[PATCH] Remove redundant path_nulls checks in setPathObject/Array
Hi all,
I noticed redundant path_nulls checks in setPathObject and setPathArray,
so here's a small refactoring patch to remove them.
- Both checks were introduced in c694701 (12 May 2015, "Additional
functions and operators for jsonb") with the functions' initial code.
- Both functions always had a single call site, `setPath`, as they are
internal implementations for the two cases (array/object) in a generic
function.
- Later, in 1edd4ec (4 Oct 2015, "Disallow invalid path elements in
jsonb_set"), the path_nulls check was propagated into setPath itself.
If desired, those conditions can be transformed into Assert()s with a
comment about setPath() validation.
---
Ilya Gladyshev // foxido.dev
The following review has been posted through the commitfest application:
make installcheck-world: tested, passed
Implements feature: tested, passed
Spec compliant: not tested
Documentation: not tested
Hello,
This patch looks good to me.
There is no SQL change: no need to check SQL compliance.
IMO no documentation change is needed.
Regards,
PF.
The new status of this patch is: Ready for Committer
Hi all,
On Mon, Sep 7, 2026 at 11:12 AM Ilya Gladyshev <ilya.gladyshev@linux.dev> wrote:
Hi all,
I noticed redundant path_nulls checks in setPathObject and setPathArray,
so here's a small refactoring patch to remove them.- Both checks were introduced in c694701 (12 May 2015, "Additional
functions and operators for jsonb") with the functions' initial code.- Both functions always had a single call site, `setPath`, as they are
internal implementations for the two cases (array/object) in a generic
function.- Later, in 1edd4ec (4 Oct 2015, "Disallow invalid path elements in
jsonb_set"), the path_nulls check was propagated into setPath itself.If desired, those conditions can be transformed into Assert()s with a
comment about setPath() validation.
Thank you for the patch.
I reviewed and tested this patch. The path_nulls[level] check is
already done in setPath() before calling setPathObject() or
setPathArray(). I also checked the call sites and confirmed that both
functions are only called from setPath(), so the checks removed by the
patch are indeed redundant.
I didn't find any issues with the patch. It looks good to me.
Regards.
Solai
On Aug 26, 2026, at 20:03, Ilya Gladyshev <ilya.gladyshev@linux.dev> wrote:
Hi all,
I noticed redundant path_nulls checks in setPathObject and setPathArray,
so here's a small refactoring patch to remove them.- Both checks were introduced in c694701 (12 May 2015, "Additional
functions and operators for jsonb") with the functions' initial code.- Both functions always had a single call site, `setPath`, as they are
internal implementations for the two cases (array/object) in a generic
function.- Later, in 1edd4ec (4 Oct 2015, "Disallow invalid path elements in
jsonb_set"), the path_nulls check was propagated into setPath itself.If desired, those conditions can be transformed into Assert()s with a
comment about setPath() validation.---
Ilya Gladyshev // foxido.dev<0001-Remove-redundant-NULL-checks-in-setPathObject-Array.patch>
Looks reasonable to me, since setPath() already checks path_nulls[level] and raises an error for a NULL path element.
I think this patch does more than just remove a redundant check, it also eliminates a potential source of confusion. setPath() treats path_nulls[level] being true as an error, while setPathObject() currently treats the same condition as an indication of done.
Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/