solaris build environment should include libsendfile/liblgrp

[jclulow]

As of version 0.2.120 of the libc crate, the solaris target now requires
some additional libraries to be present in the sysroot. Note that the
solaris target doesn't really build against files from Solaris, but
rather against some files from DilOS (a platform similar to both Solaris
and illumos). Pull in the extra libraries and their compilation links
from that apt repository.

This aims to assist with #94052.

[rust-highfive]

r? @Mark-Simulacrum

(rust-highfive has picked a reviewer for you, use r? to override)

[jclulow]

Note that I've looked in the apt repository I believe this docker image uses, and found the files I would expect to see, via:

#!/bin/bashset -o errexit
set -o pipefail
REPO=https://apt.dilos.org/dilos
A=dilos2
B=main
printf'loading package manifest...\n'
pkgs=$(curl -sSf -L "$REPO/dists/$A/$B/binary-solaris-i386/Packages.gz"| gunzip)functionfile_for {
awk -v "p=$1"' $1 == "Package:" && $2 == p { q = 1; } /^$/ { q = 0; } q && $1 == "Filename:" { print $2; }'<<<"$pkgs"
}
files=(
$(file_for libsendfile)$(file_for libsendfile-dev)$(file_for liblgrp)$(file_for liblgrp-dev)
)
td="$(mktemp -d)"forfin"${files[@]}";do
b=$(basename "$f")printf'package: %s\n'"$b"
curl -sSfL -o "$td/$b""$REPO/$f"
7z -so x "$td/$b"2>/dev/null | tar tvf - | sed 's/^/ /'printf'\n'done
rm -rf "$td"

This gives output:

loading package manifest...
package: libsendfile_2.0.2.48_solaris-i386.deb
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./lib/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./lib/amd64/
-rwxr-xr-x 0/2 13904 Oct 30 10:20 2019 ./lib/amd64/libsendfile.so.1
-rwxr-xr-x 0/2 8556 Oct 30 10:20 2019 ./lib/libsendfile.so.1
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/lib/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/lib/amd64/
-rwxrwxrwx 0/0 35 Oct 30 10:20 2019 ./usr/lib/amd64/libsendfile.so.1 symbolic link to ../../../lib/amd64/libsendfile.so.1
-rwxrwxrwx 0/0 26 Oct 30 10:20 2019 ./usr/lib/libsendfile.so.1 symbolic link to ../../lib/libsendfile.so.1
package: libsendfile-dev_2.0.2.48_solaris-i386.deb
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./lib/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./lib/amd64/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/lib/
-rwxr-xr-x 0/0 0 Oct 30 10:20 2019 ./usr/lib/amd64/
-rwxrwxrwx 0/0 16 Oct 30 10:20 2019 ./lib/amd64/libsendfile.so symbolic link to libsendfile.so.1
-rwxrwxrwx 0/0 16 Oct 30 10:20 2019 ./lib/libsendfile.so symbolic link to libsendfile.so.1
-rwxrwxrwx 0/0 35 Oct 30 10:20 2019 ./usr/lib/amd64/libsendfile.so symbolic link to ../../../lib/amd64/libsendfile.so.1
-rwxrwxrwx 0/0 26 Oct 30 10:20 2019 ./usr/lib/libsendfile.so symbolic link to ../../lib/libsendfile.so.1
package: liblgrp_2.0.2.48_solaris-i386.deb
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/lib/
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/lib/amd64/
-rwxr-xr-x 0/2 20984 Oct 30 10:19 2019 ./usr/lib/amd64/liblgrp.so.1
-rwxr-xr-x 0/2 19468 Oct 30 10:19 2019 ./usr/lib/liblgrp.so.1
package: liblgrp-dev_2.0.2.48_solaris-i386.deb
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/lib/
-rwxr-xr-x 0/0 0 Oct 30 10:19 2019 ./usr/lib/amd64/
-rwxrwxrwx 0/0 12 Oct 30 10:19 2019 ./usr/lib/amd64/liblgrp.so symbolic link to liblgrp.so.1
-rwxrwxrwx 0/0 14 Oct 30 10:19 2019 ./usr/lib/liblgrp.so symbolic link to ./liblgrp.so.1

Will the CI pipeline produce the new docker image (dist-various-2) and use that to build the solaris target for this change, or do we need to do something extra to test that?

[Mark-Simulacrum]

@bors r+ rollup=iffy

Will the CI pipeline produce the new docker image (dist-various-2) and use that to build the solaris target for this change, or do we need to do something extra to test that?

Yep, this'll automatically rebuild the docker image and use that going forward.

[bors]

📌 Commit 0dcb175 has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit 0dcb175 with merge 442ea85c0cd3580482c8770a136fcc5cf7feb1e4...

[bors]

💥 Test timed out

[rust-log-analyzer]

A job failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

[Mark-Simulacrum]

@bors retry dist-i586-gnu-i586-i686-musl timed out in ui tests

[bors]

⌛ Testing commit 0dcb175 with merge 311b962c39fc7eb88e6d977272d45253ef4489b3...

[bors]

💔 Test failed - checks-actions

[tmiasko]

@bors retry x86_64-apple-2: The hosted runner: GitHub Actions 47 lost communication with the server.

[rust-log-analyzer]

A job failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

[bors]

⌛ Testing commit 0dcb175 with merge 183090f...

[bors]

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing 183090f to master...

[rust-timer]

Finished benchmarking commit (183090f): comparison url.

Summary: This benchmark run did not return any relevant results.

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

@rustbot label: -perf-regression