Re: Proposed update to bmake-20240711

[Simon Gerraty <sjg%crufty.net@localhost>]

On Fri, 12 Jul 2024 18:37:42 +0100, Jonathan Perkin writes:
>* On 2024-07-12 at 17:47 BST, Simon Gerraty wrote:
>
>>You can use --with-defshell=sh=/path/to/mksh if you want to force it to
>>use the sh specification - which may or may not provide better results
>>(though I would not be too hopeful).
>
>Well, it's different:

FYI I have updated

https://www.crufty.net/ftp/pub/sjg/beta/bmake-20240711.tar.gz

after applying the patches I gleaned from github,
and expanding BROKEN_TESTS for ksh and mksh
to the ones I could not easily fix.

I noticed that shell-csh segfaults with mksh

fwiw I tested with

boot-strap --with-defshell=ksh	(pdksh on my machine)
boot-strap --with-defshell=/path/to/mksh

both now pass.

>
>   ===> Testing for bmake-20240711
>   cd /work/git/pkgsrc/devel/bmake/work/bmake/unit-tests && MAKEFLAGS= /work/g
>it/pkgsrc/devel/bmake/work/bmake/bmake -r -m / test
>   bmake[2]: "/work/git/pkgsrc/devel/bmake/work/bmake/unit-tests/Makefile" lin
>e 47: warning: "mkdir -p /var/folders/w7/88jp9ds51v7f5pl61gzkhmt40000gn/T//uid
>501" exited on a signal
>   bmake[2]: "/work/git/pkgsrc/devel/bmake/work/bmake/unit-tests/Makefile" lin
>e 512: warning: Skipping broken tests: sh-flags shell-ksh
>   *** Signal 11
>
>In true macOS fashion I don't know what is generating this, I've added 
>codesign signatures to both bmake and mksh but neither is generating a 
>coredump.

Fwiw I don't have to anything special for bmake on Darwin

>
>I wouldn't worry too much, we only have to use mksh in pkgsrc because 
>using a system shell breaks System Integrity Protection, so this is very 
>much a corner case.  It's clear that bmake is functioning ok though.
>
>-- 
>Jonathan Perkin   -   mnx.io   -   pkgsrc.smartos.org
>Open Source Complete Cloud   www.tritondatacenter.com