Plugin：disable-overlay-close

[hugo529100]

https://github.com/hugo529100/disable-overlay-close

Prevents closing dialogs by tapping outside or pressing ESC. This is especially useful for mobile users where accidental gestures may trigger unwanted dialog closure. Only the close button (X) is allowed to dismiss dialogs.

It doesn't seem to be very useful, but it solves the problem of accidentally closing the file when renaming or adding comment on my phone. It feels like a fifth wheel, but at least it is still hanging on the back like an off-road vehicle, so it will be useful sometimes.

[rejetto]

hey, you don't need a plugin, just this script in admin-panel / custom-html / script

document.body.addEventListener('click', ev => {
  if (!ev.target.classList.contains('dialog-backdrop')) return
  ev.preventDefault()
  ev.stopPropagation()
}, true)

[hugo529100]

wow!!!!good solution! ! ! thanks!!!!!!

and by the ways I trying to let the files show ,just one click the icon ，and click the name is the files menu，but stuck because some happen the block......any same the scrip？

[hugo529100]

Your solution perfectly solved some weird plugin conflict issues I had, nice!

[hugo529100]

finall I find a way with GPThelper did it , the auto-show just click the icon, and click the file name is the menu:

(function () {
  const imageExts = ['jpg','jpeg','png','gif','webp','bmp'];
  const videoExts = ['mp4','webm','ogg','mkv','avi','mov'];
  const isImage = ext => imageExts.includes(ext);
  const isVideo = ext => videoExts.includes(ext);
  const isMedia = ext => isImage(ext) || isVideo(ext);
  const getExt = name => name.split('.').pop().toLowerCase();

  function bindMediaIcons() {
    document.querySelectorAll('li.file').forEach(li => {
      if (li.dataset.bound === '1') return;

      const a = li.querySelector('a[href]');
      const name = a?.textContent?.trim();
      const ext = getExt(name);
      if (!isMedia(ext)) return;

      // 根據檔案類型挑對的圖示選擇器
      let icon = null;
      if (isImage(ext)) {
        icon = li.querySelector('img.thumbnail');
      } else if (isVideo(ext)) {
        icon = li.querySelector('span.icon.fa-video');
      }

      if (!icon) return;

      icon.style.cursor = 'pointer';
      icon.title = '點擊預覽媒體';

      icon.addEventListener('click', e => {
        e.preventDefault();
        e.stopPropagation();

        const list = window.HFS?.state?.list || [];
        const entry = list.find(e => e.n === name);
        if (!entry) return console.warn('[custom.html] 找不到 entry:', name);

        const mediaList = list.filter(e => isMedia(getExt(e.n)));
        HFS.state.selected = mediaList.map(e => e.n);

        setTimeout(() => {
          console.log('[custom.html] 預覽媒體:', entry.name);
          HFS.fileShow(entry, { startPlaying: true });
        }, 0);
      }, true);

      li.dataset.bound = '1';
    });
  }

  bindMediaIcons();

  const observer = new MutationObserver(() => bindMediaIcons());
  observer.observe(document.body, { childList: true, subtree: true });
})();

[rejetto]

i think you could have done something based on my previous short script instead.
ask gpt and tell "can't i do something similar to this, instead? 👀"

as a side note, the current helper is only trained on backend plugins, while you are doing something that is frontend.

[hugo529100]

Ah, yes, it seems I strayed from the original path later on. I wanted to add convenient daily-use features based on a plugin format, but eventually, I ended up working on frontend modifications to bypass certain limitations.

Good idea—I’ll keep following up and learning these techniques, trying out the methods you mentioned. Since my primary language is Chinese (specifically Cantonese) and my thinking is a bit old-fashioned, my expressions tend to lean toward classical or archaic wording. I’ve never really tried this kind of questioning approach for development before, but I’ll keep experimenting.

I’m actually a little worried that my constant trial-and-error might affect the training purity of this development assistant—in simple terms, it might introduce some code pollution.

If there's any risk of code interference in this regard, I'll switch to using DeepSeek for experimentation. After all, I recently learned that frontend development doesn't require overthinking API or backend architecture—just providing targeted frontend code and some prompts can often get the job done. It even bypasses some core aspects of HFS3, significantly reducing my coding learning barriers. Most of what I'm doing involves very basic small features and CSS modifications for appearance, so it should still be relatively fundamental.

