When attempting to remove multiple directories containing many files and subdirectories (let's call them $bigdirectory1, $bigdirectory2 and $bigdirectory3), using the following command:

rm -frv $bigdirectory1 $bigdirectory2 $bigdirectory3

rm throws the following errors:

...
rm: $directory1: Directory not empty
rm: $directory2: Directory not empty
rm: $directory3: Directory not empty
rm: $directory4: Directory not empty
rm: $directory5: Directory not empty
rm: $directory6: Directory not empty
...

Where $directory1 to $directory6 are subdirectories in $bigdirectory1 to $bigdirectory3 that are left empty and require a rerun of the command:

rm -frv $bigdirectory1 $bigdirectory2 $bigdirectory3

to remove them completely.

(To recreate the issue just grab any Linux kernel tarball (e.g. version 5.5.5), extract the tarball and run rm -frv on the extracted directory and notice how it rm stops midway through the deletion and throws the Directory not empty error)

Could this be possibly related to the following lines in rm.c?

...
  // Intentionally fail non-recursive attempts to remove even an empty dir
  // (via wrong flags to unlinkat) because POSIX says to.
  if (dir && !(toys.optflags & (FLAG_r|FLAG_R))) goto skip;
...

So, I have tried this in my 64 bit linux vm and also in arm64 android, the ls -Z doesn't show the selinux content I used the binaries from here: http://landley.net/toybox/bin/

Also, the binaries provided in the android OS seems to print selinux without any issue

Hi! When I call any applet with the --help option, it would be good if you could please start the output with the following lines, then a newline.

Toybox multi-call binary $(git describe) Web page: http://landley.net/toybox/ Bug reports: https://github.com/landley/toybox/issues Feedback: http://lists.landley.net/listinfo.cgi/toybox-landley.net Contributors, see: https://landley.net/toybox/cleanup.html#intro

Some users have Toybox because it came preinstalled on their device. The above information will make it easier for such users to figure out where and how to submit bug reports and patches.

git describe looks at tag names and commits, and hopefully outputs a version number like "0.7.1-106-g409a8e0" or "0.7.2".

I had to make getdirname return "." instead of the file when the path has no '/', like dirname does, for my code to work correctly. I checked the new flag and toybox ln in general against GNU's ln tests. Use cases: https://git.archlinux.org/svntogit/packages.git/tree/trunk/update-ca-trust?h=packages/ca-certificates#n38 https://git.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/linux#n201

Apologies if this has been explained elsewhere... I'm new to this...

The AARCH64/ARM64 binary won't run on macOS Monterey 12.x and I can't seem to build it from source either.

Should it be working? Trying to build Android and I think it's dependent on toybox. Can't find a working build of toybox for Apple Silicon anywhere.

Thanks.

Building with make fails with errors from clang on MacOS. The reason is that gsed is not found and there is no warning or instruction to install it.

I think, that either a warning should be added (possibly into scripts/portability.sh) or a note into the "Building toybox" section of README. Better both.

P.S. Also, to compile toybox I had to set CPUS=1, comment #include "generated/newtoys.h" in lib/help.c, and comment quite a few commands in .config, ending up with 136 of them. (Maybe it helps someone out there.)

OS version: MacOS Mojave 10.14.6

This also doesn't work with install -D file1 file2 dir, where dir doesn't exist, but busybox also messes that up (differently, instead of making dir and putting file[12] there, it makes dir's parents and makes puts file2 as dir) so I'm not sure what should be done in that situation. Either way, with any install I've seen that has -t, it creates the dir given to -t when -D is also passed.

Infrequent Crash for id|toybox on iOS 14.4 | byte read Translation fault | possible pointer authentication failure

Toolchain Details Apple clang version 12.0.0 (clang-1200.0.32.29) Xcode 12.4 Build version 12D4e -sdk iphoneos14.4

Built from https://github.com/landley/toybox/commit/f1be076b52adcad7b2419315cc148009a1bc2fec.

