div><div><br /></div><div>----------------</div>
<div>Кому: KES (kes-kes(at)yandex(dot)ru);<br /></div>
<div>Копия: pgsql-docs(at)lists(dot)postgresql(dot)org;<br /></div>
<div>Тема: Unexpected behaviour: it was documented to return the same value;<br /></div>
<div>13.07.2025, 18:17, "David G. Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com>:<br /></div>
<blockquote>On Sunday, July 13, 2025, KES <<a href="mailto:kes-kes(at)yandex(dot)ru">kes-kes(at)yandex(dot)ru</a>> wrote:<br /><blockquote class="f13ca48719c8a60033905b23b39675agmail_quote" style="border-left-color:#ccc;border-left-style:solid;border-left-width:1px;margin:0 0 0 0.8ex;padding-left:1ex"><div>I suppose that DEFAULT syntax should be used to set setting to its default value. Eg.</div><div><div>set_config( 'my.option', DEFAULT, false )</div><div>This is clear what should happen here. Also this syntax will correspond to other places, eg. INSERT INTO xx (name) VALUES (DEFAULT).</div><div> </div><div>Using NULL to set reset the value to the default is very very confusing.</div><div><div>set_config( 'my.option', NULL, false )</div><div>Expected: this call should make option unrecognizable again.</div></div></div></blockquote><div><br /></div><div>I don’t see us changing this even if someone were willing to write a patch. You will have to make do with the behavior that exists.</div><div><br /></div><div>David J.</div>
</blockquote>