diff options
Diffstat (limited to 'inxi.changelog')
| -rw-r--r-- | inxi.changelog | 283 |
1 files changed, 276 insertions, 7 deletions
diff --git a/inxi.changelog b/inxi.changelog index 56644c2..51a5202 100644 --- a/inxi.changelog +++ b/inxi.changelog @@ -1,4 +1,261 @@ ================================================================================ +Version: 3.3.35 +Patch: 00 +Date: 2024-06-18 +-------------------------------------------------------------------------------- +RELEASE NOTES: +-------------------------------------------------------------------------------- + +Added more Wayland wm/de updates: wm, tools, lockers, etc. This list will grow +slowly over time, I'm sure. A new but very old distro, T2 SDE added, for distro +ID, mine package manager. + +Adding support for something as uncommon as T2 has a primary value that such +systems, particularly very old ones, tend to expose certain corner case data +alternate syntaxes (clang --version, /etc/issue), which then makes inxi overall +more robust. + +For example, because T2 used /etc/issue to show an ascii login, the actual first +line was just two literal lines, so the distro name appeared as two lines. This +was because there was an assumption that the first line of /etc/issue would +always be the distro name. Added filters to not use issue value if it does not +contain at least 3 characters. + +Also added in premilinary Loongson CPU/GPU support, but for GPU, have only 1 +product ID to match, and no other meaningful data. CPU may work, but needs real +hardware based testing to confirm. Obviously I can't do it, so if any Loongson +users want to test and report, feel free. Or not, up to you. Framework for full +support it there now. + +-------------------------------------------------------------------------------- +SPECIAL THANKS: + +1. mrmazda, as is often the case, for tracking and spotting issues with graphics +and desktop data. He also noticed the weird output for urpm repos, which is now +corrected, as well as finding some errors with rpm package counts, which led to +a small refactor of the rpm logic in PackageData. + +2. GRAPHICS: gl_data(): gfxstrand on issue #303 for figuring out what was +breaking OpenGL Mesa driver version detection. (was extra (git-..commit-ID) at +end of version string). + +3. DistroData: codeberg user r1w1s1 in issue #304 for pointing out that the '+' +part of the distro name can be meaningful, as in Slackware 15.0+, where the + +means it's Slackware current. + +-------------------------------------------------------------------------------- +KNOWN ISSUES: + +1. GRAPHICS: GPU device IDs: Intel device ID sources are not being reliable, and +some IDs vanished from the raw data files. This is too hard to work around, so +leaving as is. In a sense, Intel is just getting too hard to support, though +I'll keep trying. But it's not cute anymore. + +This applies to Intel CPU Ids too, for some reason, Intel just can't seem to +manage to create a new CPU family ID, and so are actually running out of their +256 Family 6 model IDs (2^8, or 16x16 hex). This is bizarrre to see such a huge +corporation fail to do something so utterly trivial, and suggests that there may +still be internal engineering issues lingering despite the recent attemps to +resolve them. + +2. PACKAGES: RPM remains unacceptably slow to get a simple package count query, +even with --nodigest / --nosignature, on one fedora 39 took 7 seconds, and in +mageia/fedora takes 0.35-0.80 seconds without those flags, which is easily 10x +slower than most package managers. There is nothing I can do to fix this that I +am aware of. If anyone knows of other options I can pass rpm to get it to list +the packages in a sane timeframe, let me know. + +During previous testing, nodigest/signature did speed it up, but that seems to +have regressed now again in one fedora test case, but does work in most others. +For now, --rpm flag will be required to see rpm package counts for core rpm +distros, otherwise people will blame inxi for being slow. + +-------------------------------------------------------------------------------- +BUGS: + +1. Nothing to speak of beyond some fixes. + +-------------------------------------------------------------------------------- +FIXES: + +1a. REPOS: For slpkg, adapted to new syntax for each repo. Retained support for +older syntaxes. + +1b. REPOS: Fixed output for urpm, pisi, and eopkg. These all had the same error, +which was creating a new full repo type line per repo listed by these software +tools. Now shows as expected: + +[repo name] ~ [repo location] + +2. main::set_path: added hack to put back in /opt/trinity/bin if distro forgot +to add it to PATH. + +3a. SYSTEM: DesktopData: Added fallback test for twin for Trinity version +detections. This goes along with path fix for this extreme corner case (so far +mageia only one who didn't add trinity location to PATH). + +3b. SYSTEM: DistroData: Added filter in case line 1 of /etc/issue doesn't +contain any valid characters. That's for T2 failure. + +3c. SYSTEM: DistroData: