aboutsummaryrefslogtreecommitdiff
path: root/tests
AgeCommit message (Collapse)Author
2019-08-09find: fix dangling symlink behavior.Elliott Hughes
ENOENT is ignored, but other errors are reported.
2019-08-05du: disable the tests if we have SELinux.Elliott Hughes
ext stores extended attributes in a way that makes all the numbers in the test incorrect. For now, just disable the test in that case.
2019-07-27losetup.test: bail out if loopback devices are already in use.Elliott Hughes
There's probably a way to filter out the existing ones, but it's likely to be pretty painful and isn't immediately necessary. (I hit this case on Android cloud x86 devices.)
2019-07-27modinfo.test: bail out early if there are no modules.Elliott Hughes
We need /proc/modules *and* a directory of modules.
2019-07-27ifconfig.test: actually skip tests if dummy0 up fails.Elliott Hughes
Rather than just claim we're doing so. Android cloud emulators don't have dummy0 like regular devices do.
2019-07-26truncate.test: work around ext extended attributes space usage.Elliott Hughes
Weaken the test slightly so that the extra 4KiB used for SELinux is ignored. Unfortunately all the du tests suffer the same problem, and I'm less sure what to do there. Mounting a read-only file system without the extended attributes might be the best idea? But since this test modifies the file system...
2019-07-26hostname: fix tests for non-root on Android.Elliott Hughes
On Android you can't even get the hostname unless you're root, to avoid device fingerprinting/tracking. Also modernize by using skipnot, which fixes the incorrect test name in the SKIP: output.
2019-07-25killall: implement -w.Elliott Hughes
This isn't currently used in AOSP, but it is used in some of the other codebases I can easily search, and it's a better fix for test flakiness than inserting a sleep.
2019-07-25killall.test: fix flakiness.Elliott Hughes
And when I say "fix", I mean "hack around in the worst possible way". Android+mksh is often slow enough that we can do the pgrep before the killed process has actually exited. (I'm seeing roughly 100% failure rate on cloud x86 emulators, and about 75% on a real Pixel 2.)
2019-07-24tests: fix for empty /etc/passwd or /etc/group.Elliott Hughes
It turns out some Android devices have an empty /etc/passwd and/or /etc/group, which was defeating the previous workaround. Switch to testing the intention more directly: we'll try the file in /etc, and if that didn't work, we'll assume we need a workaround.
2019-07-22test.tar: fix tar tests on Android.Elliott Hughes
There's no /etc/group on Android. Loop devices are in different places, but that test doesn't seem to offer anything beyond the block/char special device tests later, so remove that rather than fix it. Fix the block/char special tests (the expectations were missing, but they're never normally run on the host because the mknod fails). Likewise add the missing expectation to the "ownership" test, which isn't normally run on the host because the chown fails. The "sparse without overflow" test was broken because of the mksh/bash disagreement about whether in `VAR=val shell_function` that's a temporary assignment to VAR or one that outlives the command. Switch SUM to being a function like LST before it. Also rewrite the TARHD stuff because Android (a) doesn't have hd and (b) doesn't have process substitution in its shell. (I wish I'd noticed TARHD earlier --- I'd been manually sending the tar files to xxd for debugging instead!) With this patch, all the tar tests now pass for me on both the device and host.
2019-07-16Fix unaligned access, tweak test suite.Rob Landley
2019-07-16grep: fix two bugs found by hwasan.Elliott Hughes
The first bug appeared as a memory overwrite, but was actually visible without hwasan: basically any `grep -F` that let to multiple matches on the same line was broken. The second bug was another memory overwrite, visible when I ran the existing grep tests. Bug: http://b/137573082
2019-07-12diff: implement --strip-trailing-cr.Elliott Hughes
Used by some ART tests and also some LLVM tests. (The motivating example is the latter, but I noticed the former when looking for other users.) Bug: http://b/137298656
2019-07-12pidof: fix default behavior, add -x.Elliott Hughes
Before this patch, we're effectively doing `pidof -x` all the time. This patch changes names_to_pid() to allow us to say whether or not we want to include scripts, and adjusts the callers appropriately. Also add tests for `pidof` versus `pidof -x` which pass after this patch, without regressing the existing killall tests.
