@guy038

More good information. Thank you!

Barry

@fml2 said in Jumping between matching brackets is not idempotent:

What I’m at is the fact that, after two Ctrl+B I’m not back where I was before.

The priority goes to character before caret and if no brace found so check other side.

https://github.com/notepad-plus-plus/notepad-plus-plus/blob/b5ce090bff42e6761f168592a6193eb591b8d021/PowerEditor/src/Notepad_plus.cpp#L2988-L3002

and same with SciTE.

https://sourceforge.net/p/scintilla/scite/ci/default/tree/src/SciTEBase.cxx#l628

It might be sloppy behavior rather than strict behavior, though there is intelligence if you can recognize it as it is not random. It is misfortunate that the caret | with [(ab)c] goes from )|c to [|( and then to c|] though it is the priority that determines where the caret goes. At [|(, the priority of the match goes to [ before the caret, not to ( after the caret.

@Eric-Wiseman said in Bug in sorting lines, staying there for years:

Line Operations -> Sort lines works ONLY if you convert Edit -> EOL Conversion ->Unix.

Unless I am missing something this looks like a bug…

As others have already explained, this is not a bug but rather a known and documented limitation, that all line endings must be the same.

If Sort Lines In Locale Order Ascending and Descending do what you want, those sorts implemented differently and are not dependent on consistent line endings. The ending for each line will remain with the text of the line it follows.

The Locale sorts can also sort using a rectangular selection for the keys even if there are tabs or characters of differing byte widths (such as can happen with non-ASCII characters in UTF-8) to the left of the selection, which does not work with the other sorts.

@Todd-Marshall ,

Some comments, based on your two most recent posts:

We had no way to know how exactly you used the AI – it could have been either of the following: you prompted it to “take a common problem found in Notepad++ and propose a solution” This version was why I suggested it might have been hallucination Since you didn’t say earlier, I had no way to know this wasn’t what had happened. you prompted it with "brace matching is not working for me in Notepad++, with details X, Y, Z. could you re-write this as a “proposal for improvement” This version was why I suggested that it was bad data fed into the prompt (ie, that you were wrong) This turned out to be closer to the truth, as you described. Both in regards to what you actually asked the AI to do, and with regard to the fact that you were wrong or incomplete in your description to the AI This was a lot of extra effort on your part. Instead of explaining to the AI what you wanted to say, and then have it re-write your text in its own words, you could have just explained it to us in your own words. And it would have been better, because it wouldn’t have introduced the AI-garbage uncertainty. And you wouldn’t have been violating the forums rules for posting. So it would have been better all around. Your posts don’t even match the AI text you posted. In your two most-recent posts, you claim you suggested Ctrl+D, but your AI text actually suggested Ctrl+] … that’s two different keystrokes. So by going through the AI intermediary, you didn’t even know what you posted. Re: Documentation: I have no idea what you mean by “mired in COIK” As the primary maintainer of the Notepad++ documentation, if you have concrete examples of how the Online User Manual is “poorly organized”, feel free to create an issue in the user manual repo with concrete descriptions… but understand that the basic structure is pretty set-in-stone at this point. But if you really meant, “I have been burned by other documentation before, so I don’t bother reading any documentation”, the onus is still on you to give each new application’s documentation a chance before rejecting it out-of-hand. I really don’t care about how good or bad the documentation for a major corporation’s huge flagship product is, and you basing your decision on whether or not to read any documentation based on a bad experience with a huge, gigantic application like that is rather unfair of you to any other documentation Re: the root cause being mismatched braces: you cannot really expect a feature that says “go to matching brace” to work as you described you wanted when the braces aren’t matching. And, in fact, the way that Notepad++ does work helps to indicate that the braces are mismatched, now that you know what that symptom means.

To sum up: don’t post AI-generated content in a Community of human Users of Notepad++; it’s not allowed, and it didn’t work well for you when you did.

@SammyBolt said in autoupdater and connection temp.sh:

Regarding the WinGup fix in v8.8.8, Was the temp.sh (or similar) exfiltration vector reported in v8.8.4 possible on a clean, official installation of Notepad++?

The temporary fix prevents the hijacking attempts that redirect to the domain other than github.com. However, it cannot protect against malware that is hosted on github.com.

The full fix will be included in the upcoming version, as mentioned below.

