GitLab

Kai Uwe Broulik authored Oct 08, 2021 and Volker Krause committed Oct 13, 2021

Otherwise timezone information is lost

We must not apply the same codec twice there, as the input to the HTML
encoding is UTF-8 in that case, not what might be specified in the HTML
content-type header.

This is a problem for e.g. Japanese and Korean codecs in particular, where
doing this wrongly completely destroys the content.

Kai Uwe Broulik authored Sep 10, 2021 and Volker Krause committed Sep 15, 2021

This seems to be a popular infrastructure provider for tourism
discount cards in Austria.

BUG: 441361

On top of the generic IATA BCBP extraction this now adds times and airport
names.

Some tickets diverge so far from the standard layout that all our recovery
attempts fail. We now also try a set of regular expression patterns on
the layouted text.

This makes MAV (Hungarian state railway) produced RCT2 tickets readable.

This for example protects against accidental infinite loops in extractor
scripts.

Evangelos Foutras authored Aug 02, 2021 and Volker Krause committed Aug 02, 2021

parseDateString() now takes "const GooString *" as the date argument in
order to support unicode date strings. Adjust parsePdfDateTime() to use
the same string type (instead of "const char *") when compiling against
poppler 21.8.0.

https://gitlab.freedesktop.org/poppler/poppler/-/merge_requests/880

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

Using signed char is documented to be undefined behavior.