@guy038 said: This time, I would avoid any option that excludes the EOLs and I propose this new sub-menu : A Invert only the marked text B Invert only the bookmarks D Invert bookmarks, then mark all text and EOL of the bookmarked lines C Invert marked text, then bookmark all lines containing marked text Note that running, consecutively, the A and B actions is equivalent to the D action No, it isn’t. If you search for something using the Mark command, only the text that actually matches is marked. Invert marked text marks everything that wasn’t marked and unmarks everything that was marked, so it marks the text surrounding the matches but leaves gaps where the matches were. Potentially useful if you want to search only text that the last search didn’t match. If you had Bookmark lines when marking text checked when you did that Mark command, there will have been a bookmark on every line that contained any marked text. If you then Invert bookmarks, the bookmarks will be only on lines that had no marked text at all. Put the two together and you’ll have marked text with gaps where the original matches were, and bookmarks only on lines that have no gaps. Invert bookmarks and mark only bookmarked lines would bookmark the same lines as the above process, but every bookmarked line (that is, every line that originally contained no part of any match) would have all its text marked. The details of which parts of lines matched and which didn’t would be completely lost. An additional complication is zero-length matches. It is impossible to mark a zero-length match; but when marking with Bookmark lines when marking text checked, the lines containing zero-length matches are still bookmarked. When you invert bookmarks, a line containing one or more zero-length matches will then not be bookmarked. If you invert bookmarks first and then mark the bookmarked lines, that line won’t be bookmarked or marked. If you invert the marks and the bookmarks separately, the line will be fully marked (because it had no marked text), but not bookmarked. You might have meant to say that A followed by B is equivalent to C, rather than D. However, in the zero-length match case, option C would bookmark the line that previously contained only zero-length matches (since it would now be fully marked). I’m still not sure what will be useful and what won’t be; I just wanted to note that it does make a difference whether you invert marks and bookmarks separately or whether you invert one and then derive the other by synchronizing. And of course, if you have marked text that wasn’t marked with bookmarking on, or if you added or removed bookmarks by hand, that would also lead to different results for the various operations.

Thanks a lot @guy038 for running the whole thing again, this time it didn’t finish in seconds ;-) Nice to see the results panel still usable with almost 10 million hits, even the jump to line 690 in that DLL at the very end. Interesting that walking the whole drive for those 130 PDFs takes almost the entire run, 5m45 of 5m48 for MultiReplace, and it’s the same picture with the other tools. The 4 “not decodable” are most likely UTF-16 files with an odd number of bytes. N++ simply drops the last byte there, the next MultiReplace build will do the same in Find in Files (Replace leaves them alone, writing them back would lose that byte). Speed is next on my list.

Feature request created: https://github.com/notepad-plus-plus/notepad-plus-plus/issues/18429 Put the several possibilities given my lack of knowing how the backend works and the feedback I’d gotten on this thread of possible feasibilities. I do stuff on both a Work computer using IT’s selected version of Notepad++ as well as home use, so even if I could get AutoIt working on my personal computer, I probably can’t get it running on my Work computer (permissions) I’ll try to train myself to alt+d

@Paul-Adamczyk said: Still much to learn. Useful References: Notepad++ Online User Manual: Searching/Regex FAQ: Where to find other regular expressions (regex) documentation

Here are the latest binaries of SmartMath. This is a fully-working unofficial build that is going to be added to the official git repository… unless more improvements are suggested :) What’s new: the archive structure corresponds to the expectations of the Plugins Admin; the plugin searches for its documentation file near the SmartMath.dll, as well as in the “plugins\doc” folder; the plugin shows a warning when it does not find its syntax files in the expected locations under the “userDefineLangs” and “autoCompletion” folders of Notepad++. 64-bit version: https://drive.google.com/file/d/1pTS_cBdU_Q0vkG03cubqWZ9TDdYiSsXW/view?usp=drive_link 32-bit version: https://drive.google.com/file/d/1qsI5Hsx4XmyRlhVeGsplAoTX391WEnOv/view?usp=drive_link

@deleelee said: Can anyone tell me why the search result for line 9 displays only a single space after the asterisk instead a full tab space like all the other lines? I’m certain there is a tab after the asterisk in line 9 in the search results window. You can verify that by highlighting the asterisk through and including the following “A”, copying, then pasting that into your document (or a new tab). The visual results are a combination of the way tabs work in Notepad++ — so-called “typewriter tabs” — and using a proportional font. Tab stops are at fixed positions, spaced by the distance equal to the width of a chosen number of spaces in the font you’re using. (In document tabs, that number is chosen in Settings | Preferences… | Indentation | Indent size. However, regardless of that setting, in the search results window the number is always 4.) A tab character renders as a space of whatever width is needed to put the next character at the next tab stop. Notice that the colons after single digit and double digit line numbers don’t quite line up. That is because you are using a proportional font. Notepad++ adds a space to make up for the missing digit, and with a fixed width font, that would line them up. You are using a proportionally spaced font, and a space is not as wide as a digit; so the colons don’t line up. After the colon comes a space, then an asterisk, then a tab. The difference in placement of the colons happens to be just enough that the tab for lines with single-digit numbers finds the next tab stop at a different position than the tab in lines with two-digit numbers.

