tools/coreutils: update to 9.1#9347
Conversation
|
This bugfix sounds important and I think should be patched: https://git.savannah.gnu.org/cgit/coreutils.git/commit/?id=e8b56ebd536e82b15542a00c888109471936bfda. Arch, Alpine and Void import this patch. |
|
Prefix title with tools/ |
708e850 to
b017602
Compare
|
Does anyone have some insight into why the macos build failed? |
|
I'm surprised no logs are stored. I'll look into that later |
|
@aparcar - At least there is some output now: Is there a way to build with the verbose setting? I do not have a MacOS device. Perhaps it can be virtualized. Never looked into that. |
diff --git a/.github/workflows/tools.yml b/.github/workflows/tools.yml
index 5b400b616d69..fcf7c04ceedc 100644
--- a/.github/workflows/tools.yml
+++ b/.github/workflows/tools.yml
@@ -114,12 +114,13 @@ jobs:
run: |
cd "$WORKPATH"
pwd
+ printf 'CONFIG_DEVEL=y\nCONFIG_BUILD_LOG=y\n' > .config
make defconfig
- name: Build tools
run: |
cd "$WORKPATH"
- make tools/install -j$(nproc) BUILD_LOG=1
+ make tools/install -j$(nproc) || make tools/install V=s -j1
- name: Upload logs
if: failure()
You've access to GH Actions so you've access to such machine.
Via GH Actions you've already access to the Linux/Windows/macOS virtual machines. The question is, for how long we're going to enjoy this free lunch :-)
|
Long time. Microsoft has enough money to subsidize CI.
|
|
Please rebase this onto a5eeac8 since it fixes the log issue for macOS.
@ynezz I thought of this to however it may result in a failure of the parallel build which is fixed in a second run and thereby hidden. I prefer the build log artifact variant since a build finishes and can't silently fail. |
|
I added the blocked tag, this shouldn't be added prior the branch. |
|
@graysky2 you did, I find the CI artifacts right here: |
|
|
There are two issues to be resolved.
Suggest solution:
|
|
Since there are two I am following this wiki page. |
|
@zhanhb - Made the patch and implemented your edits but still choked on macos. |
|
File |
669ade4 to
adc3530
Compare
|
Can you rebase this? The macOS patch should already be in master branch. |
I removed that line but masos-latest still fails |
|
News on this? |
|
Can you push -j1 V=s in another commit on the CI just to see the macos error? |
|
Where are we at with this PR? |
|
Needs a rebase. |
|
macos test is still failing it seems :/ |
|
@graysky2 Please remove upstreamed patch: 001-chmod-fix_exit_status_when_ignoring_sylinks.patch. |
|
@lovehackintosh - why do you request this? please see conversation string above for rationale to include. |
|
@graysky2 I downloaded https://ftp.gnu.org/gnu/coreutils/coreutils-9.1.tar.xz and unpacked, |
|
@lovehackintosh - you are right, fixed |
|
001-m4.patch requires refresh, please run |
|
|
Try to run |
|
Same error |
|
I refreshed it on macOS, you just need to apply this patch |
|
The "dirty" patch must still cleanly apply or else the two checks that ran would have errored out, no? https://github.com/openwrt/openwrt/actions/runs/4082720887/jobs/7037402090 |
|
What is the meaning of |
|
I am just pointing out that neither of these checks detected a dirty patch. |
I think we need to make the patch apply cleanly, so a refresh is required. |
|
I applied it, but not sure why the build checks did not flag it. I am also not sure why I cannot refresh it myself. |
|
If you want to refresh it on Linux, you need to run |
|
@hauke Can you merge this? |
In addition to version update, this commit applies a fixup to allow building on MacOS involving renaming: [gt_TYPE_WINT_T] --> [gt_TYPE_WINT_T_GNUTLS] suggested by zhanhb. Build system: x86_64 Build-tested: bcm2711/RPi4B Signed-off-by: John Audia <[email protected]>

In addition to the version update, this commit also applies a fixup to allow building on MacOS involving renaming: [gt_TYPE_WINT_T] --> [gt_TYPE_WINT_T_GNUTLS] suggested by zhanhb.
Build system: x86_64
Build-tested: bcm2711/RPi4B
Signed-off-by: John Audia [email protected]