-
Notifications
You must be signed in to change notification settings - Fork 852
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Files in DrvFs still reporting wrong type? #2464
Comments
@Warblefly - That issue should be fixed. Can you give us some reproducible steps so we can check this out locally? Adding @SvenGroot. |
This actually looks like an LxFs issue, since it's referring to a file under /usr. LxFs was not affected by the previous issue, and hasn't received any real changes recently. As @benhillis said, we need a repro to look into this. |
OK, I'll try. Just got this, though, in a
|
And every time I try to compile GCC from a directory on the DevFs system, a different file has been transformed into something wrong:
|
As for steps to a reproducible error, mine are to compile GCC 7.2.0 as a cross-compiler, using |
Completely uninstalled the subsystem, deleted lxss, removed all working directories, re-installed, and
|
The same problem occurs on Build 16281, which landed on my machine today. The host GCC is the supplied Ubuntu GCC, which is version 5.4.0
|
Is this issue #1735 again, perhaps? As before, in this issue, not only is the file 'missing' but the path searched is wrong. This also occurs using a freshly-compiled gcc-7.2.0 as the compiler, as well as the stock Ubuntu 5.4.0.
|
Compiling GCC on the 'native' Linux filesystem entirely within lxss does not fail. But on DrvFs, it does. |
Experiencing the same issue on build 16299.64. Compiling Linux kernel 3.10 with GCC 7.1 custom toolchain. The compiler errors on a different file each time but the error code is always the same.
A better formatted version: https://hastebin.com/qigevusihu.coffeescript The source for the same is here https://github.com/nutcasev15/Holo-N/tree/RW The issue is resolved if I delete all the files in the git repo with rm -rf * & then force checkout the branch. But it only stays as long as the laptop is on. If I shut down and try to compile after rebooting the issue appears again. The source & the TC are both on /mnt/d i.e. DrvFs. |
This just started happening to me too, build 16299.rs3_release. I keep getting |
Got a similar error here when building a snap package under Ubuntu 16.04 WSL(Windows 10 version 1803 build 17134.254):
The file should be in lxfs instead of drvfs, though. |
What is underneath drvfs? Ntfs or? |
Much has changed in the WSL world since this was filed. If you still see this issue, please file a new one. Thanks! |
In Build 16278, I'm seeing plenty of compilations fail with GCC complaining in ways such as:
/usr/include/x86_64-linux-gnu/sys/select.h:33:25: fatal error: lto/bits/sigset.h: Invalid argument
...which is the type of error that used to happen when DrvFs-mounted drives were borked a few flights ago. "Invalid argument" meaning a file is a wrong type (e.g. directory) not merely missing. Is this still happening?
The text was updated successfully, but these errors were encountered: