unix: annotate unstable *COUNT constants#5121
Conversation
d36f780 to
70cd6ca
Compare
*COUNT consts fitting rust-lang/libc#3131
*COUNT consts fitting rust-lang/libc#3131*COUNT consts fitting #3131
b2e7f52 to
1b7a551
Compare
This comment has been minimized.
This comment has been minimized.
|
CI is failing for reasons unrelated to the changes introduced in this patch. A rerun should do it. |
This patch is a follow up from rust-lang#5121. That PR deprecated the symbols for a stable release, while this patch ensures the symbols are altogether removed. See the accompanying PR for more details.
This patch is a follow up from rust-lang#5121. That PR deprecated the symbols for a stable release, while this patch ensures the symbols are altogether removed. See the accompanying PR for more details.
This patch is a follow up from rust-lang#5121. That PR deprecated the symbols for a stable release, while this patch ensures the symbols are altogether removed. See the accompanying PR for more details.
This patch is a follow up from rust-lang#5121. That PR deprecated the symbols for a stable release, while this patch ensures the symbols are altogether removed. See the accompanying PR for more details.
1b7a551 to
2580f8d
Compare
This comment has been minimized.
This comment has been minimized.
|
CI actually passes. There seems to be an issue with a glob import that is not used, but this has not |
This patch is a follow up from rust-lang#5121. That PR deprecated the symbols for a stable release, while this patch ensures the symbols are altogether removed. See the accompanying PR for more details.
2580f8d to
2ca368a
Compare
This comment has been minimized.
This comment has been minimized.
2ca368a to
209065b
Compare
This comment has been minimized.
This comment has been minimized.
|
For updating doc comments like the others, @rustbot author |
|
Reminder, once the PR becomes ready for a review, use |
209065b to
0debe88
Compare
|
This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed. Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers. |
These caused SemVer issues. They can change in upstream releases. This patch aleviates that. It links to usage advice.
aa9d097 to
07fba21
Compare
*COUNT consts fitting #3131*COUNT constants
Description
This PR adds documentation. It links to usage advice. Some symbols need this. They can change between upstream releases. They are not considered SemVer-breaking.
I have some doubts. They are concerned with the GNU Hurd. The
DLFO_STRUCT_HAS_EH_COUNTconstant was not found. It has not suffered any changes.Sources
Haiku sources showing how the
B_MEDIA_STALE_CHANGE_COUNTis not the type of constant we would want to deprecate. The code upstream itself does not seem to use it as a delimitter value, as can be seen in the second search result of the below reference.DragonFly BSD sources showing how the
CPUCTL_CPUID_COUNTconstant is not one we would want to deprecate, even though it may seem like it could be deprecated (because it appears as a*COUNTsymbol after a bunch of other symbols.) This constant is actually used for literal CPU counts inhttps://github.com/DragonFlyBSD/DragonFlyBSD/blob/master/sys/dev/misc/cpuctl/cpuctl.c#L145.
OpenBSD sources showing how the
KERN_TTYCOUNTis actually used as a counter for TTY devices and not for other constants (just another enum-like "variant".) This does not qualify it for deprecation.OpenBSD sources showing how the
KERN_EVCOUNTis used as an event counter (with use akin to that of the above list item.) This does not qualify it for deprecation.NetBSD sources showing how the
DCCP_OPT_NDP_COUNTconstant is used as just a regular constant (not meant for deprecation.)Linux sources showing how the
TIOCGICOUNTconstant seems to be used not as a counter for other symbols but as (possibly runtime) interrumpt counter (not meant to be deprecated.)FreeBSD sources showing how the
IFMIB_IFCOUNTconstant is promptly liable to upstream changes as it denotes the number of configured interfaces.XNU kernel sources showing how the
RTV_HOPCOUNTis not the type of constant we would want to deprecate. It seems to be more of an enum "variant" among other similar constants.Illumos sources showing that the
IPV6_MINHOPCOUNTis not dependent of other source code chagnes (and thus does not require deprecation.)L4RE sources showing how the
SIOCGIFCOUNTconstant is more of an enum "variant" among other values, and thus not a targe of these deprecation efforts.Illumos sources showing how the
LGRP_RSRC_COUNTconstant is meant to count the number of resource types, which for the time being are the two constants below it.Linux sources mentioning that the constants declared prior to the
NFNL_SUBSYS_COUNTsymbol could have been in an enumeration where the last value is the current "limit."Android sources replicating similar behavior as in Linux.
Checklist
libc-test/semverhave been updated*LASTor*MAXare included (see #3131)cd libc-test && cargo test --target mytarget); especially relevant for platforms that may not be checked in CI@rustbot label +stable-nominated