2019-07-12grep: add -R as well as -r.Elliott Hughes
On BSD these are actually the same, and there's a -S that you need in addition. So strictly this is a behavior change for Android (which is going from BSD grep to toybox grep), but it's a behavior preserving change for the AOSP build (which is going from GNU grep to toybox grep), and the latter actually has a checked-in use of -R where the former doesn't.
2019-07-11env.test: fix new test for toybox echo.Elliott Hughes
The test relied on the non-builtins not being toybox (or another multicall binary). A multicall binary will actually look at the symlinked name (in this case "true") and do what it says on the tin. Use a tiny shell script instead.
2019-07-10dd: iflags, oflags, fix ^C, fix the fundamental loop.Elliott Hughes
Investigating why the toybox tar tests fail on Android with toybox dd, I realized I was rewriting a part of dd I'd rewritten before! This is a re-send of my 2019-02-22 patch, rebased against the current ToT... This patch was originally motivated because after suggesting to the author of https://stackoverflow.com/questions/17157820/access-vdsolinux/54797221#54797221 that he could tell dd to work in bytes rather than blocks, I realized that our dd doesn't actually support that. But the rewrite of the main loop is necessary to fix the incorrect output from the dd calls in the tar test. Without this patch, `yes | dd bs=65536 count=1 > fweep` basically gives random output, based on how many bytes the pipe feels like giving you in your first read. (As far as I know, dd *without* bs= was fine, but I can't guarantee that that's true, just that I haven't seen it fail.) Also switch to TAGGED_ARRAY and comma_* for conv rather than add two more copies of an undesired idiom. It turned out -- contrary to the belief of cp(1) -- that comma_scan isn't suitable for this because of its magic handling of "no" prefixes. (It's actually harmless in cp because none of the --preserve options begin with "no", but some dd options do.) To this end, comma_remove is a less-magic comma_scan. I've also changed an `if` to a `while` because other implementations allow things like `--preserve=mode,mode` or `conv=sync,sync`. (If we decide this is a bug rather than a feature, we should at least fix the error message to be clear that we're rejecting the *duplication*, not the option itself.) I've also fixed the ^C behavior by simply adding a direct SIGINT handler rather than trying to be clever inside the read loop (which is why we weren't handling the SIGINT until the read returned). I've also removed `strstarteq` and just added the '=' to each literal when calling regular `strstart`. Plus basic tests.
2019-07-09dd: fix seek= on stdout.Elliott Hughes
2019-07-06killall: better handling of long names.Elliott Hughes
Change names_to_pid() so that we can actually match shell scripts with long names (the code to get the shell script's name was correct, but there was an extra test preventing us from actually comparing it to the sought name). In kill.c itself, remove a dead test for -l and switch to the FLAG() macro. Also extend the tests to explicitly cover long and short names.
2019-07-06ifconfig.test: disable the pointopoint tests.Elliott Hughes
These tests don't work for me as root on either my Debian desktop or my Android devices. The original mail thread implies that they were as close as the original comitter could get to _something_ that seemed to work, even if the kernel doesn't seem to bother with this: http://lists.landley.net/pipermail/toybox-landley.net/2014-November/003795.html toybox ifconfig *is* missing support for `-pointtopoint` and `pointopoint` (without an addess), similar for `broadcast` according to the man page. But since we don't appear to have a way to test this (other than looking at strace output!) I'm leaning towards YAGNI anyway...
2019-07-06Add comment and test.Rob Landley
2019-07-03nl: nl -vEric Molitor
Implement nl -v supporting negative and zero starting values.
2019-06-28modinfo: various small fixes.Elliott Hughes
Improve Android support (see code comments). Remove unnecessary fixed-length limits. Show error if module not found (plus test). Expand help text. Switch to FLAG macro. Stop hard-coding module assumptions in the tests.
2019-06-28split.test: don't rely on bash process substitution.Elliott Hughes
mksh doesn't support process substitution, so use an explicit temporary file.
2019-06-28file, stat: various small improvements.Elliott Hughes
file now shows the target of a symbolic link and calls out broken symbolic links. file now shows the device type for block/character special files. file now shows specific reason when it can't open. stat now includes the device type, plus a little more space between the number of blocks and the human-readable file type. Adjusted tests accordingly, which actually makes more of them pass on the host as a convenient side-effect, but I actually made these changes because I've been finding the desktop file and stat output more convenient in these cases.