You mentioned the fix landed in v8.8.8, but can you confirm if this vulnerability was introduced specifically in v8.8.4 (due to changes in the updater), or were versions prior to 8.8.4 (like 8.8.2/8.8.3) also susceptible to this specific gup.exe hijacking technique? What about 8.8.5-8.8.7?

We still cannot determine the exact method used by the attacker. But it is not a case of “a vulnerability introduced in version X”.

For users currently running v8.8.4 who want to verify they haven’t been impacted before updating: Aside from monitoring network traffic for temp.sh, are there specific things we can check for, like modifications to the update xml file or other modified files/logs in the program directory?

I do not have complete information to provide specific advice. However, here’s an article from a security expert who appears to have investigated this issue:
https://doublepulsar.com/small-numbers-of-notepad-users-reporting-security-woes-371d7a3fd2d9

My suggestion is download & install v8.8.8 manually from the official website.
The upcoming release v8.8.9 will be available in a few days. In this release the code signing certificates will be verified on the downloaded binary before update installation.
Then please update to v8.8.9, in which the vulnerability is fully addressed.

@h-jangra

It could be, yes, because I downloaded it and ran setup_arm64.bat in a normal cmd shell, and it worked.
But it could also be due to how you originally set up Rust. I also played around with this a bit, but couldn’t reproduce your problem.
I had MSVC installed via the Visual Studio installer, and then I installed the Rust toolchain. Some weeks later I add the arm extensions for VS and added the target to Rust, and it works for me.

@Ekopalypse I tried simple plugin now using go. It did worked, let see if I can improve. If its good then I should make template for it.

@h-jangra

I doubt that you can implement this in Scintilla in a reasonable way, but the other idea of having a preview is there, and there are at least two plugins, here and here that can give you some ideas.

@PeterJones I think I found the culprit behind this.

Looks like I had my Avast autoupdater to ON, and probably it was the reason behind quiet updates for both Notepad++ and my Mozilla Thunderbird 42.

It became a routine for me to reinstall old versions over and over and over agian.

Thank you for your response.

Really appreciate your help and detailed approach.

@PeterJones
Thank you for the reply. Yes, I understood that inference. I was hoping there was additional tweaks to try. Yes, I could do a reply. It was triggered on a very minor edit. Maybe the site support would be interested in the issue. But it’s not important to me now.

@Rick-Kosters

You can try the free Agent Ransack utility from https://www.mythicsoft.com
It works with .doc, .docx and .pdf files too.

@Lois said in Removing lines help - Confusing ,don’t know how to explain but will try and but need help please:

And I’m basically needing the lines that start their own line with a capital word joined into the upper line to link to that sentence above

Well, I would think you could potentially have any sort of alphabetical characters at the start of a line, not necessarily just a word starting with a capital letter. So this expression looks for any line NOT starting with a number and combines it with the previous line. Again have a run with using the Find Next, then Replace on the first run to check that it solves your need.

Find What:\R([^\d])
Replace With:\x20${1}
Note that the \x20 is the code for a space. Often it’s better to show something rather than the absence of a visible character, but you can just type a single space in that position if you understand it is a position in the expression.

Terry

EDIT. Actually in hindsight I think this looks better. It tests for the characters at the start of a line not being a 4 digit one such as the year as the other lines have. If the expression is true (not a year following) then it replaces the carriage return\line feed with a space.
Find What:\R(?!\d{4})
Replace With:\x20

@KingRustamus

I’ve encountered such stuff before and it was caused by the Windows ransomware protection “Controlled folder access” in action. (build in Notepad.exe works in such a case because all the Microsoft own signed binaries have a build-in exception).

So check (try to set temporary off or add a new Notepad++.exe exception there) the Windows Settings > Privacy & Security > Windows Security > Virus & threat protection > Ransomware protection > Controlled folder access.

If it doesn’t help, I can prepare a testing build of the notepad++.exe binary showing us exact system error code (and it’s not the ERROR_ACCESS_DENIED for sure, otherwise in the v8.8.6+ you saw the UAC-elevation prompt 1st) leading to this SaveFailed N++ messagebox.

@PeterJones

Thanks for pointing me to that Github issue. I somewhat researched the topic (using VS code/Electron) repos and added a suggestion there.

The issue, I think, is that instead of using a dedicated shell API, Notepad++ tries to manually launch an Explorer process.

@James-Burke-0 ,

Also, please don’t waste our time by asking the same question repeatedly in different ways. I told you this in early November, in this post, where I said, “instead of providing you a Save button…”, and went on to explain the same thing.

