[EN] inxi bugs or failures - please report

[h2]

i'm doing a big cleanup of inxi and have found a lot of failures, changed paths, changed or new syntax for things, I'm fixing them as I go along, but let me know if you have any data where something fails to get reported, for example, before inxi 2.1.4, lspci output for graphics cards that used: 3D controller OR Display Controller were not getting handled so those cards would go non recognized.

I found these failures by chance while checking some inxi stuff on the web, and also found and fixed some other bugs, like -d not reporting the optical device name/brand and rev number. That should now be fixed, the /sys/ path had changed, easy to fix with xiin output.

So if you see something fail, get missed, etc, let me know and I'll get it updated.

For those who care about such things, inxi help now wraps dynamically to fit the size of the terminal, takes a bit more time to show help, but it will fit whatever size terminal you use, no matter how narrow. Max width is a global set at 115 columns, but you can change that in /etc/inxi.conf if you want.

I'm aware of certain drivers reporting as FAILED when they aren't, but I believe that is an issue in Xorg.0.log, but if you have a case where Xorg.0.log reports successful load after it reports fail, then post the log and I should be able to resolve that issue. I believe it's a bug however somewhere that is not inxi, but it could also be some change in nourveau, which is where I see the issue most.

[vilde]

I didn't know that inxi could give so much information, I have only used -3 before. Now I checked the -help and found out

Didn't know that inxi was from h2 either, long time since I heard anything from you. I have no info about bugs and failures but anyway tanks for inxi h2

[ayla]

update to 2.1.7 came in today, checked the manpage and found option -v7.

Simply nice.

Need a chip ID? Internal or external IP? The uuid/label/size of a certain device? Drivers in use? .....

No need to search through /dev, /etc or remember a lot of commands, just ask inxi.
Detailed, fine ordered overview.

Thanks for that
ayla

[mylo]

h2 is back! Did not know up to now that he is behind inxi.
Thansk for that tool!

[h2]

inxi 2.1.10 now features more dynamic line wraps, with fixes for some of the ones that didn't work right. -C works better now for example with > 2 cpu core speeds listed, though the single core speed is not wrapping fully et at very narrow widths, ie, 80 columns.

--recommends and -c 94-99, the color selector tool, now wrap well.

Some error messages for missing data etc is now working better for 80 column widths.

And of course, what everyone has been waiting for, inxi now supports showing supybot/limnoria/gribble version information if inxi is started via irc client with shell command, but don't ask me how to do it, all I know is it works. I guess that lets you get data from the host server system or something, not sure, someone wanted it, so there it is.

-G wraps a lot better now too. -D -d -N and -A still aren't fully wrapping. -p/-P now I think should wrap ok unless the mount point is a very long path, but it's too hard to get those wraps perfect so I'm calling it good enough. It's all a lot better than it was so I"m not worried about perfection, but each line takes a lot of testing to get to wrap right with all data combinations.

Sometimes there may appear to be a blank line, but if you highlight the console you'll see that actually it was a forced wrap of a trailing space, which I am having a very hard time getting rid of internally, but that issue  usually only appears at exactly 80 columns, if the line is 79 columns wide including the trailing space/character.

[melmarker]

2.1.10, revision 2242 is now in the repo - thanks h2

[h2]

And with 2.1.12 I think I'll have to slow it way down. It fixes another line wrapping issue, -D, which is now ok.

Also fixed is zfs raid issue on some systems, but that's mostly going to be bsd so probably doesn't matter to linux users.

Also fixed is an old bug where if you use -c 0 and have a FAILED driver in G your console turns all red, forgot to unset the red variable.

[cryptosteve]

Thansk for that tool!

Definitely +1 

[melmarker]

uploaded 2.1.12

[dibl]

Very nice work -- thanks for this, h2!

[h2]

I think this is pretty much it for the line wrapping, at least this is as much as I will do:

2.1.15 now fixes wraps on -d, -n,-i,-N, and -A

There are some negatives to the tight wrapping at 80 columns, because if you have multiple cards or whatever one line may wrap and another not depending on their lengths, but that's the cost of doing true dynamic line wraps.

Except for bug fixes, this is now as done as it's going to get I hope anyway.

Things should be wrapping well down to 80 columns now, with some exceptions where the data is just too long, for example long card names, or certain data sets where the total for a second line still is greater than 80 columns, but it's crazy to do more in my opinion. It could be well argued that it was already crazy to do this much, but that's life, we have to waste our time here on earth somehow....

[h2]

And with 2.1.17 I can see the end in sight. This introduces the obvious option of -y <number >= 80> which lets you override on a one time basis all width settings. Useful for internal debugger use, that way I can set the desired line width in inxi output, and I can test the widths easily without physically resizing my terminal, and people can spit out any width they want as long as it is greater than 80. Width is in columns, that's how console/terminal width is set. IRC is also overridden, which can be useful if you want to send a narrow width to make sure the stuff doesn't wrap wrong.

Also updated man page and cleaned it up.

Usually when I cleanup the man page, it means the main project is pretty much close to done, the last step usually is updating the wiki pages if changes are required.

-y has some restrictions, it won't work with long options like --help or --recommends, but short ones are all fine, like -h.

-y also only works on the line version of inxi, not the short output version of: inxi
which always just fits the screen width no matter what. Put -y first on the list of options to make sure it works as intended, though usually it's fine, except for -h and -c stuff, where the script exits after running that option.

Now off to do real work, heh.

[der_bud]

Very nice work -- thanks for this, h2!

+1
No bug or failure, don't know if this is the right place for suggestions. Inxi gives so many useful information, would be nice to see in its output where grub currently is installed.
Sometimes when new grub versions are updated one gets a debconf dialog to choose device or partition, obviously one should use the current location again but on some multidisk/multiboot-setups it is hard to remember what the current setting was. Last time I probed my devices with

Code: [Select]

dd bs=512 count=1 if=/dev/sda 2>/dev/null | strings
that gives amongst others the word 'GRUB' as output for the device if grub is found. I am sure you could find a way to let inxi say something like "Grub currently on: /dev/sdX" (and hopefully you do not run out of letters for options).

[h2]

With a big groan, I've now completed all the lines, 2.1.18

I also fixed a weird decision I'd made a long time ago to not show the /dev - remote location of the partitions unless labels or uuid was used.

I needed to get this finished so I can move on to actual work projects, so now it's pretty much completely done.

Note that -p/-P, -D, and -o, now show ID-<number>: instead of ID: or 1: id:, ie, they are all the same now. I find it easier to read if they are numbered like that with a lot of partitions or drives.

Try 2.1.18 and see, all the lines more or less will wrap right down to 80 columns, except sometimes with -p and a few times with -S when the kernel name is super long. Maybe -R too isn't right in some cases, but I don't have access to an md-raid system at the moment to test that.

Further fixes are too hard in terms of the length being totally perfect so I call this done.

dd bs=512 count=1 if=/dev/sda 2>/dev/null | strings

requires root, I'm usually not very fond of implementing core features that require root to run, that's why -m memory is not started yet for example.

I agree it's a nice to have feature but it's not as easy as you think, grub can be on a partition boot sector, and I have no idea about EFI stuff and grub, as usual with ideas related to computers and linux in particular, the actual implementation where it actually works for almost all users would take probably 100 times longer than it takes to run a simple one liner on a system where all the data is known. At least 100 times I'd guess.

[melmarker]

inxi in our repos is now 2.1.18 in our repos, thanks h2

And now to something completly different: h2, is it possible to forbid inxi -h output for irc use? (to be true - this is disappointing )