Adélie Linux issueshttps://git.adelielinux.org/groups/adelie/-/issues2023-10-25T06:39:06Zhttps://git.adelielinux.org/adelie/packages/-/issues/1093system/easy-kernel: ppc64: miscellaneous items in need of attention2023-10-25T06:39:06ZZach van Rijnsystem/easy-kernel: ppc64: miscellaneous items in need of attentionFound on ppc64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
...
BOOTCC arch/powerpc/boot/fdt_sw.o
BOOTCC arch/powerpc/boot/fdt_wip.o
DTC arch/powerpc/boot...Found on ppc64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
...
BOOTCC arch/powerpc/boot/fdt_sw.o
BOOTCC arch/powerpc/boot/fdt_wip.o
DTC arch/powerpc/boot/dts/ps3.dtb
BOOTAR arch/powerpc/boot/wrapper.a
WRAP arch/powerpc/boot/zImage.pseries
WRAP arch/powerpc/boot/zImage.maple
WRAP arch/powerpc/boot/dtbImage.ps3
WRAP arch/powerpc/boot/zImage.pmac
WRAP arch/powerpc/boot/zImage.epapr
ld: warning: div64.o: missing .note.GNU-stack section implies executable stack
ld: NOTE: This behaviour is deprecated and will be removed in a future version of the linker
ld: warning: arch/powerpc/boot/dtbImage.ps3 has a LOAD segment with RWX permissions
egrep: warning: egrep is obsolescent; using grep -E
GEN Module.symvers
CC [M] arch/powerpc/crypto/md5-ppc.mod.o
CC [M] arch/powerpc/crypto/sha1-powerpc.mod.o
CC [M] arch/powerpc/kvm/kvm-hv.mod.o
CC [M] arch/powerpc/kvm/kvm-pr.mod.o
CC [M] arch/powerpc/kvm/kvm.mod.o
...
```
Does this warrant a closer look? If not, feel free to close the issue.https://git.adelielinux.org/adelie/packages/-/issues/1092system/easy-kernel: aarch64: missing config options2023-10-25T06:32:14ZZach van Rijnsystem/easy-kernel: aarch64: missing config optionsFound on aarch64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
*
* Restart config...
* ...Found on aarch64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
*
* Restart config...
*
*
* ARMv8.4 architectural features
*
Enable support for the Activity Monitors Unit CPU extension (ARM64_AMU_EXTN) [Y/n/?] y
Enable support for tlbi range feature (ARM64_TLB_RANGE) [Y/n/?] (NEW)
*
* ARMv8.5 architectural features
*
Branch Target Identification support (ARM64_BTI) [Y/n/?] y
Enable support for E0PD (ARM64_E0PD) [Y/n/?] y
Enable support for random number generation (ARCH_RANDOM) [Y/n/?] y
Memory Tagging Extension support (ARM64_MTE) [Y/n/?] (NEW)
```https://git.adelielinux.org/adelie/packages/-/issues/1091system/easy-kernel: x86_64: missing config options2023-10-25T06:32:14ZZach van Rijnsystem/easy-kernel: x86_64: missing config optionsFound on x86_64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
*
* Restart config...
* ...Found on x86_64 as of 651822a4fe0f86e54441a7fd6c7faffd7d4d1ede:
```
*
* Restart config...
