LibreOffice integration BUG - treating 4 preceding characters as part of citation breaks ITEM

piotr1364

SYSTEM INFO

Version: 24.2.0.3 (X86_64) / LibreOffice Community
Build ID: da48488a73ddd66ea24cf16bbc4f7b9c08e9bea1
CPU threads: 8; OS: Windows 10.0 Build 22000; UI render: Skia/Raster; VCL: win
Locale: pl-PL (pl_PL); UI: pl-PL
Calc: CL threaded

Zotero 6.0.35
Zotero LibreOffice Integration 6.0.4 (installing 6.0.4.oxt from GitHub gives v6.0.3 in LO).

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/x3sb0siqy3yp61apky0m.png

piotr1364

https://u.pcloud.link/publink/show?code=kZycnF0ZjOUs6FoYoLfLclS5zASQ4b1VO1dk

Folder containing "Zotero_bug_citation_keeps_4_characters_before.odt" with detailed description and another files.

adomasven

Can you reproduce this with LibreOffice 7?

piotr1364

Do you wish some specific version? (I would need to install parallel one).

adomasven

The latest version 7 release.

piotr1364

Yes, it's also reproductible in:

Version: 7.6.5.2 (X86_64) / LibreOffice Community
Build ID: 38d5f62f85355c192ef5f1dd47c5c0c0c6d6598b
CPU threads: 8; OS: Windows 10.0 Build 22631; UI render: Skia/Raster; VCL: win
Locale: pl-PL (pl_PL); UI: en-US
Calc: CL threaded

Zotero 6.0.36
LibreOffice integration 6.0.4

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/shm079i6ouah61pmsq06.png

adomasven

I am unable to reproduce this on my setup. Could you send the document which causes the bad output before conversion to support@zotero.org?

piotr1364

The same behaviour in:

Version: 7.6.5.2 (X86_64) / LibreOffice Community
Build ID: 38d5f62f85355c192ef5f1dd47c5c0c0c6d6598b
CPU threads: 8; OS: Windows 10.0 Build 22631; UI render: Skia/Vulkan; VCL: win
Locale: pl-PL (pl_PL); UI: en-US
Calc: CL threaded

Zotero 6.0.36
LibreOffice integration 6.0.4

This is a second computer using the same libraries synced via WebDAV.

If I'll be able, then I'll put the result from yet another computer that has never installed Zotero. So that it should work as a "fresh setup".

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/7z3nbyjtpm768e2t3sf9.png

piotr1364

The third setup behaves the same:

Version: 7.6.5.2 (X86_64) / LibreOffice Community
Build ID: 38d5f62f85355c192ef5f1dd47c5c0c0c6d6598b
CPU threads: 4; OS: Windows 10.0 Build 19045; UI render: Skia/Vulkan; VCL: win
Locale: pl-PL (pl_PL); UI: en-US
Calc: CL threaded

Zotero 6.0.36
LibreOffice integration 6.0.4

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/et4je9m4repqzl4ospxm.png

piotr1364

"Fresh" Zotero library:

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/ftm51ixwhww08eik92b5.png

adomasven

Thanks for testing, but this isn't really useful. We just need a document where the issue is reproducible before the conversion, or steps to reproduce.

piotr1364

Aa, OK. Source documents are provided in the folder linked in the second comment. Steps to reproduce are described in:

Zotero_bug_citation_keeps_4_characters_before.odt 570 629 13.03.2024 19:03 ----


Each of the following are before conversion:

Citations from imported RDF's.odt 17 390 13.03.2024 18:32 ----
LO7.6.5.2_1st_comp.odt 63 696 15.03.2024 10:04 ----
LO7.6.5.2_2nd_comp_A.odt 64 963 15.03.2024 10:09 ----
LO7.6.5.2_3rd_comp_A.odt 61 681 15.03.2024 12:18 ----

adomasven

Unfortunately when I try to download anything from your cloud provider it gives me an error. Could you send one of those documents to support@zotero.org with a link to this thread?

piotr1364

Sure, 5.67 MB attachment has been sent to support@zotero.org.

adomasven

You have only sent in documents after conversion that contain unrelated text in converted field codes. If you are saying that you are having issues converting the document back after having manually modified the exported text, then that is not really a bug. You should not be modifying the converted field codes. If you are seeing a bug where converted field codes include text that they should not, please send in the document before the conversion.

