[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Regtest profiling differences
From: |
Jonas Hahnfeld |
Subject: |
Re: Regtest profiling differences |
Date: |
Sat, 01 Feb 2020 17:26:07 +0100 |
User-agent: |
Evolution 3.34.3 |
Am Samstag, den 01.02.2020, 11:17 -0500 schrieb Dan Eble:
> On Feb 1, 2020, at 10:15, Jonas Hahnfeld <
> address@hidden
> > wrote:
> > How distracting are the profiling differences for your work? I'm
> > obviously not a big fan of reverting bc8a3fa7e4, but I also don't want
> > to block anybody else from making progress.
>
> The log differences are more annoying than the profile differences. I'd
> prefer to see your hash patch ushered through the countdown faster rather
> than see your Python 3 changes reverted.
I just found a possible workaround for the short-term:
---
commit 04a0dc897629d1606d12d54b5c0e15e9cd003652
Author: Jonas Hahnfeld <address@hidden>
Date: Sat Feb 1 17:02:49 2020 +0100
Disable Python's hash randomization
This avoids overly many differences when profiling regtests.
diff --git a/stepmake/stepmake/generic-vars.make
b/stepmake/stepmake/generic-vars.make
index 3c482bb415..e65c4407ba 100644
--- a/stepmake/stepmake/generic-vars.make
+++ b/stepmake/stepmake/generic-vars.make
@@ -63,6 +63,10 @@ endif
LANG=
export LANG
+# disable Python's hash randomization until 'make check' is fixed
+PYTHONHASHSEED=0
+export PYTHONHASHSEED
+
INFO_DIRECTORIES = Documentation
---
If this is ok, I can push the commit to staging and we can follow the
normal process for #5721.
Regards,
Jonas
signature.asc
Description: This is a digitally signed message part