Please fix this mess !
I took a quick look at the installer and it seems that this script is responsible for displaying the wrong partition sizes, but I might be wrong.
/sbin/probe: Code:
#!/bin/sh Cheers |
If that's what you mean, IIRC the unit in sizes displayed by fdisk has changed at some point in time. Should be easy to check and fix. I will investigate further in a few hours if nobody did since then
|
|
Thanks for the heads-up, Matteo. I will propose a patch later today, in this thread as I think it's better to put all suggestions in the same place.
PS Done. |
Quote:
Usually it is much easier. One just has to leach an ARCH or Gentoo patch and stick a copyright on top of the slackbuild :D Cheers |
Quote:
Quote:
Then, fixing this issue is not simple as you seem to think[1]. Trying to fix it then reading Pat's answer made me realize that a real fix would need to be verified in all situations considered in /sbin/probe, and that needs hardware that I personally do not own. [1]But if it is that simple, why didn't you propose a fix yourself? |
Looks like the problem is fixed. Here's today's ChangeLog.
Code:
usb-and-pxe-installers/usbboot.img: Rebuilt. Niki |
Patches get shared around lots of distributions. Add Red Hat, Fedora, LFS, and others to the vast resource of patchworks out there.
The sarcasm wasn't needed on any level and is a bit rude towards the Slackware team. |
Where does Ivandi think that Arch and Gentoo get their patches? Many of them are just cherry-picked commits from upstream.
One side of Ivandi's brain thinks copyright is a ridiculous barrier to this sharing (he's right, so we'll call that the "right" side of his brain), but the other side of his brain seems to think... well, forgive me if I'm wrong, but the implication seems to be that snarfing patches is somehow demeaning? Larry Wall disagrees. "The three chief virtues of a programmer are: Laziness, Impatience and Hubris." Sharing patches is normal and expected. One of the things Lucas Nussbaum said he wanted to work on after his Debian Project Leader stint was improving patch-sharing between distros. Also, acknowledging where patches originated is not merely polite, it is also an audit trail against the introduction of backdoors, and a lazy/impatient/hubristic way of referring other people to further information. I'd love to know where Ivandi thinks I found this, this and this. There is no copyright notice on those patches... I'd love to know what Ivandi thinks that means, under the Bern conventions. |
Hi,
While trying to compile Nixnote2, got a problem : With Code:
pkg-config --libs hunspell Code:
-lhunspell-1.3 Code:
-lhunspell Regards |
1 Attachment(s)
Quote:
|
Quote:
This function: Code:
# get_part_size( dev ) - Return the size in K, M, G, or T of the named partition. Is there any special reason for using expr instead of awk to do the math. For example this: Code:
get_part_size() { Cheers |
I am using numfmt in a script to get human readable sizes from a download speed provided by curl (allows me to test the speed of many mirrors and then find the fastest one available). From my limited scripting experience, it looks like what both awk and expr are doing, but with a much simpler command... although I'm not sure if it is included (or would be hard to include) with the Slackware installer (it's part of coreutils on a full Slackware install). This would require no changes to support larger drives (although, PB drives are still pretty far away) and would require no if/then/elif statements. Although, to be fair, I haven't looked through the script at all, and I haven't plugged in any numbers except for the four ivandi provided. It does only do one decimal place and a with quick look through the man page I didn't see anything to change it. Sorry if this is a stupid suggestion, I am still pretty noobish when it comes to the types of scripting that occurs in the installer.
Code:
get_part_size() { Code:
jbhansen@craven-moorhead:~$ numfmt --to=iec 2147483648 2147483647 7515144192 2000000000 |
Greetings. What about using bc as in the draft below? It won't truncate the decimal part and "scale" allows to set the needed number of decimal digits. I assumed from the original function that for sizes less than 1G decimals are not that important (1-1023K, 1-1023M, 1.0-1023.9G, 1.00T-etc.). Please don't take this post as if I'm trying to teach something to anybody: on the contrary, I'm just curious and interested in learning from you guys.
Code:
get_part_size() { Philip |
Quote:
|
All times are GMT -5. The time now is 03:25 PM. |