Hi Thorsten, On Sun, Oct 27, 2024 at 04:30:43AM GMT, Thorsten Glaser wrote: > Hi Alejandro, > > >On Sun, Oct 27, 2024 at 01:40:23AM GMT, mirabilos wrote: > >> Not too sure what the root context of this thread is, but in BSD land > > > >The root context is that > > > >- _Lengthof was added to C2y in the Minnesota meeting. It was proposed > > where/how would this be used for? Also, do you have the nXXXX.pdf link > handy? I don’t actively follow C other than what thephd posts.
Yes; there are four nXXXX papers. See below, plus the history of why four papers. <https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2529.pdf> That first paper wasn't mine. I didn't know about the existence of that paper until a few months after that paper I sent a patch to glibc adding the usual macro, and I think Joseph told me about the existence of the paper. I tried to contact the author of the paper, to ask him to change the name, since I thought it would be dangerous as I said here. I couldn't reach him in many years, so at some point I decided I would author a revision of the paper. I believe stuff should go into the standard after it exists in an implementation, so I first prepared a patch for GCC: <https://inbox.sourceware.org/gcc-patches/20240728141547.302478-1-...@kernel.org/T/#t> Joseph Myers asked me to present a paper to WG14 to see what they think before he would finish the review of my patch set, and so I prepared the paper, which was submitted as n3313. By the time I wrote this GCC patch, I had relaxed my feelings about the dangers of the name _Lengthof, so I didn't think too much about it and followed the original proposal, also since Martin Uecker --who helped me develop the GCC patch set-- had a preference for that name. <https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3313.pdf> While iterating on the patches, at some point I reminded myself of actual bugs I had fixed in the recent past in which precisely this naming confusion was involved. I decided to rename the operator to _Nelementsof, to avoid having the term length, and because the standard uses quite often the terms "number of elements in an array" and "one past the last element of an array object", so it just felt the natural term to use. Since the name of the operator was too long, and some people complained about that, I proposed an alternative name _Neltsof. I then submitted n3325. <https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3325.pdf> The paper was discussed in the Minnesota meeting of WG14, where I was invited to defend my paper. The semantics were agreed upon, but there wasn't strong consensus on the name. In the end, _Nelementsof was rejected in a poll with majority but no consensus (only slightly more YESs than NOs). Then we voted to see what name should be used, and the name _Lengthof was voted. I had to present n3369. I wasn't the only voice concerned about _Lengthof. The name _Countof seemed to have also some traction in the people that didn't want _Lengthof, so it's the name I'm considering now, but either something derived from _Nelementsof or _Countof or anything not "length"y would be good IMO. BTW, during the discussions, _Nelemsof seemed to be strongly preferred over _Neltsof. I don't care at all about those trivial details, though. <https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3369.pdf> And this paper was voted in for C2y, so here we are. I'm still trying to overrule that paper by implementing it with a different name. Since I have the advantage that I already have the patches working, I have months before anybody implements _Lengthof in another compiler and sets the name in stone. It would be great if authors of other compilers would agree on the dangerousness of this name and pushed for a different name. > >- I'm sending a patch to GCC proposing __countof__, since overloading > > length for both string length and array number of elements is (IMO) > > going to promote off-by-one bugs in string-handling code (and I've > > Definitely! The string length is one less than the amount of array > elements it has (works for char and wide strings). Thanks! > >- _Nelementsof > >- _Nelemsof > > If you want to shorten elements, elts is probably seen more > often than elems, at least in my experience. But this is very > low-grade bikeshedding, so just me pointing it out but putting > not much weight behind it. That was my original proposal, but everybody I've talked so far told me they preferred _Nelemsof. I don't care about it at all, so whatever. > bye, > //mirabilos > PS: Got any contacts in the OpenBSD and NetBSD worlds that can > add input? They invest a lot into good C code as well, in > the OpenBSD case rather heavily (if opinionated). I don't. If you do, please let them know. Theo doesn't seem to have a good opinion of me, but maybe he could help. Have a lovely day! Alex -- <https://www.alejandro-colomar.es/>
signature.asc
Description: PGP signature