It’s very annoying to go to all the trouble of giving a detailed answer, and then to get proof that the answer is completely ignored by the person who asked.

@Jeff-Esposito

Hi, I had the same problem.

In my case it was (alsmost certain) caused by heavily editing the DarkMode-Default theme that got a size of over 500Mb. I Suspect it was caused by using many custom colors.

It has nothing to do with BitDefender. We have the BitDefender Endpoint Security running on our machines too.

Switching from dark mode Default to another theme completely solved the problem for me.
\Editing that theme with default colors did not slow down startup.
Give it a try.

Tip:

You can keep your personal theme by coping and renaming it. It will them show up in the themes
list. You can find the themes in C:\Users\Username\AppData\Roaming\Notepad++

Gerard Kiekens
(Key-Can Computer Services The Netherlands)

@bvklaveren said in sha512 certificate cannot be validated:

I have seen the instructions, but installing the self-signed certificate on all computers in the organization is not a practical solution.

Okay. Then ignore the self-signed, since there’s a commercially-signed alternative that can give you confidence that the binary is being distributed by the one who it claims to be.

It only takes one certificate to give the confidence, but it also means that you can ignore any other certificate on the same file, because having two certificates (one self-signed that you choose to distrust and the other commerically-signed which you choose to trust) gives you exactly the same amount of confidence that the binary is valid as does having just the commercially-signed certificate with none others. Whereas for someone who does trust the self-signed, they can choose to trust that self-signed and ignore the commercially-signed as being tainted by money.

Also, why would you include a self-signed certificate if you also have “real” one that can be validated without extra steps?

The “why’s and wherefore’s” are all available in the public, in various discussions on this forum (search for certificate) and the relevant section of the User Manual and in the various announcement pages in the official website. But I’ll sum it up, from my outside perspective:

Because for months, there wasn’t a commercial certificate available, because the corporations behind the certificates are set up to make things difficult to open source projects. And once he did get the new globalsign cert, he didn’t feel it was worth it to undo all the effort he had put into creating the self-signed certificate.

Lastly, it is not in the msi, so why is it in the exe?

apples and oranges?

The MSI is a container, which can be signed, and is signed separately from any EXE or DLL that might be contined therein. So the developer apparently chose to only sign that MSI with the commercially-derived – probably because his system was already setup to sign the EXE with both (because it was setup to sign with just the self-signed, then he added the globalsign cert on top of that once it was available), whereas the MSI didn’t come until a month or two after the globalsign was available, so he probably didn’t think it was worth the extra effort to figure out how to sign the MSI with the self-signed as well.

@pbarney said in PHP Linter PythonScript:

my impression is that Pythonscript ships with 2.8

From Plugins Admin, PythonScript v2.1.0 uses Python v2.7.18 (as far as I am aware, there is no such thing as Python v2.8).

However, if you go to the PythonScript releases page, you will see that there is also an “alpha” track of PythonScript v3.0.x (currently at 3.0.24), which use Python v3.12.xx. This is the route I personally recommend (caveat below).

The maintainer of PythonScript has chosen not to take PythonScript 3 out of “alpha” until it fully supports non-unicode encodings (ie, until it supports the 1980s-style character set encodings), so until that point, to get a maintained version of the Python interpreter for PythonScript, you have to manually install an alpha version of PythonScript – essentially, you have to put up with legacy syntax of Python if you want to deal with legacy non-unicode encodings of text. I personally have no need of supporting 80s technology for file encodings (and I pity anyone for whom their circumstances demand that they deal with file encodings that haven’t been brought into this millennium anytime in the last quarter century), so I use the 2020s version of Python for my PythonScript. (My personal opinion is that PythonScript should start shipping PythonScript 3 as non-alpha, and that they just prominently make it clear that if you have to deal with the old character sets that you should instead install the outdated PythonScript2 version from the releases page. But my opinion on such things holds no weight.)

Thank you for all of the information. While it might be fun to dig into the Windows APIs again, I was hoping for something simple I could run with NppExec, etc. without a lot of coding. Since it’s not a lot of extra work to move the splitter manually, I’ll live with that for now. And maybe the MarkdownPanel team will find my request worth implementing, though it’s possible it can’t be done since one side is a NPP tab and the other is a panel.

@PeterJones this worked perfectly! Thank you very much