*
*
* Mitigations for speculative execution vulnerabilities
*
Mitigations for speculative execution vulnerabilities (SPECULATION_MITIGATIONS) [Y/n/?] y
Remove the kernel mapping in user mode (PAGE_TABLE_ISOLATION) [Y/n/?] y
Avoid speculative indirect branches in kernel (RETPOLINE) [Y/n/?] y
Enable return-thunks (RETHUNK) [Y/n/?] (NEW)
Enable UNRET on kernel entry (CPU_UNRET_ENTRY) [Y/n/?] (NEW)
Enable IBPB on kernel entry (CPU_IBPB_ENTRY) [Y/n/?] y
Enable IBRS on kernel entry (CPU_IBRS_ENTRY) [Y/n/?] y
Mitigate speculative RAS overflow on AMD (CPU_SRSO) [Y/n/?] (NEW)
Force GDS Mitigation (GDS_FORCE_MITIGATION) [N/y/?] n
*
* General architecture-dependent options
*
Kprobes (KPROBES) [Y/n/?] y
Optimize very unlikely/likely branches (JUMP_LABEL) [Y/n/?] y
Static key selftest (STATIC_KEYS_SELFTEST) [N/y/?] n
Static call selftest (STATIC_CALL_SELFTEST) [N/y/?] n
Enable seccomp to safely execute untrusted bytecode (SECCOMP) [Y/n/?] y
Show seccomp filter cache status in /proc/pid/seccomp_cache (SECCOMP_CACHE_DEBUG) [N/y/?] n
Stack Protector buffer overflow detection (STACKPROTECTOR) [Y/n/?] (NEW)
Strong Stack Protector (STACKPROTECTOR_STRONG) [Y/n/?] (NEW)
Link Time Optimization (LTO)
> 1. None (LTO_NONE)
choice[1]: 1
Number of bits to use for ASLR of mmap base address (ARCH_MMAP_RND_BITS) [28] 28
Number of bits to use for ASLR of mmap base address for compatible applications (ARCH_MMAP_RND_COMPAT_BITS) [8] 8
Provide system calls for 32-bit time_t (COMPAT_32BIT_TIME) [Y/n/?] y
Use a virtually-mapped stack (VMAP_STACK) [Y/n/?] y
Randomize kernel stack offset on syscall entry (RANDOMIZE_KSTACK_OFFSET_DEFAULT) [N/y/?] n
Locking event counts collection (LOCK_EVENT_COUNTS) [N/y/?] n
```https://git.adelielinux.org/adelie/horizon/-/issues/370efivarfs isn't managed correctly2023-10-22T02:26:21ZA. Wilcoxefivarfs isn't managed correctlyhscript/meta.cc:859:
```c++
if(mount("efivarfs", efipath.c_str(), "efivarfs", MS_NOEXEC |
MS_NODEV | MS_NOSUID | MS_RELATIME, nullptr) != 0) {
```
We need to use `MS_REMOUNT` if efivarfs is already mounted. ...hscript/meta.cc:859:
```c++
if(mount("efivarfs", efipath.c_str(), "efivarfs", MS_NOEXEC |
MS_NODEV | MS_NOSUID | MS_RELATIME, nullptr) != 0) {
```
We need to use `MS_REMOUNT` if efivarfs is already mounted. However, sometimes it isn't mounted, especially if Horizon is running in a chroot or from non-Adélie media. In fact, we _dropped_ `MS_REMOUNT` in bad8487c53 because it wasn't mounted at all on the ThinkPad being used for EFI testing.
What we really need to do is:
1. Check if `efipath` is a mount point or not. (Probably should check if it's even extant.)
2. If not, make the above mount call (no change).
3. If it is, use `mount_setattr(2)` to clear `MS_RDONLY`. Unfortunately, that is only available as a raw syscall.A. WilcoxA. Wilcoxhttps://git.adelielinux.org/adelie/packages/-/issues/1090user/{cbindgen,ripgrep,xsv,zola}: stray 'usr/.crates2.json'2023-10-24T06:24:07ZZach van Rijnuser/{cbindgen,ripgrep,xsv,zola}: stray 'usr/.crates2.json'Per #1089, some stray and conflicting files exist in our Rust packages.
The root cause is that they were built without the `cargo install --no-track` flag, which, when omitted, can create unwanted files.
From https://doc.rust-lang.org/...Per #1089, some stray and conflicting files exist in our Rust packages.
The root cause is that they were built without the `cargo install --no-track` flag, which, when omitted, can create unwanted files.
From https://doc.rust-lang.org/cargo/commands/cargo-install.html:
> `--no-track`
>
> By default, Cargo keeps track of the installed packages with a metadata file stored in the installation root directory. This flag tells Cargo not to use or create that file. With this flag, Cargo will refuse to overwrite any existing files unless the `--force` flag is used. This also disables Cargo’s ability to protect against multiple concurrent invocations of Cargo installing at the same time.
