Bug 96078 - rsid should not be added to automatic styles
Summary: rsid should not be added to automatic styles
Status: UNCONFIRMED
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: filters and storage (show other bugs)
Version:
(earliest affected)
4.4.6.3 release
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-11-26 06:53 UTC by Oliver Specht (CIB)
Modified: 2025-01-23 06:33 UTC (History)
3 users (show)

See Also:
Crash report or crash signature:

Attachments
resulting odt file (19.42 KB, application/vnd.oasis.opendocument.text)
2015-11-26 06:53 UTC, Oliver Specht (CIB) 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Oliver Specht (CIB) 2015-11-26 06:53:58 UTC 
Created attachment 120805 [details]
resulting odt file

The import of words rsid results in odt files with lots of automatic styles that only differ by rsid. The rsid should be stored as attribute of the element they belong to instead of storing them as style information.

The resulting change of the file format should be documented at https://wiki.documentfoundation.org/Development/ODF_Implementer_Notes/List_of_LibreOffice_ODF_Extensions

example styles:
<style:style style:name="T9" style:family="text">
<style:text-properties fo:font-weight="normal" style:font-weight-asian="normal" style:font-weight-complex="normal"/>
</style:style>
<style:style style:name="T10" style:family="text">
<style:text-properties fo:font-weight="normal" officeooo:rsid="001c075f" style:font-weight-asian="normal" style:font-weight-complex="normal"/>
</style:style>
<style:style style:name="T11" style:family="text">
<style:text-properties fo:font-weight="normal" officeooo:rsid="001d0c2e" style:font-weight-asian="normal" style:font-weight-complex="normal"/>
</style:style>
Comment 1 Buovjaga 2015-11-28 18:44:26 UTC 
I trust you and will set this to NEW..
Comment 2 QA Administrators 2017-01-03 19:37:04 UTC Comment hidden (obsolete)
Comment 3 Toms 2018-07-06 16:00:21 UTC 
I'd like to note that this behaviour of creating a new automatic style just because of a changed officeooo:rsid tag makes it difficult for other programs to identify tags. For example, translation software usually provides all tags so a translation can retain meaningless formatting in the target document (bold/italic etc.).

However, in OmegaT, for example, due to officeooo:rsid tags a sentence may look like: LibreOffice <f0>simply </f0>create<f1>s</f1> too many tags <f2>. And they are too hard to understand</f2> when translating.

Some tags may hold some formatting, but most are just rsid tags.

Therefore it makes it challenging to translate ODF documents that contain rsid tags because a user has to guess which tags are responsible for formatting and which are just empty rsid tags.

Ideally, the software would be able to simply delete and/or hide rsid tags, but it seems very difficult if they are baked in styles, because a style can represent different formatting too.

Correct me if I'm wrong and they shoudn't cause problems. I'm not really a developer. I know that it's possible to avoid saving these tags, but It would be much nicer if this mechanism of rsid tags would be more transparent and other software could simply deal with them.
Comment 4 QA Administrators 2019-07-07 03:40:12 UTC Comment hidden (obsolete)
Comment 5 QA Administrators 2021-07-07 04:03:08 UTC Comment hidden (obsolete)
Comment 6 Jérôme 2022-05-01 13:07:35 UTC 
I'm experiencing this bug in the below version.
Version: 7.3.2.2 / LibreOffice Community
Build ID: 454130fadb9a820d3728b86ccb63c8f359d70528
CPU threads: 4; OS: Linux 4.9; UI render: default; VCL: gtk3
Locale: fr-FR (fr_FR.utf8); UI: fr-FR
Calc: threaded
Comment 7 Jérôme 2022-05-01 13:22:52 UTC 
In addition, this design induced an other bug on a 7.0 release (that I couldn't reproduce it with the 7.3). The formatting was odd : changing the style properties make appear the inserted/modified part with a different format. The user had to clear the direct formatting everywhere even if he/she didn't use direct formatting.
Comment 8 Jim Avera 2023-09-05 20:20:55 UTC 
Has more clarity emerged about how to fix the RSID problem?

The presence of rsid properties in many styles and the associated spans is wreaking havoc with downstream applications.  In my case I need to clone existing content, such as an existing paragraph or table, but that later crashes LO because, I guess, LO expects rsid-bearing styles to be used exactly once.

An old bug https://bugs.documentfoundation.org/show_bug.cgi?id=68183 says that LO 4.5.0 got an option to prevent using rsids altogether, or at least remove them when saving a document.   However if that feature still exists is is not documented -- the Help system returns nothing when searching for "rsid".

I think the following is needed:

1. If an option to omit rsid exists, document it so it can be used.

2. Evaluate the original request in this bug.   If the function can be handled without polluting styles used for other purposes then it would be a big improvement.

Thanks.
Comment 9 Mike Kaganski 2023-09-06 15:28:56 UTC 
(In reply to Jim Avera from comment #8)
> The presence of rsid properties in many styles and the associated spans is
> wreaking havoc with downstream applications.

It is what needs to be fixed in downstream applications.

> In my case I need to clone
> existing content, such as an existing paragraph or table, but that later
> crashes LO because, I guess, LO expects rsid-bearing styles to be used
> exactly once.

This needs a reproducer and a proper crash fix.

> An old bug https://bugs.documentfoundation.org/show_bug.cgi?id=68183 says
> that LO 4.5.0 got an option to prevent using rsids altogether, or at least
> remove them when saving a document.   However if that feature still exists
> is is not documented -- the Help system returns nothing when searching for
> "rsid".

https://help.libreoffice.org/master/en-US/text/shared/optionen/01040800.html?DbPAR=SHARED#bm_id481597340419434

There seems to be no need to provide any entry for "rsid", just the same way as no need to provide help articles for "p", "span", "sequence-decl", or hundreds of other XML entry names, that are not user-visible. There is a "random number" UI element, and it has respective entry in help.
Comment 10 Mike Kaganski 2023-09-06 15:37:43 UTC 
It is unclear why the proposal would be any improvement. In ODF, all properties are applied using styles (in case of direct properties, that's automatic styles). RSID is just one such property. IMO, the proposal needs to discuss the problems that it addresses, to allow discussion. Its "NEW" status should be "NEEDINFO"...
Comment 11 PrimaWill 2024-03-14 10:22:01 UTC Comment hidden (spam)
Comment 12 steve 2024-12-15 20:41:56 UTC Comment hidden (spam)
Comment 13 DGC BPO Service 2025-01-10 21:26:10 UTC Comment hidden (spam)
Comment 14 Lina Kuri 2025-01-23 06:29:50 UTC Comment hidden (spam)
Comment 15 Lina Kuri 2025-01-23 06:33:18 UTC Comment hidden (spam)