Reproduction Case: Using Terminal and ssh'd into the device as root, Executed the command: id

id

... Hardware Model: iPhone12,1 Process: toybox [489] Identifier: toybox Code Type: ARM-64 (Native) Parent Process: sh [453] OS Version: iPhone OS 14.4 (18D52) ... Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Subtype: KERN_INVALID_ADDRESS at 0x6b726f7774656e50 -> 0x0000007774656e50 (possible pointer authentication failure) VM Region Info: 0x7774656e50 is not in any region. Bytes after previous region: 32017575505
REGION TYPE START - END [ VSIZE] PRT/MAX SHRMOD REGION DETAIL commpage (reserved) 1000000000-7000000000 [384.0G] ---/--- SM=NUL ...(unallocated) --->
UNUSED SPACE AT END

Termination Signal: Segmentation fault: 11 Termination Reason: Namespace SIGNAL, Code 0xb Terminating Process: exc handler [489] Triggered by Thread: 0

Thread 0 name: Dispatch queue: com.apple.main-thread Thread 0 Crashed: 0 libsystem_platform.dylib 0x00000001e953bbc4 _platform_strlen + 4 1 libsystem_c.dylib 0x00000001a6e8c16c __vfprintf + 4328 2 libsystem_c.dylib 0x00000001a6eaee0c __v2printf + 396 3 libsystem_c.dylib 0x00000001a6eaf044 __xvprintf + 240 4 libsystem_c.dylib 0x00000001a6e8af80 vfprintf_l + 68 5 libsystem_c.dylib 0x00000001a6e897b8 printf + 84 6 id 0x00000001042ea540 0x1042bc000 + 189760 7 id 0x00000001042ce930 0x1042bc000 + 76080 8 id 0x00000001042ce320 0x1042bc000 + 74528 9 id 0x00000001042ce978 0x1042bc000 + 76152 10 libdyld.dylib 0x000000019d6c66b0 start + 4

Thread 0 crashed with ARM Thread State (64-bit): x0: 0x6b726f7774656e5f x1: 0x6b726f7774656e50 x2: 0x0000000000000000 x3: 0x000000016bb432c3 x4: 0x000000016bb42e80 x5: 0x000000016bb43810 x6: 0x0000000000000031 x7: 0x0000000000000f70 x8: 0x6b726f7774656e5f x9: 0x00000001a6ec5aac x10: 0x0000000000000001 x11: 0x0000000000000003 x12: 0x0000000000000000 x13: 0x0000000000000000 x14: 0x0000000000000010 x15: 0x0000000000000001 x16: 0x00000001e953bbc0 x17: 0x00000001a6e8dde4 x18: 0x0000000000000000 x19: 0x000000016bb42d70 x20: 0x0000000000000003 x21: 0x0000000000000073 x22: 0x0000000104307526 x23: 0x000000000000000a x24: 0x0000000000000000 x25: 0x0000000104307524 x26: 0x6b726f7774656e5f x27: 0x0000000000000000 x28: 0x000000016bb42e88 fp: 0x000000016bb43140 lr: 0x00000001a6e8c16c sp: 0x000000016bb42d70 pc: 0x00000001e953bbc4 cpsr: 0x20000000 esr: 0x92000004 (Data Abort) byte read Translation fault

Binary Images: 0x1042bc000 - 0x104307fff id arm64e ... 0x19d6c5000 - 0x19d6fffff libdyld.dylib arm64e <21b19919133438bcb233896e929945e0> /usr/lib/system/libdyld.dylib ... 0x1a6e49000 - 0x1a6ec7fff libsystem_c.dylib arm64e <961a8fb0de8a3567b4375d99cd549fae> /usr/lib/system/libsystem_c.dylib 0x1e9536000 - 0x1e953cfff libsystem_platform.dylib arm64e <88fa2f54074d32c49f4879eb67e67b7b> /usr/lib/system/libsystem_platform.dylib EOF

The answer could be that this functionality is not expected to work on iOS 14.4. Please let me know if any additional details can be provided.

This UB breaks builds of sed with HWASan enabled in Android, because the check of the while (command) is optimized away.

Adding to NULL pointer and subtracting back is undefined behaviour, as it is pointer arithmetic beyond the bounds of the object. The compiler is free to assume that ptr - 1 is non-NULL.

Adding prefix to toybox (e.g. renaming toybox to mytoybox) produces the following error:

toybox: Unknown command ./mytoybox (see "toybox --help")

Is there anyway around to this problem?

toybox sed crashes with a segfault when given certain substitution commands that specify to only substitute the nth match where n > 1.

Steps to reproduce

echo "e" | toybox sed -e "s/e//2"

Analysis

In sed.c:498, sed decides correctly that, since it is looking at the 1st match but should only replace the nth match, the current match should be skipped. To do so, it copies the remainder of the current line not covered by the match into l2. The memory for l2 is allocated in line 526, when carrying out a substitution. However, if the first match already needs to be skipped, there have been no substitutions yet and l2 is still a null pointer, leading to the crash.