As an example, we also do not need to remove `user/.crates.toml` manually:
```diff
diff --git a/user/ripgrep/APKBUILD b/user/ripgrep/APKBUILD
index be18ce2e1..68b4c0732 100644
--- a/user/ripgrep/APKBUILD
+++ b/user/ripgrep/APKBUILD
@@ -124,8 +124,7 @@ check() {
package() {
export CARGO_HOME="$srcdir/cargo-home"
- cargo install --features pcre2 --path . --root="$pkgdir"/usr
- rm "$pkgdir"/usr/.crates.toml
+ cargo install --no-track --features pcre2 --path . --root="$pkgdir"/usr
}
_zshcomp() {
```
See also:
* https://wiki.archlinux.org/title/Rust_package_guidelines#Notes_about_using_cargo_installhttps://git.adelielinux.org/adelie/horizon/-/issues/368Partition Manager doesn't work in live desktop environments2023-10-24T06:52:13ZA. WilcoxPartition Manager doesn't work in live desktop environmentsThe partition manager does not launch under at least LXQt and MATE (have not tested KDE, kdesu5 may be working properly there). A single window pops up with all controls disabled, appearing to want to authenticate as root but being unab...The partition manager does not launch under at least LXQt and MATE (have not tested KDE, kdesu5 may be working properly there). A single window pops up with all controls disabled, appearing to want to authenticate as root but being unable to do so. `apk add kdesu` wasn't enough to make it work.0.9.8A. WilcoxA. Wilcoxhttps://git.adelielinux.org/adelie/packages/-/issues/1089[meta] find and fix conflicting files between all packages2023-12-02T02:20:31ZZach van Rijn[meta] find and fix conflicting files between all packages#1086 and #1088 are recent examples of documentation conflicts.
Create a script or tool that can scan a package repository and identify conflicts.
Possible starting point: `tar -tf /path/to/file.apk | grep -vE '(/$|^\.SIGN|^\.PKGINFO)'`.#1086 and #1088 are recent examples of documentation conflicts.
Create a script or tool that can scan a package repository and identify conflicts.
Possible starting point: `tar -tf /path/to/file.apk | grep -vE '(/$|^\.SIGN|^\.PKGINFO)'`.https://git.adelielinux.org/adelie/packages/-/issues/1088system/{gdb,binutils}: ERROR: gdb-doc-12.1-r0: trying to overwrite usr/share/...2023-10-24T06:24:07ZZach van Rijnsystem/{gdb,binutils}: ERROR: gdb-doc-12.1-r0: trying to overwrite usr/share/info/ctf-spec.info owned by binutils-doc-2.41-r0.Found on a real installation (`20231016` media) with tree as of fe2bdb57ff4f50f7ff7bb953342d92adfb78824a:
```
Adelie-E3C972 # apk add gdb
(1/2) Installing gdb (12.1-r0)
(2/2) Installing gdb-doc (12.1-r0)
ERROR: gdb-doc-12.1-r0: trying t...Found on a real installation (`20231016` media) with tree as of fe2bdb57ff4f50f7ff7bb953342d92adfb78824a:
```
Adelie-E3C972 # apk add gdb
(1/2) Installing gdb (12.1-r0)
(2/2) Installing gdb-doc (12.1-r0)
ERROR: gdb-doc-12.1-r0: trying to overwrite usr/share/info/ctf-spec.info owned by binutils-doc-2.41-r0.
Executing mandoc-1.14.6-r0.trigger
1 error; 3927 MiB in 1428 packages
```https://git.adelielinux.org/adelie/packages/-/issues/1087user/{systemsettings,kuserfeedback}: infinite loop / hang2023-10-30T13:40:06ZZach van Rijnuser/{systemsettings,kuserfeedback}: infinite loop / hangFound on an installed x86_64 system from `adelie-inst-x86_64-1.0-beta5-20231016`.
