HN user

vadiml

194 karma
Posts10
Comments87
View on HN

I've been working on extending CDC File Transfer to support Linux→Linux and macOS builds. This is an open-source project born from Google Stadia – fast file sync tools (cdc_rsync) and a streaming filesystem (cdc_stream) that use Content Defined Chunking (FastCDC) to transfer only the changed portions of files.

Originally Windows→Linux only, designed for game developers to iterate on 40+ GB builds. The CDC-based diffing is up to 30x faster than rsync (1500 MB/s vs 50 MB/s).

What I added (40 commits):

Linux Client Support:

Ported cdc_rsync client to run on Linux (was Windows-only) Added process_linux.cc (~700 lines): Process management using fork/execve Added port_manager_linux.cc: Port detection using POSIX shared memory Now you can sync Linux→Linux, not just Windows→Linux macOS Support:

Full build support for macOS (ARM64 and x86_64) Fixed 17+ platform-specific issues: /tmp, /var, /etc are symlinks to /private/* – breaks path comparisons POSIX shared memory rounds ftruncate() up to 16KB, rejects shrinking BSD fts_* API reports Bazel runfiles symlinks as FTS_SL not FTS_F strerror(0) returns different strings per platform stat64 doesn't exist on macOS (just use stat) All 41 tests now pass on macOS Infrastructure:

GitHub Actions CI for Linux and macOS (self-hosted runner for macOS) Added comprehensive CLAUDE.md for AI-assisted development The project uses Bazel with gRPC, protobuf, and FUSE. Dependencies via git submodules or http_archive.

GitHub (original): https://github.com/google/cdc-file-transfer My fork: https://github.com/Vadiml1024/cdc-file-transfer

I've been using Claude Code (Anthropic's AI coding assistant) for most of this work. Happy to discuss the experience or the technical challenges.

I don't understand really why these patents are granted. Software is math and math formulas are not patentable. The patent (and copyrights) system today is actually subverted by greedy players and instead of promoting progress and sciences is serving to enrich the middlemen by artificially SLOWING the progress.

I'm really baffled by all this discussion on copyrights in the age of AI. The Copilot does not 'steal' or and reproduce our code - it simply LEARNS from it as a human coder would learn from it. IMHO desire to prevent learning from your open source code seems kind of irrational and antithetical to open source ideas.

Well, a strategic goal to let Europeans suffer during winter in order to shatter our resolve to support Ukraine, while avoiding contractual penalties for unilateral cessation of gaz delivery, can explain the benefit of exploding the pipeline. This can explain also the fact that the fourth pipe was left intact - as a potential carrot to Europe... I do not pretend that the above is truth, but it seams to be a plausible explanation.

The classic representative democracy suffers from a problem that it promotes a-holes to power. Because they have no scrupules to lie and cheat they have to get and hold power. And this explain the fact that the percentage of a-holes in political elites is way bigger then in rest of populations. (btw this specific problem could be easily solved by replacing elections by some form of lottery) What is relatively unknown about Swiss democracy is that on the referendums on national level are extremely rare, but a possibility of a referendum looms like a Damocles sword over political elites and prevent them from abusing their legislative power.

If i read the code published on github and later i do not infringe on any license. If later i'm hired to implement some function chances are that i'll produce code which is similar to what i've learned from github. Or if somebody asks me on how to implement some function and i can respond: Hey that and that project on github already doing it, just use it, i'm not infringing on any license. Now, replace "I" in the above paragraph by "Copilot". It seems that Copilot is not infringing on anything...

If we REALLY want to reform DMCA there is relatevily simple way to do it: Organize a massive simultaneous DMCA takedowns of congressmen web pages... The reform will come in days.