@PeterJones Thank you now as for my other questions …

@rdipardo , I got a statement of sorts in regard to this request, specifically, this comment: Using AI to create a PR for an open source project - is it a good thing or a bad thing? I have no judgement on that. For me, the only criterion that matters is whether the merged code will be maintainable in the future. On this point, I have to say that AI-generated PRs (I can usually tell during code review), often do not meet the critera in their current state - but many human devs don’t either.

@PeterJones Oh!! Understood!! thank you!!!

XBrackets Lite v2.0.4 Glory to Ukraine! Glory to the heroes! improved: a manually typed closing bracket steps over an auto-completed one, even after typing text inside the pair. For example, typing ( automatically inserts ), producing (|), where | indicates the caret position. Typing abc places the characters inside the pair, resulting in (abc|). Typing the final ) then steps over the existing ), resulting in (abc)| instead of adding a duplicate closing bracket (abc)). changed: while searching for “XBrackets.txt”, the plugin first checks the “plugins\XBrackets” folder, and only then checks the “plugins\doc” folder. a few internal corrections. https://github.com/d0vgan/npp-XBracketsLite/releases/tag/XBrackets_v204

@docpomus , I am sorry no one replied yet. Usually when that happens, it means that either gup.exe or nppPluginList.dll (in subdirectories of your Notepad++ installation folder) still have the “mark of the web”. To check (and clear) the “mark of the web”, exit Notepad++; right click on the exe or dll file; select Properties; in the General properties tab, if there is a message at the bottom with “unblock”, checkmark that checkbox, and click OK. Do that for either/both files. After that, the plugins should be able to update from Plugins Admin again.

@Adam-Narkiewicz , (sorry for the slow reply: when I first read your post, I was in the middle of something, and couldn’t go looking for the issues, so I hoped that someone else would reply with similar information. Since no one has, I am doing it now.) Honestly, I believe the bug is in step 5: it shouldn’t do that “check for underlying file existence” when you tell it to Close All (or when just closing one file, as in Issue #11866; or when you tell it to reload a file, as in Issue #18718). The developer tried (per this comment) to find a way around this, but hasn’t found a viable solution yet. After asking whether to keep a file in the editor, Notepad++ closes it no matter what my answer is. For me, since you already said “close all”, I don’t see a problem in Notepad++ closing it. I am not convinced it should’ve asked in the first place. Fortunately, if there are unsaved changes – so if you’ve edited the file, so there are unsaved changes to the file that doesn’t exist anymore – it will interrupt the process and ask you to save it, and when it fails, it stops the process (not the best solution, but better than losing data). So this oddity at least won’t let you lose unsaved changes accidentally when doing Close All for files that have disappeared. So that’s something. asks me whether to keep some deleted files open, but not all That does seem problematic, to me. I guess after doing the single check, it finally believes your request to Close All meant to Close All. In conclusion: Since Issue #11866 exists for the single file-close, I don’t think this needs a separate official issue. I’m glad it’s “not particularly annoying” to you, because it could be a long time (if ever) before a solution can be found.

I said: In case anyone else is interested, you can try previewing plain-text wireloom files with this development build of Preview HTML. This feature has now been merged, with additional support for offline Markdown previews that will render wireloom code blocks: [image: 1790348955006-wireloomx898.gif] All rendering is done by client-side JavaScript modules, so MS Edge is still required. @Keith-Collyer said: What would be really good would be if the developers of these re-merged their plugins! Don’t look now, but no fewer than three new Markdown plugins have appeared in the past two weeks — and they all depend on the WebView2 runtime (i.e. Edge).

@peterjones Installed Notepad++ v8.9.8.1, and this problem went away [image: 1790343129511-%D0%B1%D0%B5%D0%B7%D1%8B%D0%BC%D1%8F%D0%BD%D0%BD%D1%8B%D0%B9.png] p.s. And I don’t need to unblock files received from the Internet, because after a clean install of any Windows, I will block this option through the registry.