Navigate to this page and try to drag the slider above "Disabled":
![VirtualBox_test1_16_10_2023_17_29_55](/uploads/8c4d9c67843097f2f056cf01e34d8c64/Virt...Found on an installed x86_64 system from `adelie-inst-x86_64-1.0-beta5-20231016`.
Navigate to this page and try to drag the slider above "Disabled":
![VirtualBox_test1_16_10_2023_17_29_55](/uploads/8c4d9c67843097f2f056cf01e34d8c64/VirtualBox_test1_16_10_2023_17_29_55.png)
Now you have a frozen/laggy system.https://git.adelielinux.org/adelie/packages/-/issues/1086user/groff: ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/ma...2023-10-24T06:24:06ZZach van Rijnuser/groff: ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/man7/roff.7.gz owned by groff-doc-1.22.4-r0.Found using `adelie-inst-x86_64-1.0-beta5-20231016`:
1. Run through the installer to create `/etc/horizon/installfile` (see example from image#404)
2. Replace `x86-1` and `x86-2` with `x86_64/x86_64-1` and `x86_64/x86_64-2` respectively...Found using `adelie-inst-x86_64-1.0-beta5-20231016`:
1. Run through the installer to create `/etc/horizon/installfile` (see example from image#404)
2. Replace `x86-1` and `x86-2` with `x86_64/x86_64-1` and `x86_64/x86_64-2` respectively
3. Remove `libreoffice` and `thunderbird` from the package list, for now (unrelated issues)
4. Unmount `/target` by running this until there is no more output: `mount | grep /target | awk '{print $3}' | while read k; do umount $k; done`
5. Run `hscript-executor /etc/horizon/installfile` and wait
![VirtualBox_test1_16_10_2023_16_16_24](/uploads/b091623e723319ce4cc6883ad738b14d/VirtualBox_test1_16_10_2023_16_16_24.png)
![VirtualBox_test1_16_10_2023_16_25_09](/uploads/e8cc887da314abd3c47b652a9a8c0244/VirtualBox_test1_16_10_2023_16_25_09.png)
![VirtualBox_test1_16_10_2023_16_28_45](/uploads/e2558765a45ce4dc6122b9ee4beb6bdb/VirtualBox_test1_16_10_2023_16_28_45.png)
Minimal reproducer:
```
adelie # apk add mandoc-doc groff-doc
(1/2) Installing groff-doc (1.22.4-r0)
(2/2) Installing mandoc-doc (1.14.6-r0)
ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/man7/roff.7.gz owned by groff-doc-1.22.4-r0.
```https://git.adelielinux.org/adelie/horizon/-/issues/367user/groff: ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/ma...2023-10-16T21:54:38ZZach van Rijnuser/groff: ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/man7/roff.7.gz owned by groff-doc-1.22.4-r0.Found using `adelie-inst-x86_64-1.0-beta5-20231016`:
1. Run through the installer to create `/etc/horizon/installfile` (see example from image#404)
2. Replace `x86-1` and `x86-2` with `x86_64/x86_64-1` and `x86_64/x86_64-2` respectively...Found using `adelie-inst-x86_64-1.0-beta5-20231016`:
1. Run through the installer to create `/etc/horizon/installfile` (see example from image#404)
2. Replace `x86-1` and `x86-2` with `x86_64/x86_64-1` and `x86_64/x86_64-2` respectively
3. Remove `libreoffice` and `thunderbird` from the package list, for now (unrelated issues)
4. Unmount `/target` by running this until there is no more output: `mount | grep /target | awk '{print $3}' | while read k; do umount $k; done`
5. Run `hscript-executor /etc/horizon/installfile` and wait
![VirtualBox_test1_16_10_2023_16_16_24](/uploads/2381ed6edb3d218390d31454245c4f6e/VirtualBox_test1_16_10_2023_16_16_24.png)
![VirtualBox_test1_16_10_2023_16_25_09](/uploads/d443bcf1cf1ea7a48600e80ee2f0da6b/VirtualBox_test1_16_10_2023_16_25_09.png)
![VirtualBox_test1_16_10_2023_16_28_45](/uploads/652a564e7eb6c12b7937201ca0a2dfce/VirtualBox_test1_16_10_2023_16_28_45.png)
Minimal reproducer:
```
adelie # apk add mandoc-doc groff-doc
(1/2) Installing groff-doc (1.22.4-r0)
(2/2) Installing mandoc-doc (1.14.6-r0)
ERROR: mandoc-doc-1.14.6-r0: trying to overwrite usr/share/man/man7/roff.7.gz owned by groff-doc-1.22.4-r0.