piotr1364

Sorry, I'm confused about your response. Maybe I'm wrong, but I think you meant the attachment from mail titled:
"[BUG] Zotero can't restore citations in foot-&-endnotes when no preceeding text before citation ITEM"
containing file:
112849 convert citation Writer to Word.zip 362 516 15.03.2024 16:47 ----

...but earlier I have sent to you also mail titled:
"[Zotero Forums] adomasven commented on LibreOffice integration BUG - treating 4 preceding characters as part of citation breaks ITEM"
containing file:
112785 Zotero citation eats 4 characters before ITEM (support@zotero.org).zip 5 942 217 15.03.2024 15:08 ----

In this ~6 MB attachment there is file that has detailed description:
Zotero_bug_citation_keeps_4_characters_before.odt 570 629 13.03.2024 19:03 ----


Link for that files
https://filedn.com/l1AJpXWiF2LRTsLNsykIFE4/112785 Zotero citation eats 4 characters before ITEM/


additional screen caption showing steps to reproduce this bug:
Steps to reproduce Zotero forum bug 112785 2024-03-18 16-35-06.mp4

adomasven

Ok, thanks for the report. It's caused by your item in the RDF containing a complex unicode character for the italic letter "μ" in the abstract. Interestingly when saving the item from the item page with the Connector the item is not imported with it. How did you get this item in the first place?

I've created an issue to track it, but we cannot promise any fast resolution. For now the best workaround is to remove any such characters in the item metadata, pressing Refresh in the Zotero tab in LibreOffice, and the convert the document for transfer.

adomasven

@piotr1364 Thanks for the report. LibreOffice will include the fix for this issue in the upcoming versions of 7.6.7 and 24.2.3 (to be released by Apr 21).

piotr1364

@adomasven Thank you for handling this.

[EDIT]:
https://bugs.documentfoundation.org/show_bug.cgi?id=160278

piotr1364

Another issue is that citation ITEM can't be recovered in Word, when this citation ITEM contains soft hyphen (U+00ad). SHY is displayed in Word by negation mark "¬" (U+00ac) https://support.microsoft.com/en-us/office/show-or-hide-tab-marks-in-word-84a53213-5d02-404a-b022-09cae1a3958b.

https://filedn.com/l1AJpXWiF2LRTsLNsykIFE4/112785 Zotero citation eats 4 characters before ITEM/2024-03-26tue - another case/


https://s3.amazonaws.com/zotero.org/images/forums/u4845974/frpgyjx45ar7rscwlztn.png

adomasven

@piotr1364 Where are you getting these items with these characters? Is this the normal item metadata when saved with the connector, or imported via DOI or similar?

piotr1364

@adomasven I don't know when and how the soft hyphens (U+00ad) were introduced.

If about creating item metadata, I'm nearly 100% sure that it was saved by Firefox Zotero Connector (in the past: 2023-03-12 01:28:31). I'm thinking so because today's import (2024-03-27 10:34:51) downloaded PDF and filled all common fields in the item form. So I assume that in the past it was similar, and finally imported automatically by connector without any manual editing.

https://s3.amazonaws.com/zotero.org/images/forums/u4845974/xtywrv9wky1tvv6658g0.png
Today's import by web browser connector does not contain soft hyphens.

piotr1364

Importing from below sources gives "Abstract" field without SHY:

https://www.researchgate.net/publication/351198409_Effect_of_silica_nano-additive_on_flash_point_pour_point_rheological_and_tribological_properties_of_lubricating_engine_oil_an_experimental_study

https://go.gale.com/ps/i.do?id=GALE|A694675678&sid=googleScholar&v=2.1&it=r&linkaccess=abs&issn=13886150&p=AONE&sw=w&userGroupName=anon~3997e4e4&aty=open-web-entry

https://scholar.google.pl/scholar?q=kashefi+2022+effect+of+silica+nano-additive&hl=pl&as_sdt=0&as_vis=1&oi=scholart

https://link.springer.com/article/10.1007/s10973-021-10812-4

https://www.semanticscholar.org/search?q=Effect of silica nano-additive on flash point, pour point, rheological and tribological properties of lubricating engine oil: an experimental study&sort=relevance

https://www.semanticscholar.org/paper/Effect-of-silica-nano-additive-on-flash-point,-pour-Kashefi-Saedodin/6a30ad58de7bf40fa7dc888d6288bad78a28ec7f