啊 是的，到後來我的路綫好像走歪了，因爲我想在這個基礎上添加方便日常使用的功能，并且是基於插件形式的，但是後期就因爲需要繞開這些限制而走到了前端的編寫修改。

好辦法，我會繼續跟進學習這些技巧，參考你提及到的方法進行嘗試，因爲我的語種是中文而且是粵語，加上我的思維比較老舊，所以表達也是比較古代守舊的古文，所以沒有嘗試過這種提問的形式進行開發，我會繼續嘗試的。

我還真是有點擔心我這邊的不斷嘗試而影響了這個開發助手的訓練純净度，簡單來説可能會產生一些代碼污染。

如果有這方面的代碼干擾的話，我會轉過去使用deepseek進行嘗試，畢竟我近期才知道前端的開發不需要考慮太多關於api和後臺的運行結構，只需要提供靶向的前端代碼和一些提示就可以嘗試做出來了，甚至是繞過了hfs3的一些核心，只是讓我減少了不少學習代碼的阻礙，畢竟我做這些很多都是很基礎的一些小功能，還有外觀關於css的修改，所以應該還是比較基礎的。

[rejetto]

i don't think the risk of "pollution" you talk about is real, so don't worry.

I recently learned that frontend development doesn't require overthinking API

this not strictly true. Knowing the specific frontend APIs makes the work easier and better, and sometimes it's even necessary to have something that really works.

[rejetto]

also, since you are not a frontend developer, you don't know when a solution is overcomplicated, even we don't consider HFS' api, but just the browser's api

[hugo529100]

It's good that there's no pollution, because to me, the plugin-helper is like a "witch's cauldron"—constantly throwing various attempts into the "pot" over and over again...

That’s for sure. APIs are definitely necessary, but there are a very few cases where certain functionalities can be "worked around" based on native capabilities. At least, in my experience, many complex features still can’t be achieved, surely because the underlying requirements can’t be bypassed, or perhaps additional API requests need to be made to make them possible.

As you say, I’m not a frontend developer, so naturally, I have no knowledge of the entire underlying architecture. All I can do is probe around the edges through trial and error...

But for me, this process of exploration is also part of learning and fun—even if it’s chaotic and noobish, at least I’ve picked up a few tricks. Besides, some of my weird requirements can still be solved through my own tinkering, which reduces the trouble for big brothers. I only report when there are clear functional gaps or bugs.

I try my best to explore and solve these odd feature requests on my own, because sometimes, after using them for a while, I realize certain functionalities aren’t really that meaningful in the end, and I end up scrapping them and reverting to the original version...

Like recently, I successfully wrote a script to "display remaining disk space" on the login page, but I eventually abandoned it because constantly obsessing over hundreds of GB of free space seemed a bit fool... It’s much more straightforward to just check it directly on the screen.

After all, your software is already close to perfect...

沒有污染那就好，因為plugin-helper對我來說就好像一個“巫婆的大鍋”，不斷反復地往“鍋”里進行各種嘗試……

那是肯定的，API肯定是需要的，只是有很少一部分功能是可以基於原生存在的基礎上進行“繞過”，至少我目前遇到不少複雜的功能還是實現不到的，肯定是因為無法繞過最底層的要求，也可能需要提出額外添加api的請求才能做到。

正如您提到的，我不是前端開發者，那麼可以肯定的是中間的整個架構肯定是不知情的，只能在外圍進行試探性的摸索……

不過對我來說，摸索的過程也是學習和娛樂的一部分吧，雖然很chaos和菜鳥的，至少還是學到一些小技巧。 況且我一些奇怪的要求還是可以通過自己嘗試來解決，這樣減少對您們的煩擾，除了明顯的功能缺點和bug我才會獨立report。

這些奇怪的功能要求還是盡量自己嘗試摸索解決，畢竟有時候用到最後好像又覺得某些功能沒有太大的存在意義，而最後捨棄返回原版……就像我近期成功寫了個“顯示硬盤剩餘空間”的腳本放在登錄頁面內，最後還是捨棄了，因為時刻操心盯住幾百gb的剩餘空間好像有點傻了……直接打開屏幕看看還更直接。

畢竟你的軟件已經是接近完美了……

[rejetto]

did you use the plugin-helper version 2 for your recent projects?

[hugo529100]

already use ,and recent projects is cut-sync-plugin, mean same side can store the cut list with diffrent pages, but stuck in someting unknow bugs.....

[hugo529100]

Or it would be faster for you to implement it. After all, this involves the storage of cookies and APIs, which seems a bit beyond my knowledge, unless I find a breakthrough.

[rejetto]

i don't understand what it is supposed to do

[hugo529100]

in simple： If I'm in a photo album and need to categorize different pictures by moving them to different folders, the original method is very slow. Creating a plugin that can sync browser cookies would allow to open multiple pages (different directories) simultaneously for faster moving and categorization.

An advanced version would involve storing the "cut" list in an account for synchronization, enabling cross-device file management.

Unfortunately, I'm already stuck at the basic functionality—the cut interface disappears after refreshing—let alone the more advanced cross-device access. This is over my current capabilities.