| Summary: | Plenty of list-related character styles created on DOCX export which clearly not needed | ||
|---|---|---|---|
| Product: | LibreOffice | Reporter: | Telesto <telesto> |
| Component: | Writer | Assignee: | Not Assigned <libreoffice-bugs> |
| Status: | NEW --- | ||
| Severity: | normal | CC: | eyalroz1, jluth, libreoffice, michelle, nemeth, telesto, vmiklos |
| Priority: | medium | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | All | ||
| See Also: |
https://bugs.documentfoundation.org/show_bug.cgi?id=133410 https://bugs.documentfoundation.org/show_bug.cgi?id=147673 |
||
| Whiteboard: | |||
| Crash report or crash signature: | Regression By: | ||
| Bug Depends on: | |||
| Bug Blocks: | 108769 | ||
| Attachments: |
Example file
Simple reproducer file The reproducer document saved to docx The docx version of the reproducer saved back to odt The original reproducer and its docx version The docx version and the odt saved from it |
||
|
Description
Telesto
2021-02-07 11:11:14 UTC
Created attachment 169550 [details]
Example file
@Justin We had a bug about this somewhere.. And if I the styles code was to horribly broken (and to complex). But this might give an approach to to get rid of most of the junk styles without a big re-factoring Specifically, these are character styles ListLabel1 - ListLabel171. A second round-trip brings it to ListLabel225. A similar report for DOC is bug 133410 which says, see LO 5.0.6's tdf#95213 DOCX import: don't reuse list label styles. which was mollified somewhat by LO 6.3's tdf#92335 DOCX: fix multiplying of "ListLabel" styles. Especially important is "However, making a change here would be fraught with danger." @Miklos I do perceive the whole Style creation as an rather substantial issue (usability, navigator mess + slow file opening at some point). However Justin already pointed out this being 'fraught with danger'. So this needs someone with lots of expertise (and loving to clean-up potential fall-out) However no clue who is up to the task (nor how much time/budget this would involve). So my way of 'raising awareness' (say ESC). In the hopes someone will accept the challenge (some day) (In reply to Justin L from comment #3) > Specifically, these are character styles ListLabel1 - ListLabel171. A second > round-trip brings it to ListLabel225. > > A similar report for DOC is bug 133410 which says, > see LO 5.0.6's tdf#95213 DOCX import: don't reuse list label styles. > > which was mollified somewhat by LO 6.3's tdf#92335 DOCX: fix multiplying of > "ListLabel" styles. > > Especially important is "However, making a change here would be fraught with > danger." If you want to put something to the ESC agenda, just reply to the agenda emails on the dev list and come to the ESC call. Thanks. The current one is here: https://lists.freedesktop.org/archives/libreoffice/2021-March/086942.html If you don't feel like coming to that call, then just send a mail to the dev list, so more people can take part in the discussion. I think the problem here is at import time. Doing this: Steps to Reproduce: 1. Open the attached file 2. Save as DOCX 3. File reload 4. Go to character styles.. notice a whole list of additional styles does not result in extra character styles being saved to the docx files styles.xml (since bug #92335 was fixed). These extra styles appear only at docx import time, then these can be saved to odt (unfortunately). Created attachment 170749 [details]
Simple reproducer file
Created attachment 170750 [details]
The reproducer document saved to docx
Created attachment 170751 [details]
The docx version of the reproducer saved back to odt
This has the extra character styles saved to sytles.xml
Created attachment 170752 [details]
The original reproducer and its docx version
The problem starts at the docx import.
Created attachment 170753 [details] The docx version and the odt saved from it Once these fake ListLabel character styles are created they are saved to odt. Version: 7.2.0.0.alpha0+ (x64) / LibreOffice Community Build ID: 2fb274950e5207ca55f4f52325fb522bd44024e1 CPU threads: 4; OS: Windows 10.0 Build 18363; UI render: Skia/Raster; VCL: win Locale: en-US (hu_HU); UI: en-US Calc: CL (In reply to NISZ LibreOffice Team from comment #11) > Once these fake ListLabel character styles are created they are saved to odt. I don't think we would want to do anything to try and stop this though. However, one possibility would be to create a "clean up document" tool that goes through and removes excess styles, obsolete direct formatting, etc. (This, of course, is a HUGE enhancement idea, and so far we can't even get someone to write a compress-all-pictures tool...) *** Bug 143237 has been marked as a duplicate of this bug. *** (In reply to Justin L from comment #12) > (In reply to NISZ LibreOffice Team from comment #11) > > Once these fake ListLabel character styles are created they are saved to odt. > > I don't think we would want to do anything to try and stop this though. Why? I mean, if styles are generated artifically, why not prevent that from happening? Or at least, use more stringent criteria to decide when a style is generated? > However, one possibility would be to create a "clean up document" tool that > goes through and removes excess styles, obsolete direct formatting, etc. > (This, of course, is a HUGE enhancement idea, and so far we can't even get > someone to write a compress-all-pictures tool...) and it would also be a separate bug IMHO. |