#317: bibdocfile: replace ligatures when textifying
-------------------------+--------------------------------------------------
Reporter: simko | Owner: jani
Type: enhancement | Status: new
Priority: major | Milestone:
Component: *general* | Version:
Keywords: |
-------------------------+--------------------------------------------------
To work around Firefox/Mac ligature display problem, a quick fix was
introduced recently in d24b8dd3c7c30c958536b24ad03f6f311ea9d5ef.
However, the ligatures should be replaced at the textification time
already, because the full-text indexing and searching behaviour is
influenced by ligatures too, e.g. one gets on INSPIRE:
{{{
fulltext:coefficient - 38,412 hits
fulltext:coefficient - 787 hits
}}}
which is not good, because users are not probable to type UTF-8 `ffi'
character at the search time :). By expanding ligatures at the full-text
textification time, the indexer and the searcher would work in ASCII-like
mode so `ffi' would be used in these words everywhere, and Firefox/Mac
workaround would not be needed anymore.
So the `sed` expressions should be basically ported to `bibdocfile`
textification time.
BTW note also that the Firefox/Max workaround currently generates lots of
messages in Apache err file like:
{{{
sed: couldn't write 1377 items to stdout: Broken pipe
sed: couldn't write 3264 items to stdout: Broken pipe
sed: couldn't write 58 items to stdout: Broken pipe
sed: couldn't write 399 items to stdout: Broken pipe
sed: couldn't write 757 items to stdout: Broken pipe
sed: couldn't write 255 items to stdout: Broken pipe
}}}
--
Ticket URL: <https://invenio-software.org/ticket/317>
Invenio <http://invenio-software.org>
