Description: Hello :-) I was doing a search and replace on a file of 190+ blocks of text (the doc started off as one column of words and numbers (file attached) but I used Find & Replace (F&R) to convert the CR to tabs (more about why later). Then I wanted to find all text-items (the names of the countries) and make each one be the start of a line. I used the word-boundary regular expression "alpha" to find the words (i.e. \b[:alpha;] ) That did work, BUT ONLY ON SOME OF THE WORDS! It skipped a large chunk of the file and found another word. I tried \bSlo because I had words Slovakia and Slovenia on the page, but separated by lines with several other words. The F&R DID find them. I was beginning to think "What on Earth is going on here?" I looked at the source webpage for clues to formatting - no help. I tried "Clear all direct formatting": no improvement. I tried [:alpha:][:alpha:] - and that's when I realised a bug report was called for! "Find Next" meandered down the page highlighting pairs of characters at random intervals - no discernible pattern! Large jumps and no similarity between choices of letters! By the way, the reason I was doing all this was to get a table of statistics in the same form as the one in the newspaper (I just highlighted the table on the webpage and did Ctrl-C, then pasted into the Writer document). It gave me 5 lines per country, so I converted all the end-of-lines to tabs, intending to then make all the names (the text-items) begin with "newlines". Then I could read it into Calc as a "Tab-sep-variable" file. If there is a better way to get a copy of a table, please let me know :-)) 'Bye for Now, Brian Howe Steps to Reproduce: 1.Load my (attached) file into Writer 2.Use Find & Replace (with Regular Expressions selected) to "Find" \b[:alpha;] 3. repeat with [:alpha;][:alpha:] Actual Results: Only some of the names of countries were found. Many were skipped over Expected Results: I expected each word to be highlighted after each click of the "Find Next" button. In the second example (step 3 above) I expected the highlighting to step along two characters at a time. Reproducible: Always User Profile Reset: No Additional Info: Version: 7.2.5.2 (x64) / LibreOffice Community Build ID: 499f9727c189e6ef3471021d6132d4c694f357e5 CPU threads: 4; OS: Windows 10.0 Build 19043; UI render: Skia/Raster; VCL: win Locale: en-GB (en_GB); UI: en-GB Calc: threaded The attached file contains the Web-link of the page from which the "table" was copied
Created attachment 177633 [details] Blocks of text with country stats, some not separated I just did a check to ensure that the file I sent (one of 6 stages in the testing process) DID give the behaviour I described (it was the original) and I could not reproduce the misbehaviour! I substituted a later file in the series (one in which I had done some Find & Replace operations, and that is attached now) but I still could not see the misbehaviour! I wonder if it has something to do with the "Replace" attempts I did that went wrong. It may have been cleared by me closing the files. Perhaps when I reopened the files all was well? I don't have the time right now to go through all those "didn't work" operations, but there must have been some reason for the strange behaviour.
(This should probably be marked as NOTABUG.) You have an error in your Regular Expressions. The correct regex for "alphabetic" characters is: - [:alpha:] Notice TWO COLONS. Yours was accidentally using a SEMICOLON at the end. For a full list of what can be used in LibreOffice 7.2 help page: "List of Regular Expressions" https://help.libreoffice.org/latest/en-US/text/shared/01/02100001.html - - - Note: What your original regular expression was looking for: - [:alpha;] "Hey, look for these letters: - a colon - OR 'a' - OR 'l' - OR 'p' - OR 'h' - OR semicolon" That's why you were getting all these weird, "random", combinations appearing.