```https://git.adelielinux.org/adelie/horizon/-/issues/365Web Browser tool: does not open2023-10-19T03:43:29ZZach van RijnWeb Browser tool: does not openThis button does nothing:
![VirtualBox_test1_16_10_2023_13_54_29](/uploads/77fbe6b7804fecbcd8588e892fb77d86/VirtualBox_test1_16_10_2023_13_54_29.png)This button does nothing:
![VirtualBox_test1_16_10_2023_13_54_29](/uploads/77fbe6b7804fecbcd8588e892fb77d86/VirtualBox_test1_16_10_2023_13_54_29.png)0.9.8https://git.adelielinux.org/adelie/horizon/-/issues/364adelie-inst-x86_64-1.0-beta5-20231016: UI does not redraw when Terminal is open2023-10-17T00:50:11ZZach van Rijnadelie-inst-x86_64-1.0-beta5-20231016: UI does not redraw when Terminal is openThe UI will not redraw when another window (e.g. Terminal) is open, even if it is minimized:
![VirtualBox_test1_16_10_2023_13_50_07](/uploads/58f31a8363bd06d12dbc69b5bc187d68/VirtualBox_test1_16_10_2023_13_50_07.png)
![VirtualBox_test1_...The UI will not redraw when another window (e.g. Terminal) is open, even if it is minimized:
![VirtualBox_test1_16_10_2023_13_50_07](/uploads/58f31a8363bd06d12dbc69b5bc187d68/VirtualBox_test1_16_10_2023_13_50_07.png)
![VirtualBox_test1_16_10_2023_13_50_22](/uploads/bb7b4a0474d5dd815446c3839811aa71/VirtualBox_test1_16_10_2023_13_50_22.png)
![VirtualBox_test1_16_10_2023_13_50_31](/uploads/e2613e6e9affa43ff75177950510f863/VirtualBox_test1_16_10_2023_13_50_31.png)
Additionally, if the window is dragged around, this happens:
![VirtualBox_test1_16_10_2023_13_50_39](/uploads/8673fb156187c79d06ff950f772c76f4/VirtualBox_test1_16_10_2023_13_50_39.png)
If the Terminal is re-focused (see #363) and closed, the UI works again.https://git.adelielinux.org/adelie/horizon/-/issues/363adelie-inst-x86_64-1.0-beta5-20231016: UI disappears when minimized2023-10-17T00:54:58ZZach van Rijnadelie-inst-x86_64-1.0-beta5-20231016: UI disappears when minimizedThe UI disappears when the minimize button is pressed:
![VirtualBox_test1_16_10_2023_13_47_13](/uploads/e3a8c7f7e1cc74cebac032a035c3e824/VirtualBox_test1_16_10_2023_13_47_13.png)
![VirtualBox_test1_16_10_2023_13_48_10](/uploads/3337a608...The UI disappears when the minimize button is pressed:
![VirtualBox_test1_16_10_2023_13_47_13](/uploads/e3a8c7f7e1cc74cebac032a035c3e824/VirtualBox_test1_16_10_2023_13_47_13.png)
![VirtualBox_test1_16_10_2023_13_48_10](/uploads/3337a608bf8d6c73ef72a9f6e09afc62/VirtualBox_test1_16_10_2023_13_48_10.png)
Workaround: `ALT` + `TAB`.
