Fix commits list problem

[kwankyu]

Seen in https://34.105.185.241/sagemath/sage-prod-2023-01-30-077/issues/34956#issuecomment-7308003

Checking regression...

[kwankyu]

Affected cases should be very rare.

It takes me long to check regression. Would you check?

[kwankyu]

Affected cases should be very rare.

No, there are many cases. In all those cases, our converter seems smarter than Trac, as seen in

-I made a change like this, but changing `2008`} to `{2008`}}.  The second brace } matches the one a few lines earlier, in the line `@manual{sage`.  (The change wasn't really necessary, but it doesn't hurt, either.)
+I made a change like this, but changing `2008}` to `{2008}}`.  The second brace } matches the one a few lines earlier, in the line `@manual{sage`.  (The change wasn't really necessary, but it doesn't hurt, either.) 

for trac #1389

[mkoeppe]

Yes, I just ran it and there's a long long diff. pr171.diff.gz

I'll try to find time this evening to review it

[kwankyu]

I checked through the diff file. There are two cases:

• positive change (most of them), like

  -whereas the parent should be `Ordered set partitions of {1,2,3,4,5,6`}
  +whereas the parent should be `Ordered set partitions of {1,2,3,4,5,6}`  

• slightly negative change (a few, where the writer made a mistake of typing 4 braces }}}} instead of 3 }}}), like

  -maybe you should add `r"""`} when the doctests contains something that can perturb the pdf doc ?
  +maybe you should add `r"""}` when the doctests contains something that can perturb the pdf doc ?    

I definitely recommend to take this PR.

[mkoeppe]

I agree, looking good. Thank you!

[kwankyu]

Thanks!