The SD-10 was less of a bad decision than it was a necessity of the time. In the early 2000s there were hardly any decent active radar-homing (ARM) AAM solutions available to the PAF. For example, there was the R-Darter, which - while an ARM - was becoming a dated solution, and Denel's next-gen BVRAAM programs (e.g. T-Darter) were just not entering development.
In fact, I was speaking to a South African engineer familiar with Denel's dealings with Pakistan a few weeks ago; the company had hoped to engage in a number of projects, but in the mid-2000s, the BRICS mantra and India's lure for lucrative contracts encouraged the company's leadership close the door on Pakistan (and in the process abandon a number of promising programs).
For one reason (e.g. corruption allegations) or another ("IP theft in India" - his words, not mine), Denel was not able to effectively enter the Indian market. "It was a regretful decision" (again, his words, not mine).
The PAF was interested in the MICA-RF, but that tunnel collapsed.
Besides that, there wasn't much of anything else, besides perhaps some even older semi-active radar-homing (SARH) solutions such as the Aspide. Unfortunately, and to
@Bilal Khan 777's point, the PAF has yet to secure a true next-gen AAM solution - i.e. a platform that will stay relevant for not a few years, but at least 10-15 years. In that regard, it is a matter of studying who is in practical possession of the best concepts.