It never happened AFAICT. I had asked for something along the lines of
to be done in the course of Proposed PEP 11 update -- policy for unsupported platforms, but that was deemed off-topic.
As it happens, I recently had to look at this again, because libc++ v22.1 is raising the minimum deployment target to 11.0, which affects us in conda-forge. Below the fold is the same analysis as above (28 days of downloads for requests on osx, split by macOS version).
| version | downloads | percentage | cumulative | percentage | users >=version |
|---|---|---|---|---|---|
| 10.1 | 2 | 0.0000% | 2 | 0.0000% | 100.0000% |
| 10.5 | 7 | 0.0001% | 9 | 0.0001% | 100.0000% |
| 10.6 | 4 | 0.0001% | 13 | 0.0002% | 99.9999% |
| 10.7 | 2 | 0.0000% | 15 | 0.0002% | 99.9998% |
| 10.8 | 3 | 0.0000% | 18 | 0.0002% | 99.9998% |
| 10.9 | 39 | 0.0005% | 57 | 0.0007% | 99.9998% |
| 10.10 | 26 | 0.0003% | 83 | 0.0011% | 99.9993% |
| 10.11 | 65 | 0.0008% | 148 | 0.0019% | 99.9989% |
| 10.12 | 205 | 0.0026% | 353 | 0.0045% | 99.9981% |
| 10.13 | 1,319 | 0.0168% | 1,672 | 0.0213% | 99.9955% |
| 10.14 | 1,470 | 0.0188% | 3,142 | 0.0401% | 99.9787% |
| 10.15 | 15,094 | 0.19% | 18,236 | 0.23% | 99.96% |
| 10.16 | 232,838 | 2.97% | 251,074 | 3.20% | 99.77% |
| 11 | 17,199 | 0.22% | 268,273 | 3.42% | 96.80% |
| 12 | 92,973 | 1.19% | 361,246 | 4.61% | 96.58% |
| 13 | 157,024 | 2.00% | 518,270 | 6.61% | 95.39% |
| 14 | 1,083,349 | 13.82% | 1,601,619 | 20.43% | 93.39% |
| 15 | 4,118,308 | 52.53% | 5,719,927 | 72.97% | 79.57% |
| 26 | 1,948,673 | 24.86% | 7,668,600 | 97.82% | 27.03% |
| 27 | 336 | 0.00% | 7,668,936 | 97.83% | 2.18% |
| Invalid | 9,095 | 0.12% | 7,678,031 | 97.94% | 2.17% |
| None | 161,146 | 2.06% | 7,839,177 | 100.00% | 2.06% |
| Sum | 7,839,177 | 100% |
Note that 10.16 is a misreported version corresponding to something >=11.0.