Regex assumed all numeric characters were part of the tag number
This is obviously false as $9 would be a likely candidate
to be removed on export.
The code can be any printing ascii character other than space
according to LoC's website.
Also changed regexp to no longer allow a zero length tag number
which is nonsensical.
The old regex would accept shorter than 3 digit tags but these were
not stripped so I've removed that option, considering it a bug
not a feature
NB assumption that the code makes that a tag is always numeric is
incorrect but works in practice. Handling non-numeric tags is
a 'To be done'
Made code dependent on the regex succeeding. Picking up results
from a previous regex on failure can lead to weird hard to
identify bugs
In the course of testing Paola Rossi <paola.rossi@cineca.it>
spotted that the delete operation was flawed, only removing
the first occurence of the specified tag/tag subfield. Reworked
the delete loop to operate on all occurences
Signed-off-by: Paola Rossi <paola.rossi@cineca.it>
Signed-off-by: Jonathan Druart <jonathan.druart@biblibre.com>
Signed-off-by: Galen Charlton <gmc@esilibrary.com>
248283fc29