# 4.4.2 breaks kio 5.116.0 on some platforms

**URL:** https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782
**Category:** Development
**Created:** [August 6, 2026, 8:23pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782 "2026-08-06T20:23:16Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![wiz](https://discourse.cmake.org/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@wiz](https://discourse.cmake.org/u/wiz)
#### Post date: [August 6, 2026, 8:23pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782/1 "2026-08-06T20:23:16Z")

</div>

The timing fits, so I suspect that the update from 4.4.0 to 4.4.2 in pkgsrc broke the build of kio 5.116.0 on a couple platforms.

Here’s an example build log:

[https://releng.netbsd.org/pkgreports/shadow/HEAD/NetBSD-11.0-x86\_64/20260802.1144/kio-5.116.0nb12/configure.log](https://releng.netbsd.org/pkgreports/shadow/HEAD/NetBSD-11.0-x86_64/20260802.1144/kio-5.116.0nb12/configure.log)

from a NetBSD 11.0/x86\_64 build; I see the same on 11.99.7, and here’s a build on SmartOS:

[https://reports.pkgci.org/SmartOS/upstream/trunk/20260802T225326Z/kio-5.116.0nb12/configure.log](https://reports.pkgci.org/SmartOS/upstream/trunk/20260802T225326Z/kio-5.116.0nb12/configure.log)

All the places I could find in the code look like this:

```auto
if(ACL_FOUND)
target_link_libraries(KF5KIOCore PRIVATE ${ACL_LIBS})
endif()

```

which I thought should be safe. Any ideas?

Thanks,  
Thomas

---

<div class="post-metadata">

### Author: ![brad.king](https://discourse.cmake.org/user_avatar/discourse.cmake.org/brad.king/32/11_2.png) [@brad.king](https://discourse.cmake.org/u/brad.king)
#### Post date: [August 7, 2026, 2:24pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782/2 "2026-08-07T14:24:13Z")

</div>

I can reproduce this by building `kio`’s `v5.116.0`tag on Debian with the `libacl1-dev` package _not_ installed. I’ve narrowed the problem to a simple example and opened [CMake Issue 28026](https://gitlab.kitware.com/cmake/cmake/-/work_items/28026). Please follow that for further updates.

---

<div class="post-metadata">

### Author: ![brad.king](https://discourse.cmake.org/user_avatar/discourse.cmake.org/brad.king/32/11_2.png) [@brad.king](https://discourse.cmake.org/u/brad.king)
#### Post date: [August 7, 2026, 2:30pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782/3 "2026-08-07T14:30:02Z")

</div>

FWIW, `kio` `5.116.0` also has a bug in [`cmake/FindACL.cmake`](https://github.com/KDE/kio/blob/v5.116.0/cmake/FindACL.cmake#L49) that helps trigger the failure. The find module leaks `CMAKE_REQUIRED_LIBRARIES` to its caller. The problem [persists in `kio` `master`](https://github.com/KDE/kio/blob/d1cb3b7b78a6000f744662e2d1a09775c12e4473/cmake/FindACL.cmake#L49).

---

<div class="post-metadata">

### Author: ![brad.king](https://discourse.cmake.org/user_avatar/discourse.cmake.org/brad.king/32/11_2.png) [@brad.king](https://discourse.cmake.org/u/brad.king)
#### Post date: [August 7, 2026, 3:19pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782/4 "2026-08-07T15:19:18Z")

</div>

Also, `kio`’s `FindACL` module shouldn’t bother with `check_symbol_exists` calls on a library that wasn’t found.

---

<div class="post-metadata">

### Author: ![Matthew\_Woehlke](https://discourse.cmake.org/user_avatar/discourse.cmake.org/matthew_woehlke/32/2345_2.png) [@Matthew\_Woehlke](https://discourse.cmake.org/u/Matthew_Woehlke)
#### Post date: [August 7, 2026, 4:56pm UTC](https://discourse.cmake.org/t/4-4-2-breaks-kio-5-116-0-on-some-platforms/15782/5 "2026-08-07T16:56:22Z")

</div>

Seconded. It happens the check should at least not throw an error if you try to feed it `<name>-NOTFOUND` in `CMAKE_REQUIRED_LIBRARIES`, but I suspect that breaks if there’s more than one library. Regardless, looking for symbols in a not-found library is indeed silly. 🙂