Files
odfstudio/writerfilter
László Németh 7dd8f8aace tdf#120351 DOCX import: fix slow endnote import
by parsing endnotes.xml only once instead
of parsing again and again for every endnotes.
This was a serious performance problem for
documents with hundreds of endnotes, where the
endnote import took minutes instead of seconds.

Note: switch off CHECK_NOTMERGED in a debug build
to measure realistic speed-up, e.g. 98 s -> 18 s
for a document with thousand of endnotes.

Follow-up of commit 9b39ce0e66
"tdf#76260 DOCX import: fix slow footnote import".

Change-Id: I88cdc5101c25041f985fdc23739a0dadf24a13e0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111030
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
2021-02-18 13:30:55 +01:00
..
2020-08-19 09:47:44 +02:00
2020-12-15 14:10:44 +01:00
…
…
…

The writerfilter module contains import filters for Writer, using its UNO API.

Import filter for docx and rtf.

== Module contents ==
 * documentation: RNG schema for the OOXML tokenizer, etc.
 * inc: module-global headers (can be included by any files under source)
 * qa: cppunit tests
 * source: the filters themselves
 * util: UNO passive registration config

== Source contents ==
 * dmapper: the domain mapper, hiding UNO from the tokenizers, used by DOCX and RTF import
   * The incoming traffic of dmapper can be dumped into an XML file in /tmp in
     dbgutil builds, start soffice with the `SW_DEBUG_WRITERFILTER=1`
     environment variable if you want that.
 * filter: the UNO filter service implementations, invoked by UNO and calling
   the dmapper + one of the tokenizers
 * ooxml: the docx tokenizer
 * rtftok: the rtf tokenizer