Suggest disabling these buttons in kiosk mode (the "inst" media).https://git.adelielinux.org/adelie/horizon/-/issues/362"Open Terminal" button does nothing2023-11-15T22:48:46ZZach van Rijn"Open Terminal" button does nothingIf installation fails, the "Open Terminal" button fails to launch:
![VirtualBox_test1_16_10_2023_13_41_17](/uploads/0da2cde0bd022608efec3318a61c808a/VirtualBox_test1_16_10_2023_13_41_17.png)
Clicking it does nothing (no visible errors)...If installation fails, the "Open Terminal" button fails to launch:
![VirtualBox_test1_16_10_2023_13_41_17](/uploads/0da2cde0bd022608efec3318a61c808a/VirtualBox_test1_16_10_2023_13_41_17.png)
Clicking it does nothing (no visible errors).
Workaround: `CTRL` + `SHIFT` + `F2` or similar.0.9.8https://git.adelielinux.org/adelie/image/-/issues/405adelie-inst-x86_64-1.0-beta5-20231016: "Open Terminal" button does nothing2023-10-16T18:46:29ZZach van Rijnadelie-inst-x86_64-1.0-beta5-20231016: "Open Terminal" button does nothingIf installation fails, the "Open Terminal" button fails to launch:
![VirtualBox_test1_16_10_2023_13_41_17](/uploads/e98b70f41ab7090ba3e0d82084ee82f9/VirtualBox_test1_16_10_2023_13_41_17.png)
Clicking it does nothing (no visible errors)...If installation fails, the "Open Terminal" button fails to launch:
![VirtualBox_test1_16_10_2023_13_41_17](/uploads/e98b70f41ab7090ba3e0d82084ee82f9/VirtualBox_test1_16_10_2023_13_41_17.png)
Clicking it does nothing (no visible errors).
Workaround: `CTRL` + `SHIFT` + `F2` or similar.https://git.adelielinux.org/adelie/image/-/issues/404adelie-inst-x86_64-1.0-beta5-20231016: UNTRUSTED signature2023-10-17T00:49:08ZZach van Rijnadelie-inst-x86_64-1.0-beta5-20231016: UNTRUSTED signatureThe installer is attempting to download nonexistent keys:
![VirtualBox_test1_16_10_2023_13_37_43](/uploads/45e4146e4580d5a14bfacd2ee676b89b/VirtualBox_test1_16_10_2023_13_37_43.png)
It should check for 404 status as well or verify that...The installer is attempting to download nonexistent keys:
![VirtualBox_test1_16_10_2023_13_37_43](/uploads/45e4146e4580d5a14bfacd2ee676b89b/VirtualBox_test1_16_10_2023_13_37_43.png)
It should check for 404 status as well or verify that keys are, in fact, available.https://git.adelielinux.org/adelie/image/-/issues/403post-binutils 2.41, post-grub 2.12, x86_64 media cannot boot2023-10-24T06:24:06ZZach van Rijnpost-binutils 2.41, post-grub 2.12, x86_64 media cannot bootUsing either of these media:
* http://foxkit.us/linux/adelie-live-kde-x86_64-1.0-beta5-foxexp.iso (2023-10-11)
* https://temp.adelielinux.org/adelie-live-kde-x86_64-1.0-beta5-20231012.iso (built with latest tree as of packages@c1026184b2...Using either of these media:
* http://foxkit.us/linux/adelie-live-kde-x86_64-1.0-beta5-foxexp.iso (2023-10-11)
* https://temp.adelielinux.org/adelie-live-kde-x86_64-1.0-beta5-20231012.iso (built with latest tree as of packages@c1026184b292497a9dd025c0138b26c391c37594 with all packages synced to distfiles except for pmmx and armv7 as of writing)
the VM fails to boot in VirtualBox, on metal, and in other hypervisors:
![image](/uploads/c296c839305d64753a434e51136e51e4/image.png)
```
17:34 < dimitsos> awilfox: There seems to be something wrong with foxexp x86_64 iso: I tried a laptop with amd cpu and graphics and one with intel cpu and nvidia graphics. In the first case, instead of grub menu I got stuck with a fuzzy screen. In the second case, I think it tried to show the grub menu, then it instantly rebooted.
...
18:07 < dimitsos> awilfox: 3rd attempt: Intel atom netbook, it gets stuck at "GRUB loading..."
...
