repl: default useGlobal to true #7795
Conversation
|
s/globlal/global/ in the commit message |
silverwind
changed the title
cjihrig
force-pushed
the
useglobal
branch
2 times, most recently
from
5859f88
to
4573a53
Compare
|
@mscdex typo fixed and added the regression test. |
|
LGTM |
1 similar comment
|
LGTM |
|
Sigh. Windows. |
|
Fixed the Windows slash issue. Ran the CI again: https://ci.nodejs.org/job/node-test-pull-request/3369/ Only one unrelated failure. |
This is a partial revert of 15157c3. This change lead to a regression that broke require() in the CLI REPL, as imported files were evaluated in a different context. Refs: nodejs#5703 Fixes: nodejs#7788 PR-URL: nodejs#7795 Reviewed-By: Anna Henningsen <anna@addaleax.net> Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
|
I'm going to opt to not land this change on |
|
Sorry to bring up an old PR but I just came across this bug and it took me ages to get to the bottom of it. My project needs to work on 6.3 (deployment is beyond my control) so that's what I've been developing locally with. It's really confusing that my test suite runs ok but then if I manually run a test in the REPL it fails (due to requiring a module that modifies Just wanted to check you were aware that this bug still occurs on 6.3.0 and 6.3.1. Would it make sense to release 6.3.2 with this fix so people using 6.3 don't get this issue or is there a reason this needs to be in 6.4? |
|
@lukechilds unfortuantely the 6.3 line is no longer maintained. There have been a variety of security bugs that have been patched since. There will not be a new release on that line. While I can respect that the deployment is out of your control, do keep in mind that v6 will be going into LTS this month and any other releases in v6 will not be supported |
|
@thealphanerd I thought that might be the case, just checking. Thanks for the quick reply 👍 |
Checklist
make -j4 test(UNIX), orvcbuild test nosign(Windows) passesAffected core subsystem(s)
repl
Description of change
This is a partial revert of 15157c3. This change lead to a regression that broke
require()in the CLI REPL, as imported files were evaluated in a different context.A known issue test is available in #7793, which I'll port over here as a normal test.