Additional Information

This behavior was detected using techniques developed by the SYMBIOSYS research project at COMSYS RWTH Aachen University. This research is supported by the European Research Council (ERC) under the EU's Horizon 2020 Research and Innovation Programme grant agreement n. 647295 (SYMBIOSYS).

On Android 12

/sdcard/jidanni/bin/myscript[246]: can't create temporary file /data/local/sh8mcytt.tmp: Permission denied

And

export TEMPDIR=...

didn't help.

Please be sure all the toybox commands don't try to use those directories that are out of bounds in Android 12, e.g., using non-rooted Samsung stock pre-installed toybox 8.4.

My shell script had some pipes and here-docs in it.

In http://landley.net/toybox/ we read

What commands are planned/implemented in toybox? The current list of commands implemented by toybox is on the status page...

Okay. That's great. But it would still be great to have a additional list of all the commands grouped by their functionality:

• networking commands
• file status commands
• data manipulation commands
• etc.

rrt[]msg* for SRV was copy+paste (from MX) wrong/confusing 

--- toybox-0.7.8/toys/pending/host.c_ORIG	2019-05-23 20:49:04.760865776 +0000
+++ toybox-0.7.8/toys/pending/host.c	2019-05-23 20:49:18.475658895 +0000
@@ -52,7 +52,7 @@
   [6] = { "SOA", "start of authority", PL_SOA },
   [12] = { "PTR", "domain name pointer", PL_NAME },
   [15] = { "MX", "mail is handled", PL_MX },
-  [33] = { "SRV", "mail is handled", PL_SRV },
+  [33] = { "SRV", "has SRV record", PL_SRV },
   [255] = { "*", 0, 0 },
 };
 

The TXT RR RDATA returned by DNS servers is rfc1035 <character-string>s
<character-string> is a single length octet followed by that number of characters. 
... notice how the first RDATA byte is usually 1-less than the RDLENGTH?

If max allowed TXT record length is 255 chars + we assume only one char string,
then we can just verify and discard the first byte of TXT RDATA. If first byte
doesn't match (1 shorter than RDLENGTH) then don't discard it.
... throw an error instead? ... ignore the response as invalid?


--- toybox-0.8.6/toys/pending/host.c_ORIG	2022-07-25 17:30:48.958174663 +0000
+++ toybox-0.8.6/toys/pending/host.c	2022-07-25 17:32:40.178209094 +0000
@@ -170,6 +170,10 @@
         dn_expand(abuf, abuf+alen, p, plname, sizeof(plname));
         break;
       case PL_TEXT:
+        if ((pllen > 1) && (pllen <= 256) && (pllen-1 == p[0])) {
+          pllen--;
+          p++;
+        }
         snprintf(plname, sizeof(plname), "\"%.*s\"", pllen, p);
         break;
       case PL_SOA:

Hi Rob~ I have some "route.c"questions need to ask and my toybox version is 0.8.8.

1. I have add route command (route add -net 192.168.1.0 netmask 255.255.255.0 gw 192.168.1.10 dev eth0), this may show crash log. I find that the problem is at "line 355" char *prefix = strtok(0, "/");
   I have changed code into char *prefix = strtok(tgtip, "/"); This changed may not crash again.

2. I have fixed the first question at my local side. Then I add the same route command (route add -net 192.168.1.0 netmask 255.255.255.0 gw 192.168.1.10 dev eth0). I find that the new routing rules are not added to the routing table. This problem I need to ask you for helping. Thank you~

3. As above problem number two that I don't understand, so I download static binaries at README. I have downloaded toybox-armv4l this binary file and tested. I use the command ./toybox-armv4l route add -net 192.168.1.7 netmask 255.255.255.0 gw 192.168.1.13 dev eth0 , but this command doesn't work at all. The routing table doesn't show the routing rules that I added, so I change the command into ./toybox-armv4l route add -net 192.168.1.7/24 gw 192.168.1.13 dev eth0. Although the routing rule have been added, I find that Netmask of the Routing table is always 255.255.255.255. Even I change Netmask 192.168.1.7/26, the Routing table is always 255.255.255.255. Is problem of Netmask a bug? If it is a bug, how to fix in "route.c"?

4. I would like to ask why the master code test results are different from the static binary file.

Can you tell me how to fix "route.c" that can let route feature working normally(add routing rules normally). Thank you and sorry for asking many problems~