00:48 <~awilfox> dimitsos: interesting. another tester reported failure with GRUB on iMac with UEFI
00:48 <~awilfox> that's not good
00:48 <~awilfox> these are the first media with the grub 2.12~rc1 update
00:49 <~awilfox> it does seem to work reliably on ppc/ppc64 and aarch64, which is progress because 2.04 didn't work on aarch64 at all and was shaky on ppc
00:49 <~awilfox> I guess now I need to debug it on x86_64
```
A log from VirtualBox (excluding parts in common with the successfully-booting `20230829` media):
* http://ix.io/4IQIhttps://git.adelielinux.org/adelie/packages/-/issues/1085user/mate-desktop: no font deps2023-10-26T05:54:07ZA. Wilcoxuser/mate-desktop: no font depsNo MATE packages depend on any fonts, so it is possible to install MATE without having any fonts on the system, which ends up rendering horrible boxes that make no sense (adelie/horizon#353 has an example screenshot).
Either mate-deskto...No MATE packages depend on any fonts, so it is possible to install MATE without having any fonts on the system, which ends up rendering horrible boxes that make no sense (adelie/horizon#353 has an example screenshot).
Either mate-desktop or mate-complete should depend on something like ttf-liberation or ttf-dejavu (+ otf-source-code-pro for nice monospace) or *something*. user/plasma-desktop (KDE) deps on ttf-liberation and ttf-noto, and the draft LXQt theme package (user/adelie-lxqt-theme) deps on ttf-liberation as well. The only problem is MATE seems to have a bit of trouble *using* the Liberation set of fonts, which I haven't poked at much yet.https://git.adelielinux.org/adelie/horizon/-/issues/361DHCP and network issue meta-ish bug/feature2023-10-24T06:52:13ZA. WilcoxDHCP and network issue meta-ish bug/featureWe have a lot of issues that are because of inadequate support of DHCP and network connections. There are a few ways forward.
Desktop images, those with KDE/LXQt/MATE (etc), should probably use NetworkManager and start NM on boot. I'v...We have a lot of issues that are because of inadequate support of DHCP and network connections. There are a few ways forward.
Desktop images, those with KDE/LXQt/MATE (etc), should probably use NetworkManager and start NM on boot. I've done some testing with aarch64 and ppc media and it works great. NM implements the logic of "determine what interfaces have a link", "try and bring up v4 and v6 and see if either have actual network connectivity, not just link-local", and so on, that we would otherwise need to reimplement ourselves.
The Horizon install image, that is `-inst`, may not need NetworkManager, but read on why I think it should probably do it anyway.
If we make dhcpcd setuid on live media only (*not* on real systems!), the security risk is low (it's a live system, nominally booted off a read-only CD) and allows the wizard to try DHCP without the wizard itself being setuid. This avoids us having to add the setuid-ok flag on the wizard which would then make it setuid-able even in other situations or on other platforms (or make a big ugly `#ifdef` tree).
When dhcpcd is setuid, and NM has already achieved a DHCP lease on the desired interface, dhcpcd is intelligent enough to see the nm management and return immediately with "lease already obtained". It's a great experience for the user because they aren't waiting any time for the interface to come up or obtain an IP; the Next button is nearly instantaneously available.
The other concern is that we really should be enabling chronyd on systems. It should be part of the default install, and perhaps even base/text-only installs. This is to support SoCs with no RTC, older machines with flat clock batteries, etc etc. I see no reason *not* to do this. And this is required to connect to the HTTPS servers in Horizon, hence why I think nm should be brought up on the `-inst` image as well. This way, chronyd can "start early" and fix up the clock before the wizard even begins.
So, to fix this bug, and all the other bugs related, this is my proposal:
* add `NetworkManager` to default run level of install images, to bring up network for live desktop;
* add `chronyd` to default run level of all images, even rootfs, to make sure RTC has correct time;
* setuid `dhcpcd` on live media, to ensure DHCP lease can be acquired by System Installation without setuid;
* setuid `horizon-run-qt5` so the `horizon-ui` script can be run as the `live` user.0.9.8A. WilcoxA. Wilcox