Created attachment 170142 [details]
LibreOffice rendering of sample input
In Math, digits in variable names seem to have leading and trailing "implied separators". E.g. a variable named "r2d2" becomes "r 2 d 2" with italic letters and plain (no character emphasis/embellishment) digits.
This may be seen as a useability improvement in many cases. To me this "second guessing the users intent" is a regression, and typical of the reasons why I still prefer Apache OpenOffice in many cases.
c3po > r2d2
7of9 neq Agent355
Attached PDFs showing how sample input is rendered in LO and OO.
Created attachment 170143 [details]
Apache OpenOffice Math rendering of sample input
Worked as expected in Version: 18.104.22.168
Build ID: 89ea49ddacd9aa532507cbf852f2bb22b1ace28
Problem reproducible in Version: 22.214.171.124
Build ID: 05dceb5d363845f2cf968344d7adab8dcfb2ba71
Although it could had been an intentional change, it is backward-incompatible and breaking existing data, so (tentatively) marking regression.
This seems to have begun at the below commit.
Adding Cc: to Frédéric Wang; Could you possibly take a look at this one?
352223248b6d0c3198e3bd741e7a47b27e395aa4 is the first bad commit
Author: Matthew Francis <email@example.com>
Date: Sat Sep 5 18:36:17 2015 +0800
Author: Frédéric Wang <firstname.lastname@example.org>
AuthorDate: Mon Jun 24 23:14:42 2013 +0200
Commit: Fridrich Strba <email@example.com>
CommitDate: Fri Jun 28 09:52:08 2013 +0000
fdo#55853 - improve parsing of concatenated numbers and identifiers.