Outputs "\r\n t" instead of letter i

It gives line break and letter t instead of letter i in some parts of the related file. Such an error may occur because the relevant text is in Turkish. However, there are places in the same text that do not give this error.
Sample:

Talep ed
tlen belgeler
t (KPSS sonuç belges
t ve yabancı d
tl b
tlg
ts
t sev
tyes
tn
t gösteren
belge har
tç) eks
tk
tbraz eden veya h
tç
tbraz etmeyenler bu belgeler
t son başvuru tar
th
tne kadar
Başkanlığımıza
tbraz ett
tkler
t takd
trde talepler
t kabul ed
tlecek, son başvuru tar
th
tnden sonra
tbraz

Related file: https://www.resmigazete.gov.tr/ilanlar/eskiilanlar/2021/11/20211126-4-9.pdf
Note:Since the file is taken from the Official Gazette, it is not subject to any copyright.

Start by reproducing the extraction problem with the linked Official Gazette PDF and trace where its text is decoded or assembled in the parser. Compare the affected Turkish characters with unaffected text and add a regression case if the relevant test location can be identified; done means the sample no longer produces line breaks and incorrect “t” characters.