main::clean_characters(): removed filter for '+' because +that can be a meaningful part of the name, as in r1w1s1's codeberg issue #304, +where the + indicates it's Slackware 15.0 Current, aka: Slackware 15.0+. + +4. GRAPHICS: gl_data(): Issue #303 exposed a possible syntax that inxi had not +seen or handled before, the OpenGL core profile version string ending with a git +commit ID in parens: +OpenGL core profile version: 4.6 (Core Profile) Mesa 24.2.0-devel (git-f0620dae) +as opposed to handled and expected: +OpenGL core profile version: 4.6 (Core Profile) Mesa 24.2.0-devel + +This just required adding another 0 or 1 occurance test for that ending git +commit string. Really for anything tacked on to the end of the string after the +mesa / nvidia driver version value. + +5. PACKAGES: Added support for Mageia's urpmi package counts, using rpm. That's +either not worked, or broke recently, I don't know which. + +Note that rpm -qa is unacceptably slow (even using --nodigest --nosignature), so +requires use of --rpm flag for primarily rpm distros. Otherwise they will think +it's inxi that is slow. + +Just to document it, mageia also says they may move to dnf in the future. But +that should not matter with the redone logic. +https://wiki.mageia.org/en/Ways_to_install_programs#.deb_packages + +6. INFO: compilers: made clang version detection more robust. + +-------------------------------------------------------------------------------- +ENHANCEMENTS: + +1. SYSTEM: DesktopData: Added the following items: + +* menus/launchers: wayland gmenu, onagre, ulauncher, wmenu. + +* greeters: slick-greeter/elephant-greeter lightdm, but those are not the same +as greetd/seatd type greeters so not using, but leaving in comments. + +* bars: hybridbar. + +* x11 wm: MaXX. + +* wayland compositors: MagmaWM, newm [already abandoned], newm-atha [fork of +newm], niri, Pinnacle. + +2. INFO: PowerData: Added waylock + +3a. GRAPHICS: new AMD, Intel, Nvidia GPU IDs. + +4a. GRAPHICS: loongson GPU arch data. Note this data is extremely incomplete, +but at least it's got the framework to support Loongson GPUs now. Hard to find +this data in any reliable format beyond process nodes for each CPU, and since +most GPUs are iGPUs, in theory those should be the same. Maybe some Chinese +users will decide to provide this info. + +Also see CODE 4b for the tools for this item. + +5. PACKAGES: added package manager type mine. That's used in T2, maybe others. +Also added tools gastone, gasgui. + +6. REPOS: added basic T2 svn repo URL handler. Not robust, but that's fine. + +7. CPU: added Loongon CPU arch support, that should work reasonably well out of +the box since also found sample set of cpuinfo files. + +Also see CODE 4d for the tool for this item. + +-------------------------------------------------------------------------------- +CHANGES: + +1. REPOS: See FIX 4. Output was silly for repo reports, that's always been +broken, now correctly shows expected (for urpm, eopkg, pisi): + +Active urpm repos: + 1: [repo-name] ~ [repo location] + 2: [repo-name] ~ [repo location] + +lines per package manager tool, instead of one full repo line per repo in the +output: + +Active urpm repo: Stable + 1: [repo location] +Active urpm repo: Testing + 1: [repo location] + +2: MAIN::clean_characters(): removed filter for '+' because that can be part of +the distro name, like Slackware 15.0+ (current, that is). Hopefully this won't +create any issues for other name strings, but it should be ok. + +-------------------------------------------------------------------------------- +DOCUMENTATION: + +1a. DOCS: inxi-desktop-wm.txt, updated for new wm, bars, LM, launcher/menus. + +1b. DOCS: inxi-power.txt: added waylock to lockers sectsion. + +1c. DOCS: inxi-repo-package-manager.txt: updated for 3 syntaxes of slpkg, with +samples. Also updated for mageia urpm, which has special package list rpm +requirements. + +1d. DOCS: inxi-cpu.txt: added more links and sources for Loongson. + +1e. DOCS: inxi-distros.txt: refactored a bit, added top distro names to section +lists, added T2 instructions. + +-------------------------------------------------------------------------------- +CODE: + +1a. MAIN: ProgramData: + +* Had redundant busybox hash key data array, first one was replaced by explicit +busybox ash version logic in shell, and running one is for init data. + +I don't actually know what Perl does when you give it a second case of same key +name for a hash. Probably worth finding out. But it worked in testing. + +* Cleaned up, made consistent ', # unverified' program data key array comments +to make them easier to find/ Ideally one day we'll get people to help figure out +these unverified program version data items. Now to find a way to motivate +them... + +1b. MAIN: PackageData: Added --dbg 67 to make rpm packages handling debuggable. +Also refactored a bit, again to make rpm stuff debuggable. + +1c. MAIN: clean_characters: refactored, now uses references, not copies, and +removed '+' from filter. This can be meaningful part of distro name. + +2. REPOS: