Re: [lmi] Tests PR 174

[Vadim Zeitlin]

On Wed, 24 Mar 2021 22:21:48 +0000 Greg Chicares <gchicares@sbcglobal.net> 
wrote:

GC> >  So could you please revert this commit using "git revert 593a7121f" and
GC> > push it to fix the current CI failure in clang build?
GC> 
GC> Done.

 Thanks again! The CI build didn't finish yet, but they should pass now.

GC> Returning to "much, much more expensive" above...okay, submodule
GC> updates are more expensive than single-file updates; but maybe
GC> they still belong in a hooks/post-[something] script.

 It's a bit confusing, but I've just replied to this in another message in
this thread.

GC> My reasoning is that submodule updates should do nothing, fairly
GC> quickly, when there's nothing to do, and in that case they really must
GC> be idempotent.

 Yes, this is true. But suppose you upgrade wx to 3.1.5 and then want to
compare behaviour with the current and new wx versions. I'd expect
switching wx submodule between the two to take noticeable time as the diff
as of right now is

 2411 files changed, 160006 insertions(+), 147282 deletions(-)

 It's probably still bearable under Linux with fast SSD because Git is
amazingly fast (under Linux; things are much, much slower under MSW), but
I'd still rather avoid doing it implicitly.

VZ

pgpJ65pv2bFGO.pgp
Description: PGP signature