[Proposal] Standardise the representation of byte sequence as a sequence of unsigned 8 bit integers #2099
Conversation
ConnorGray
changed the title
There was a problem hiding this comment.
Thank you very much for going through the proposal process for this! We're going to discuss this internally within the team this coming week and we'll circle back here. Just wanted to give you a quick update here to let you know we haven't forgotten or missed this.
In general, I'm +1 for this — we're just trying to understand some of the existing history for the rationale why Int8 was chosen for String to begin with to help guide our decision on how we want to move forward.
JoeLoser
added
needs-discussion
proposals/byte-as-uint8.md
Outdated
| # Motivation | ||
|
|
||
| Logically a byte is an integer value between `0` and `255`. Lots of algorithms make use of arithmetics ground by this assumption. | ||
| A signed 8 bit integer on the contrary represents values between `-128` and `127`. This introduces very settle bugs, when an algorithm written for unsigned 8 bit integer is used on a signed 8 bit integer. |
There was a problem hiding this comment.
Suggestion s/settle/subtle
There was a problem hiding this comment.
@mzaks after this small typo is fixed, happy to land this 👍
|
Thanks so much for going through the proposal process! We as a team really appreciate the thought and motivation presented, and are happy to accept this proposal! 🎉 🎉 🎉 Internally, we went back and forth whether to add an Regarding the alias part, since an alias is a weak type, it's only for readability sake/to denote we're conceptually working with "bytes" since you can't overload functions/APIs on weak types, and we don't yet have a strong Congrats again on the first standard library proposal to be accepted! Now that we have a process internally for discussing these, we expect to be able to triage and discuss these in a much more timely fashion going forward. Thanks for your patience with this one. |
|
I made #2317 to track the progress of the implementation |
…gned 8 bit integers Signed-off-by: Maxim Zaks <maxim.zaks@gmail.com>
mzaks
force-pushed
the
proposal/byte-as-uint8
branch
from
7adaf40
to
cd404f0
Compare
rephrase as suggested Signed-off-by: Maxim Zaks <maxim.zaks@gmail.com>
|
Hm, why is the first commit not verified? @JoeLoser do you have an idea? |
Looks like one was signed with your GPG key and one was not. The GPG signing isn't required, so I went ahead and merged this. We only require a sign-off using |
|
FYI I realized this was targeting the main branch and not nightly, but snuck in since it wasn't rebased where we have CI checking it targets the nightly branch. I'm happy to open a PR for this to target the nightly branch as well later today unless you beat me to it. The main branch will automatically get synced each time we make a new MAX release and it'll pick up all the things in the nightly branch effectively that made it into the stable MAX release. |
…s a sequence of unsigned 8 bit integers (modularml#2099) Add a proposal to standardize the representation of byte sequence as a sequence of unsigned 8 bit integers. Signed-off-by: Maxim Zaks <maxim.zaks@gmail.com> (cherry picked from commit 4688996)
See the byte-as-uint8.md file