[vim/vim] skip tempdir deletion test on Solaris (PR #20567)
Vladimír Marek
Test_delete_temp_dir() assumes Vim can detect when its cached private temp directory has been deleted, which depends on flock/dirfd support. Solaris does not provide flock(), so Vim keeps using the cached temp directory path after the directory has been deleted. Skip the test there.
BTW thank you, I have few more patches Solaris was hoarding and will be creating pull requests for those. Anything you don't like I'm more than happy to change or accept that this is not suitable for upstream.
Vlad
You can view, comment on, or merge this pull request online at:
https://github.com/vim/vim/pull/20567
Commit Summary
- d987312 skip tempdir deletion test on Solaris
File Changes
(1 file)
Patch Links:
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.
Christian Brabandt
chrisbra left a comment (vim/vim#20567)thanks
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.
Christian Brabandt
chrisbra left a comment (vim/vim#20567)While we are it and since you are familiar with Solaris, can you have a look here and update it according to your experience?
Thanks
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.
Vladimír Marek
vlmarek left a comment (vim/vim#20567)Hi, I'm happy to, but I'm not sure what to say there. I'm employed by Oracle to support opensource. Most of the time I'm the one fixing vim. Vim works great for us internally. I am heavy vim user myself. Currently there are no open bugs against vim in solaris. Neither internal or from customer. That said, we do not use vim as shipped by you. We have number of local patches. So how well vim works depends whether it is compiled by Oracle or custom from source code. My last week effort is to form the patches into shape (hopefully) acceptable by you. And alongside fix all test issues found. If you are interested, this is list of patches we ship:
https://github.com/oracle/solaris-userland/tree/master/components/vim/patches
Since you were kind enough to accept some of the pull requests the patches are a bit obsolete compared to my local workspace. Notable patches:
For a long time vim was unable to compile without https://github.com/oracle/solaris-userland/blob/master/components/vim/patches/si_id.patch . But today I was not able to reproduce the issue so I'm not sure what is going on here.
Another one is the frequent crash which is fixed by https://github.com/oracle/solaris-userland/blob/master/components/vim/patches/signalstacksize.patch . I'm not sure how much it is suitable for you even if we wrap it in #ifdef ...
Next critical patch is https://github.com/oracle/solaris-userland/blob/master/components/vim/patches/test-hang.patch . I have spent last three days on it and will be submitting it soon. I have distilled it to single line change for now. Without that the tests are hanging and there is potential for vim to hang too.
--- vim-9.2.0670/src/channel.c
+++ vim-9.2.0670/src/channel.c
@@ -4332,7 +4332,7 @@
}
// Reading a disconnection (readlen == 0), or an error.
- if (readlen <= 0)
+ if (len == 0 || readlen <= 0)
{
if (!channel->ch_keep_open)
ch_close_part_on_error(channel, part, (len < 0), func);
Rest of the patches is mostly test fixes. I run the tests in screen which makes things to have weird delays. er That I also do not feel to be suitable for you.
Overall I need to look at si_id again. If that is not an issue and the two patches mentioned are fixed in git too then I am confident to say that vim is well supported even when compiled from sources.
Last thing that comes to my mind. You are using github actions to test vim. Would you be interested in adding solaris as one of the test targets? From my recent experience you have enough test issues as it is, so I didn't think you would want to see another weird OS :)
Thank you
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.
Christian Brabandt
chrisbra left a comment (vim/vim#20567)Thanks for the feedback. I'll record this as fully supported :)
Regarding the github actions, I think a more obscure architecture can help finding bugs early. So this would be nice to have
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.
Vladimír Marek
vlmarek left a comment (vim/vim#20567)Hehe :)
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you are subscribed to this thread.