<div dir="ltr"><div dir="ltr"><br></div><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Apr 6, 2023 at 8:37 AM Rob Landley <<a href="mailto:rob@landley.net">rob@landley.net</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><br>
<br>
On 4/6/23 09:55, Rob Landley wrote:<br>
> On 4/5/23 17:13, Yifan Hong wrote:<br>
>> Hello Rob Landley,<br>
>> <br>
>> The hack has not been reverted yet, but I'll do this immediately<br>
>> (<a href="https://android-review.git.corp.google.com/c/kernel/build/+/2525840" rel="noreferrer" target="_blank">https://android-review.git.corp.google.com/c/kernel/build/+/2525840</a>). I'll<br>
>> also revert other hacks.<br>
>> <br>
>> A separate issue I see with the prebuilt tar is that it doesn't seem to extract<br>
>> a single file prefixed with `./`. This is best illustrated with the following<br>
>> example (see attached repro.sh and result.txt)<br>
>> <br>
>> That is, if an archive contains `./foo` as a entry, then extracting with `tar xf<br>
>> foo.tar.gz ./foo` doesn't seem to extract the file properly.<br>
> <br>
> Huh, looks like I broke that since the 0.8.9 release. And what's _really_ weird is:<br>
> <br>
> $ ./tar cv ./README > fweep.tar<br>
> ./README<br>
> $ ./tar tv ./README < fweep.tar<br>
> $ ./tar cv ./README | ./tar tv ./README<br>
> ./README<br>
> <br>
> Doesn't work when written to/from a file but does when piping between two<br>
> processes? Something to do with the changed exit path maybe? I wonder what<br>
> strace says...<br>
> <br>
> openat(AT_FDCWD, "/lib/x86_64-linux-gnu/libnss_nis.so.2", O_RDONLY|O_CLOEXEC) = 3<br>
> ...<br>
> socket(AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0) = 3<br>
> connect(3, {sa_family=AF_UNIX, sun_path="/var/run/nscd/socket"}, 110) = -1<br>
> ENOENT (No such file or directory)<br>
> ...<br>
> openat(AT_FDCWD, "/etc/passwd", O_RDONLY|O_CLOEXEC) = 3<br>
> <br>
> I hate glibc.<br>
> <br>
>> Could you please take a look? Thanks!<br>
> <br>
> Looking...<br>
<br>
Sorry for the delay, slack stopped working with debian "oldstable"'s version of<br>
chrome (big fullscreen "too old" tantrum page), and I went down a rathole of<br>
asking #devuan IRC why a newer one wasn't in the backports repo, and they<br>
forwarded me to the Debian guys who said Google's builds are too brittle to even<br>
_try_ building chromium on anything but the newest release with all the current<br>
compilers and libraries. So I have to do a major OS version update before I can<br>
use slack on my laptop again, and confirming that took a while. (I miss static<br>
binaries.)<br></blockquote><div><br></div><div>(yeah, the flip side of hermetic builds is that folks like debian who will neither use your binaries nor buy into "i'll use your compiler and sysroot and everything" are in for a world of pain.)</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Anyway, I broke this adding the --wildcard stuff, commit 24133d5e23d7 should fix<br>
it. Sorry about that...<br></blockquote><div><br></div><div>thanks! updating AOSP in <a href="https://android-review.git.corp.google.com/c/platform/external/toybox/+/2526229">https://android-review.git.corp.google.com/c/platform/external/toybox/+/2526229</a>, and i'll start on new host prebuilts for elsk once that's in...<br class="gmail-Apple-interchange-newline"></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
> Rob<br>
<br>
Rob<br>
</blockquote></div></div>