ICDYCT / I Can Do, You Can Too

Preface I want to use the "Fully Hide Messages from Blocked Users" feature of Ayugram . However, I feel that Ayugram has made too many modifications to the original version, and some features clearly go against the original, so I'm worried about getting my account banned. So I want to port this specific feature to Forkgram . Agent-Oriented Development Hermes interfaces with mimo-v2.5-pro The quoted blocks below are the natural language prompts I sent to the Agent Analyze the Ayugram project 分析 https://github.com/AyuGram/AyuGramDesktop 与 源项目的不同之处, 哪些代码实现了过滤用户发言的功能? 聚焦 隐藏被屏蔽用户消息 这一功能, 分析完整的功能实现流程, 注明涉及的源文件及行数 把你分析的内容总结 为 .md 报告 Develop on Forkgram 在 https://github.com/crazypeace/forkgram-tdesktop 新开一个 branch feat-hide-msg-from-blocked-users 新建一个yml文件, 基于这个branch编译 在 feat-hide-msg-from-blocked-users 分支上进行开发 功能: 在群组消息显示时, 屏蔽某个用户消息(Fully Hide Messages from Blocked Users) 参考分析报告 /root/ayugram-filter-analysis.md Agent feedback:  同意你的修改方案, 实施优先级改动一下: 优先级 1 isHidden() - 主入口 ...

前言 我想使用 Ayugram 的屏蔽某个用户消息的功能(Fully Hide Messages from Blocked Users). 但是觉得Ayugram对原版的修改太多, 而且有些功能明显与原版对着干, 担心被封账号. 所以我想把这个特定的功能 移植 到 Forkgram 上面.

Preface Previously, we compiled forkgram/tdesktop https://zelikk.blogspot.com/2026/06/oracle-arm-vps-2cpu-12ram-47hdd-forkgram-tdesktop-aqtinstall-qt.html https://zelikk.blogspot.com/2026/06/2g-x86-vps-forkgram-tdesktop-swap-8g.html So, let's experiment to see if we can use natural language to direct an Agent to help me modify telegram! For example, when the mouse hovers over the emoji button, a popup window appears. I want this popup's height to be half of the current, and the width to be 2/3 of the current. Agent-Oriented Development First, complete the compilation of the forkgram/tdesktop project, omitted. Herems connected to mimo-v2.5-pro Everything in the blockquote below is the natural language I sent to the Agent Run telegram, take a screenshot to see. The Agent installed Xvfb, took a screenshot Open a VNC, let me connect and operate The Agent opened a VNC, I connected, completed the Telegram account login operation. Now I'm logged in, take a look at the interface Th...

前言 之前 我们实现了编译 forkgram/tdesktop https://zelikk.blogspot.com/2026/06/oracle-arm-vps-2cpu-12ram-47hdd-forkgram-tdesktop-aqtinstall-qt.html https://zelikk.blogspot.com/2026/06/2g-x86-vps-forkgram-tdesktop-swap-8g.html 那么, 实验一下能不能用自然语言指挥 Agent 帮我修改 telegram 吧! 比如, 鼠标移到emoji按钮上时, 自动弹窗, 我想让这个弹窗高度是现有的一半, 宽度是现有的2/3.

Preface My VPS configuration 2 vCPU  2GB RAM 120GB Disk Want to compile the https://github.com/forkgram/tdesktop project Agent-Oriented Development Hermes connecting to mimo-v2.5-pro The content in the quote blocks below is the natural language I sent to the Agent 把 https://github.com/forkgram/tdesktop clone到本地, 并完成编译 Agent feedback:  Tried several times, always failed 分析一下你做过了什么， 得到什么结果 Agent feedback:  发现需要 40+ 外部库 (ada, openssl, ffmpeg, qt, webrtc...) 这些库需要预先编译安装到 /usr/local 或 ../Libraries - 这个过程只能全部成功或全部失败吗？ 能不能一个一个编译？ The Agent switched to compiling one by one. Result: when compiling qtgui, memory ran out and compilation was aborted. Later, the agent found a way: install Qt, then use DESKTOP_APP_USE_PACKAGED=ON Qt no longer needs to be compiled. Then, the compilation of tg_owt (WebRTC) was completed. But still, memory ran out when finally generating telegram. Tried a solution: increase swap space Changed to 4GB, still ran out of memory Changed to 8GB, compilation succeeded The co...

前言 我的VPS配置 2 vCPU  2GB RAM 120GB Disk 想编译 https://github.com/forkgram/tdesktop 项目

Preface My Oracle ARM VPS configuration: 2 cores CPU 12G RAM 47G Disk Want to compile the  https://github.com/forkgram/tdesktop project Agent-Oriented Development Hermes connected to mimo-v2.5-pro The content in the quote box below is the natural language I sent to the Agent 把 https://github.com/forkgram/tdesktop clone到本地, 并完成编译 1st Compilation Failed - Reason: Insufficient Disk Space Record Solution Opened a new session, let Hermes analyze the disk space usage, deleted some environments from other projects, and packed and moved some transferable data to other VPSes for temporary storage. Freed up 25G of disk space. 2nd Compilation Failed - Reason: Insufficient Disk Space Solution Opened a new session, let Hermes search the internet for relevant information to see if anyone has encountered the same Telegram Desktop compilation failure. Hermes searched around and found a solution: directly use aqtinstall to install (instead of compiling) QT 6.11.1, then use .so dynamic linking 3rd Com...

前言 我的 Oracle ARM VPS 配置: 2核  CPU 12G 内存 47G 硬盘 想编译  https://github.com/forkgram/tdesktop 项目

Preface We have already set up ShareX to connect to R2's object storage using the S3 API . But the R2 management page doesn't show image previews in a very intuitive way. We can use a worker to manage them Practice Considering that there may be many and large images, it is recommended to enable Cloudflare's Image Transformation feature First, you need to have a domain hosted on Cloudflare . Enable the Image Transformation feature Set a custom domain for R2 Do you remember this page? Enter a subdomain of your domain. For example, r2test.xxxx.xxxx After a while, it will look like this, indicating it has taken effect. Set the CORS policy [   {     "AllowedOrigins": [       "*"     ],     "AllowedMethods": [       "PUT",       "GET",       "HEAD",       "DELETE"     ],     "AllowedHeaders": [       "*"     ]   } ] * Note, to be honest this is a bit blunt here, but it's not a big deal. ...

前言 我们已经让 ShareX 用 S3 API 对接了 R2 的对象存储 . 但 R2 的管理页面并不那么直观地看到图片预览. 我们可以用 worker 来管理

Github https://github.com/crazypeace/Url-Shorten-Worker Transform into a File Storage Save to R2 Object Storage  First, a demo video Apply for a Cloudflare account, skipped. * During the following operations, the colored information , you'd better write it down separately. This way the operation process will be smoother. Create R2 Remember the name of this R2, take r2test as an example Set up public link Remember this R2 public link,  https://pub-*****.r2.dev Set CORS Policy [   {     "AllowedOrigins": [       "*"     ],     "AllowedMethods": [       "PUT",       "GET",       "HEAD",       "DELETE"     ],     "AllowedHeaders": [       "*"     ]   } ] * Note, to be honest, it's a little rough here, but no big deal.  Create R2 API Go back to the R2 overview page, remember this Account ID Click the Manage button Click the "Create User API token" button Select "Object Read ...