refactored completely slpkg to handle 3 possible config file syntax, +including the 2024-04-19 update which changes it completely. + +It's tempting to apply this to all multi-line repo syntax handlers, though the +cost is losing the order the repos appear in the file, since the data now goes +into a hash, which Perl does not have ordering, so it has to be sorted on +constructing of repo line arrays. + +3. DEBUGGER: Added kdesktop, kwin_wayland, twin --version data, might be useful. +kwin_wayland was an oversight, already had kwin_x11 so just never added wl. + +4a: TOOLS: made top comment header syntax consistent, added some GPL to ones +where I'd forgotten to do that. Updated copyright years. Also made empty line +use consistent in all the tool files. + +4b. TOOLS: gpu_raw.pl: Got rid of manually updated list of nv numeric keys, and +now extract it it automatically, so updates by itself when I add a new nv driver +series. Should have made it that way from start, but better late than never. +Always trying to chip away at anything manual I have to update for these +matching tables, makes maintaining the features easier. + +4c. TOOLS: gpu_raw.pl, gpu_ids.pl: added premilimary support for Loongson GPUs. +Very little data, but did at least get some device IDs for Loongson, but not +very reliable or accurate. This will need updates in future. Anyone Chinese +users want to help? + +4d. TOOLS: cpu_arch.pl: added Loongon support, first new CPU arch in a while! + +-------------------------------------------------------------------------------- +-- Harald Hope - Tue, 18 June 2024 13:16:38 -0700 + +================================================================================ Version: 3.3.34 Patch: 00 Date: 2024-04-13 @@ -10,7 +267,7 @@ This is a nice point release that corrects and updates numerous smaller, and larger, features. Most pressing in terms of numbers affected was the failure to detect in display, non console, for Wayland without Xwayland running. But updates will impact a fairly wide range of systems, though mostly are corner -case issues, but inxi is all about handling corner cases to make the handliing +case issues, but inxi is all about handling corner cases to make the handling of core stuff more robust and less prone to failure or error. -------------------------------------------------------------------------------- @@ -65,7 +322,7 @@ culprit was forgetting the 'starts with ^ in: Thanks Codebergy issue poster thatso in issue #299 for pointing this out. 2. NETWORK: services: was missing nfsd because it's a kernel process, that's -corrected. Showed as: [nfsd] +corrected. Shows as: [nfsd], so its brackets have to be stripped off first. 3. MAIN: primary DISPLAY test fails to check for WAYLAND_DISPLAY and other desktop session environmental variables only present if X/Wayland are running @@ -111,12 +368,19 @@ deepin-kwin_x11. Also for deepin-kwin_wayland. in kernel yet but may be. Also added in various filter lists just in case. Search for'nouveau' to find occurances. -4. CPU: Added more AMD, Intel CPU Ids. The Intel stuff is getting really +4a. CPU: Added more AMD, Intel CPU Ids. The Intel stuff is getting really difficult to track, they refuse to adopt sane family updates, keeping it always -at 6, and thus cram more and more into finite 2 hexadecimal numbers, leaving -marketing terms in place of actual die process engineering. There's only so much -I can do to work around this, so if you find errors, do the research, and submit -corrections, keeping in mind these things are super fuzzy. +at 6, and thus cram more and more into finite 2 hexadecimal model numbers, +leaving marketing terms in place of actual die process engineering. There's only +so much I can do to work around this, so if you find errors, do the research, +and submit corrections, keeping in mind these things are super fuzzy. + +4b. CPU: Added premilinary support for Loongson CPUs. Using regex to detect +arch/process info, but not positive how it will handle MIPS tests. No way to +fully test without running inxi on an actual Loongson system due to the early +MIPS/ARM type tests inxi runs. We'll see. So far only testing with fake cpuinfo +files. But it's a start. Any actual Loongson user should test and supply +datasets if gpu/cpu info is off. 5. DRIVES: More disk vendors, the endless list. Unhandled list grows slowly, maybe one day someone will step in and figure out what more of them are, @@ -154,6 +418,11 @@ shifting empty index 0 from @row to make values same as top sections for main ZFS Array data. The regex was simply too convoluted to work reliably, so it failed under certain cases, FreeBSD Linear ZFS mode for example. +3. CPU: Added data/cpu/loongson cpuinfo data files; added loongson support to +tools/cpu_arch.pl. We'll see if it works, won't know for a while since needs +a hardware run loongson test, this is just theoretical, but fixes should be +minor in case it doesn't work. + -------------------------------------------------------------------------------- -- Harald Hope - Sat, 13 March 2024 19:00:08 -0800 |