Revert "fix truncate_unuesed_mem dividing by 7 instead of 8"

[Forpee]

Reverts #49

[ecioppettini]

was there a reason this had to be reverted? was the code actually wrong?

[Forpee]

was there a reason this had to be reverted? was the code actually wrong?

@ecioppettini Sorry for reverting — the code looked good, but a zkEngine test failed after the change. I’m going to add some CI tests first before looking at your other PRs. Again, thanks for those

[ecioppettini]

was there a reason this had to be reverted? was the code actually wrong?

@ecioppettini Sorry for reverting — the code looked good, but a zkEngine test failed after the change. I’m going to add some CI tests first before looking at your other PRs. Again, thanks for those

No problem. I did miss the fact that a test was failing. If it helps, I think the reason it failed is because of #47. I think the highest address used in that case is close enough to the last host call witness, that the division by 7 instead of 8 is enough to cover for it.

I think probably when I tested I just had both patches at the same time, but then did split it into multiple patches and didn't consider that.

[Forpee]

was there a reason this had to be reverted? was the code actually wrong?

@ecioppettini Sorry for reverting — the code looked good, but a zkEngine test failed after the change. I’m going to add some CI tests first before looking at your other PRs. Again, thanks for those

No problem. I did miss the fact that a test was failing. If it helps, I think the reason it failed is because of #47. I think the highest address used in that case is close enough to the last host call witness, that the division by 7 instead of 8 is enough to cover for it.

I think probably when I tested I just had both patches at the same time, but then did split it into multiple patches and didn't consider that.

@ecioppettini makes sense, you can open another PR for it. Thanks again!

[Forpee]

@ecioppettini Are you going to re-open this?