There you go: github.com/dfaranha/ope...
GitHub - dfaranha/openabe: The OpenABE library - open source cryptographic library with attribute-based encryption implementations in C/C++
The OpenABE library - open source cryptographic library with attribute-based encryption implementations in C/C++ - dfaranha/openabe
github.com
July 1, 2025 at 10:13 PM
There you go: github.com/dfaranha/ope...
Another update: OpenABE provides an option in the makefiles to avoid RELIC completely (due to someting with mingw, I believe), and instead only rely on a modified version of OpenSSL 1.1.1. I was able to extend that to my build, and long story short: It works.
My TODO: Make it work with modern RELIC
My TODO: Make it work with modern RELIC
June 20, 2025 at 11:07 PM
Another update: OpenABE provides an option in the makefiles to avoid RELIC completely (due to someting with mingw, I believe), and instead only rely on a modified version of OpenSSL 1.1.1. I was able to extend that to my build, and long story short: It works.
My TODO: Make it work with modern RELIC
My TODO: Make it work with modern RELIC
Sungkyunkwan University researchers engineered AI-derived Deam-P32 into OpenABE1.1/1.2 by grafting a missing C-terminal domain and 3 key mutations, matching ABE8e's 41.7% efficiency while slashing ATC bystander edits up to 12-fold. 🧬🤖 #BaseEditing #CRISPR #OpenABE #NAR2026 #AIProteinDesign #Biotech
July 30, 2026 at 5:44 AM
Sungkyunkwan University researchers engineered AI-derived Deam-P32 into OpenABE1.1/1.2 by grafting a missing C-terminal domain and 3 key mutations, matching ABE8e's 41.7% efficiency while slashing ATC bystander edits up to 12-fold. 🧬🤖 #BaseEditing #CRISPR #OpenABE #NAR2026 #AIProteinDesign #Biotech
Let me push the changes I have (plus your patches) to an OpenABE fork, that's easier than trying to sync over here. :)
June 26, 2025 at 8:35 PM
Let me push the changes I have (plus your patches) to an OpenABE fork, that's easier than trying to sync over here. :)
Possibly last for now: I've made OpenABE compile with RELIC 0.7.0. It took a bit of renaming here and there. However, it still fails on A64, and it seems to pertain to how OpenABE uses randomness with RELIC.
CLI output in next post.
It works with OpenSSL, so we can work from there instead.
CLI output in next post.
It works with OpenSSL, so we can work from there instead.
June 22, 2025 at 9:00 PM
Possibly last for now: I've made OpenABE compile with RELIC 0.7.0. It took a bit of renaming here and there. However, it still fails on A64, and it seems to pertain to how OpenABE uses randomness with RELIC.
CLI output in next post.
It works with OpenSSL, so we can work from there instead.
CLI output in next post.
It works with OpenSSL, so we can work from there instead.
I turned on debugging with RAND=CALL and saw that OpenABE was failing the health test in one of the calls because its RNG outputs 32 zeroes (not sure why). The problem with test_pk solved itself with WSIZE=64.
June 26, 2025 at 7:17 AM
I turned on debugging with RAND=CALL and saw that OpenABE was failing the health test in one of the calls because its RNG outputs 32 zeroes (not sure why). The problem with test_pk solved itself with WSIZE=64.
Turning CHECK=on made most of the binaries in the examples folder to work, except test_cp that still gives RNG and decryption errors. I went down the RNG error and could see the OpenABE RNG outputting only zeroes, but I could not find the reason. Looks like error-handling and RNG are broken.
June 26, 2025 at 12:59 AM
Turning CHECK=on made most of the binaries in the examples folder to work, except test_cp that still gives RNG and decryption errors. I went down the RNG error and could see the OpenABE RNG outputting only zeroes, but I could not find the reason. Looks like error-handling and RNG are broken.
I did get RELIC to work in isolation on RPi. My current challenge is to get OpenABE with latest RELIC (compiled with default RAND) to work correctly, even on x64. Sorry if I have been unclear about that.
