[parted-devel] A few patches for GNU fdisk

[Leslie P. Polzer]

[Parted guys: see the third section of this mail.]


> I fixed it, the patch is attached as check-fix.patch.

Seems to work nicely, thanks.


> There was a minor leak in my new code, I attach the fix as
> leak-fix.patch. The patch also fixes the indentation of some code
> sections, as I made the previous patches with the -b option, I omitted
> it this time to fix this.

This one doesn't seem to be against current:

---
patching file src/cfdisk.c
Hunk #1 FAILED at 182.
Hunk #3 FAILED at 1171.
Hunk #4 FAILED at 2174.
Hunk #5 FAILED at 2307.
Hunk #6 FAILED at 2424.
Hunk #7 FAILED at 2458.
Hunk #8 FAILED at 2495.
Hunk #9 FAILED at 2521.
Hunk #10 FAILED at 2547.
Hunk #11 FAILED at 2600.
10 out of 11 hunks FAILED -- saving rejects to file src/cfdisk.c.rej
patching file src/fdisk.c
Hunk #1 FAILED at 1011.
Hunk #2 FAILED at 1060.
2 out of 2 hunks FAILED -- saving rejects to file src/fdisk.c.rej
---


> I've discovered some strange behavior, which I believe is an error in
> libparted, but the necessary conditions are rather strange. I get a
> segmentation fault on exiting cfdisk under the following condtitions:
>
> 1. I have libreiserfs installed, but it's unsupported by parted version
> (progsreiserfs-0.3.1-rc8).

Yeah, unfortunately.  And progsreiserfs sucks.
It would be best to scrap it until someone provides libreiserfs glue.


> 2. I run cfdisk as unprivileged user and without specifying a device on
> the command line (so I'm asked to enter one). In this case cfdisk stores
> the user input into a buffer and then does a strdup on the buffer and
> passes it to ped_device_get as a parameter. I have no idea how this is
> connected to the problem, but it doesn't happen otherwise.

Neither do I.  CC'ed to GNU Parted people.

Thanks a lot for your contributions.

  Leslie


PS: Could you send stuff to bug-fdisk at gnu.org from now on?
    It will be archived there and people can see we got activity.  Thanks.