Re: ZTST_continue (was Re: Lots of test failures when --disable-multibyte)

[Peter Stephenson <p.w.stephenson@xxxxxxxxxxxx>]

> On 06 April 2022 at 06:32 Jun T <takimoto-j@xxxxxxxxxxxxxxxxx> wrote:
> > 2022/04/06 5:29, Peter Stephenson <p.w.stephenson@xxxxxxxxxxxx> wrote:
> > 
> > This might help?
> 
> I've been preparing basically the same patch. The only differences are:
> use (( $ZTST_continue )) so that "ZTST_continue=0 make check" will work,
> rename "failures" --> "ZTST_failures" to avoid possible (future) conflicts.

You might as well go ahead.
 
> With the patch below:
> (1) if a chunk in %prep fails then the test is aborted immediately.
> (2) If the "bad placement of 'test' section" happens, we do not exit immediately
> but run ZTST_cleanup.
> Or should we run ZTST_cleanup even in (1)? (it will do no harm...)

Well, since it shouldn't happen it's hard to know what to do if it does.
Cleaning up anything that did get added in the preparation would be kind
of logical, so I can't see why not.

pws