2019-06-27touch.test: add missing `TZ=utc`s.Elliott Hughes
I don't know why we fixed the other tests but not this one. Before this patch it's failing with/without TEST_HOST. After this patch, it passes both.
2019-06-26losetup: fix Android.Elliott Hughes
Use /dev/block/loop* more uniformly, and teach the tests which to expect.
2019-06-26Fix cp.test to pass even if you're root or have a restrictive umask.Elliott Hughes
2019-06-25killall.test: don't assume the shell is /bin/sh.Elliott Hughes
The killall test still fails with toybox. Seems like right now killall and pidof share name to pid logic, but playing about with the desktop killall (which does pass this test), it seems like killall actually behaves more like pgrep than pidof (which does seem to be this limited on the desktop too)...
2019-06-25Make chgrp and chown tests work on Android.Elliott Hughes
There's no /etc/passwd and /etc/group, but there are enough users and groups that we can test with. ("bin" and "daemon" were added for LTP; "shell" is what you get with `adb shell`.)
2019-06-25blkid: adapt tests for util-linux blkid.Elliott Hughes
util-linux's blkid doesn't support reading from stdin, so move that to being a special toyonly test rather than the default. toybox blkid differs from util-linux in that it doesn't seem to find the LABEL for the checked-in example f2fs file system (the offset is wrong, but also f2fs uses LE16 strings), nor does it support the SEC_TYPE that util-linux blkid shows for several of the test file systems.
2019-06-25losetup: minor fixes.Elliott Hughes
Fix `losetup -f` to not fail with an error. Add the missing \n for `losetup -f --show FILE`. Use decimal for the device number, like the desktop losetup. Switch to the FLAG macro. Make the tests runnable as tests, and expand coverage a bit. With this patch, the tests pass both with and without TEST_HOST on the desktop. Note though that this patch is part of fixing some real-life losetup issues, not part of the "test cleanup" I'm also looking at. losetup is low down that list!
2019-06-25tar: implement --mode.Elliott Hughes
Used by build/make/tools/mktarball.sh in AOSP. (Which is why today's switch to toybox tar got reverted.)
2019-06-22Don't test shell builtin for kill.Rob Landley
2019-06-22Fix mv with trailing slash in source.Denys Nykula
Press tab, have bash complete dir name with a slash, notice musl rename() dislikes that. Replace trailing slash in the cp loop with a null character, if the command name is mv. Add the slash back if an error occurs.
2019-06-21bc: remove a power test that is now an errorGavin Howard
2019-06-21Fix cp -r dir/. symlink child.makepost
Remove the existing link before trying to re-create, passing the test. Add -p to the -r test as a regression guard, portage calls cp with both.
2019-06-11find: add -true/-false.Elliott Hughes
Used near the end of the AOSP build. Almost there! (This patch also fiddles with the help text to be able to slip the new options in without requiring so much extra space.)
2019-06-10kill: add trivial -l tests.Elliott Hughes
(`kill -l HUP` and `kill -l 1` both said "HUP" instead of giving you back the one you didn't provide, before my real-time signal patch.)
2019-06-10Add failing test for cp -r dir/. symlink child.makepost
Gentoo packages that build multiple variants, like once for every version of a slotted dependency, currently fail to install if their source dir includes a relative symlink to own child. Affects lots of Python infrastructure, for example meson and setuptools. You've already run into this issue, since cp.test has a todo. It's from 2008 though, so I guess I'll bump with an expanded test case. Libc is musl-1.1.22.
2019-06-03Clean up remaining TAIL_SEEK references.Elliott Hughes
2019-05-29Teach find -printf about %.Ns patterns, tweak help text, add tests.Rob Landley
2019-05-28Workaround for mkshRob Landley
(which doesn't handle "VAR=blah thingy" right if thingy is a function, the variable winds up staying set after the function returns.)
2019-05-26Teach sed about +N range ending.Rob Landley
2019-05-25grep: add --exclude-dir.Elliott Hughes
Used quite a lot, especially with `--exclude-dir=.git`.
2019-05-19Fix tar sparse extract with extension blocks.Rob Landley
2019-05-19More tar tests.Rob Landley
2019-05-17Fix a missing else, and an inverted test hidden by the missing else.Rob Landley
Add test to show failure case.
2019-05-16Add human_readable_long() for more than 3 digits of output.Rob Landley