Spell Check stops before the end of long documents

Do you want to: Report a bug
OS version: Win 10 Pro 22H2
App version: (x64 exe)
Downloaded from: ONLYOFFICE website

I have what could be defined as a “big document”.
214 pages, so far. 100250 words (according to OO “word count”)
size is currently 38.5 MB (40,431,277 bytes), it has pictures…

Spell check gets to page 199, and then stops. I know there are words to be flagged further on; I can see “didn’t” on page 201 (that is being flagged as a spelling error waiting on internal number - 41724 to fix), as as the document has been open now for 25 minutes, if it was going to get to it, it would have happened by now.

And I pretty sure it’s not a resource issue;
RAM is at 31% of 32Gb CPU is idle (bouncing around 0% to 8%)

I’d break the document up into smaller chunks, but that would break the ToC given, as far as I can tell, OO does not support document merge.

… although, the document could be assembled “on demand” at pdf creation via the clipboard.

Hello @DavidRGreen

That is quite strange because I wasn’t able to reproduce the issue. I’ve created brand new document and inserted the same text into it, then by the page 200 added few sentences with errors and they got highlighted after a slight delay:

Did you create the document in Desktop Editors initially? Can you run the test according to the scenario but with an empty file please?

Hi @Constantine ,

Yes, the document was built up, effectively one chapter per week.
It may not be related to the page count, but more with the “content weight” so to speak, as in more to do with the amount of content.

I can see from your screen shots that your content is fairly light, many short paragraphs, and I would guess that it is no where near the 34Mb file size of my document.

I loaded up the document just now, and spell-check has stalled at first line of page 203 this time.

Screenshot 2023-10-04 152805

This screen shot is of page 203, it has been more than 5 minutes since the document was opened. The first highlight shows “droids” as an unrecognised word, but there are three more occurrences of the same word further down the page as the spell-check has “run out of steam :smiley:” before getting to them and the rest of the document, at page 214, for now.

As an exercise, I made a copy of the document, and then went through and deleted all images. That reduced the document to ending at page 186, but the spell-check as stalled at part-way down page 174.
Without illustrations, the file is just only 260Kb. And after recycling OO and re-opening the file, spell-check, again, stalled at page 174, but got a little further this time.

But docx is a compressed container, so the actual document is over 1Mb content and XML tags. So, from observation, the spell-check threshold seems to be about 1Mb of content.

Or maybe there is a limit on how many errors can be tracked. There are a ticket still outstanding in regards to words with apostrophes being classed as spelling errors erroneously; words like “didn’t”, “couldn’t”, “I’ve”, etc. These words being tracked could also be an issue.

That’s about all I can suggest at this stage.

Thank you for valuable information. Mind if I ask for the document for analysis?

Also, do I recall correctly that you are using Desktop Editors on Windows 11?

No, still on Win 10, I don’t like Win 11.

Nevermore 3 no-illustrations.docx (262.9 KB)

This is the illustration-free version. It is a light-novel for a mature audience.

… fyi, this is a “live document” with more content to be added. And I have made some edits which have had an interesting effect.

I have eliminated a lot of flagged spelling by;

  • Adding some to the dictionary,
  • and removing contracted words that were not in dialogue; these include isn’t, couldn’t, wasn’t, etc.

Reducing the number of flagged spelling errors seems to have allowed the spell check to get to the end of the document.

I also tried adding “wasn’t” to the dictionary file, but OO was not having that, and flagged that as a spelling error in the document anyway :joy:

There are probably a couple more chapters to add, so it may well break again, but I’ll see.

That is valuable piece of information. We are checking the situation.

We checked the document and this behavior with stopping spellchecker was registered as a bug as a result. Possibly, it will take some time to find out the root of the issue, so, unfortunately, no ETA available yet.

Thank you for the report.

1 Like

I’m guessing that there is some kind of array holding all the flagged spelling errors with their position in the document which is hitting a limit.

If that other bug pending resolution about words containing an apostrophe was resolved, that may well help ease the issue too.

:joy: I now have three OO tickets related to spell-check. That has to be fairly unique. I’ll be looking forward to a resolution on those fronts.

Thanks for the update.