Setup: dpaste.org/1dyhD
Patch (current diff): dpaste.org/fGWGa#L12
Result: dpaste.org/oDo8B
Setup: dpaste.org/1dyhD
Patch (current diff): dpaste.org/fGWGa#L12
Result: dpaste.org/oDo8B
dpaste/1dyhD (Bash)
dpaste.org
June 25, 2025 at 6:40 PM
I did get RELIC to work in isolation on RPi. My current challenge is to get OpenABE with latest RELIC (compiled with default RAND) to work correctly, even on x64. Sorry if I have been unclear about that.
Setup: dpaste.org/1dyhD
Patch (current diff): dpaste.org/fGWGa#L12
Result: dpaste.org/oDo8B
Setup: dpaste.org/1dyhD
Patch (current diff): dpaste.org/fGWGa#L12
Result: dpaste.org/oDo8B
- OpenABE uses CMP_xx extensively, so it's not trivial to just use a newer version of RELIC.
- Given this, we just assume RELIC isn't the culprit for now (despite the text in readme) and see if we can import keys, cts, pts from desktop to RPi to identify where it fails and dig from there.
(3/3)
- Given this, we just assume RELIC isn't the culprit for now (despite the text in readme) and see if we can import keys, cts, pts from desktop to RPi to identify where it fails and dig from there.
(3/3)
June 20, 2025 at 12:21 PM
- OpenABE uses CMP_xx extensively, so it's not trivial to just use a newer version of RELIC.
- Given this, we just assume RELIC isn't the culprit for now (despite the text in readme) and see if we can import keys, cts, pts from desktop to RPi to identify where it fails and dig from there.
(3/3)
- Given this, we just assume RELIC isn't the culprit for now (despite the text in readme) and see if we can import keys, cts, pts from desktop to RPi to identify where it fails and dig from there.
(3/3)
Thanks, I'll check that out tonight!
How did you inspect the OpenABE RNG?
Also, how does test_pk work for you? With old RELIC, we got a segmentation fault, modified OpenSSL made it work, and new RELIC produced an infinitely (?) repeating error message for me. Looks like a canary in the mine.
How did you inspect the OpenABE RNG?
Also, how does test_pk work for you? With old RELIC, we got a segmentation fault, modified OpenSSL made it work, and new RELIC produced an infinitely (?) repeating error message for me. Looks like a canary in the mine.
June 26, 2025 at 6:58 AM
Thanks, I'll check that out tonight!
How did you inspect the OpenABE RNG?
Also, how does test_pk work for you? With old RELIC, we got a segmentation fault, modified OpenSSL made it work, and new RELIC produced an infinitely (?) repeating error message for me. Looks like a canary in the mine.
How did you inspect the OpenABE RNG?
Also, how does test_pk work for you? With old RELIC, we got a segmentation fault, modified OpenSSL made it work, and new RELIC produced an infinitely (?) repeating error message for me. Looks like a canary in the mine.
That's generous of you.
BTW, I had to change relic_core.h:484 int to size_t to match the signature of rand_stub and callback.
I'll also see if I can modify OpenABE to work with the default RAND.
BTW, I had to change relic_core.h:484 int to size_t to match the signature of rand_stub and callback.
I'll also see if I can modify OpenABE to work with the default RAND.
June 23, 2025 at 2:44 PM
That's generous of you.
BTW, I had to change relic_core.h:484 int to size_t to match the signature of rand_stub and callback.
I'll also see if I can modify OpenABE to work with the default RAND.
BTW, I had to change relic_core.h:484 int to size_t to match the signature of rand_stub and callback.
I'll also see if I can modify OpenABE to work with the default RAND.
@matthewdgreen.bsky.social Good afternoon! We're trying to make OpenABE run on RPi 3 and 5, but the compiler complains about invalid conversion from const void* to void* in zpkey.cpp (61 IIRC). -fpermissive lets it compile, but decryption fails in the examples. Do you have any insight we could try?
June 19, 2025 at 8:12 PM
@matthewdgreen.bsky.social Good afternoon! We're trying to make OpenABE run on RPi 3 and 5, but the compiler complains about invalid conversion from const void* to void* in zpkey.cpp (61 IIRC). -fpermissive lets it compile, but decryption fails in the examples. Do you have any insight we could try?
明日はきっとOpenABEが動かねえ動かねえってたぶん言ってる
December 4, 2024 at 4:19 AM
明日はきっとOpenABEが動かねえ動かねえってたぶん言ってる