Your message dated Mon, 28 Nov 2016 21:36:07 +0100
with message-id <20161128203607.d25ndgwndlntz...@home.ouaza.com>
and subject line Closing bug
has caused the Debian Bug report #774652,
regarding Improve installer options & option related code
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ow...@bugs.debian.org
immediately.)
--
774652: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=774652
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--- Begin Message ---
Package: live-build
Version: 5.0~a2-1
Having twice recently tried incorrectly but logically using 'none' for
the installer type instead of 'false', resulting in an error suggesting
not having an installer isn't currently supported, I set about making
some improvements to the installer options.
The attached set of patches: replace 'false' with 'none'; remove the
'true' option (which digging through the code revealed was an alias for
netinst); fix the incomplete aliasing of netboot = netinst; then remove
the unnecessary netinst alias; expand the list of options presented in
the cgi frontend; and apply some other misc code cleanup improvements.
There are four issues outstanding however, for which feedback is
required before I can tackle them:
1) If the user selects a hdd type image, and specifically opts for the
netboot installer, that is what they get, otherwise it defaults to the
cdrom installer. What about the 'hd-media' installer directory available
on the mirrors. Is this not the correct installer for hdd media images?
If it is, the file sizes are different, so it's certainly different to
the cdrom and thus probably a good idea to switch to by default for hdd
images.
So I just need to know whether to go ahead and make this patch or not.
2) The udeb include file copied to .disk/udeb_include; Is this
informational only, or is it actually used by an installer to determine
the udebs to use? The file copied is a fixed list which must be manually
maintained. It would be very easy to create a list dynamically from when
applying exclude filters to the full list, which is already being done.
If the file is informational only, surely this dynamically created list
would suffice? If used by the installer, again, would such a dynamically
generated list not suffice?
So can we get rid of the manually maintained udeb include lists and add
a tiny hack to generate this file dynamically?
3) The 'businesscard' type is in most ways an alias for netboot/netinst,
but unlike all other types, it does not create a
'.disk/base_installable' file in the binary disk. I cannot tell if this
is a mistake or deliberate...
4) If the file in #2 can be generated dynamically as described above,
and the lack of a file in #3 is a mistake, then the 'businesscard' type
is pointless as it will become a perfect alias for netboot/netinst, and
thus can be removed. I need answers to #2 and #3 before a decision can
be made on whether or not to remove it.
improve-installer-options.tar.gz
Description: GNU Zip compressed data
--- End Message ---
--- Begin Message ---
Hello,
I reviewed the patches in this bug report and they are all
only refactoring patches, some of them even break backwards compat by
changing the options accepted on the command line.
I'm not interested in such patches right now, they might introduce subtle
bugs and I have not found that the updated code was of much higher quality
than the initial one.
So I'm closing this ticket.
Cheers,
--
Raphaël Hertzog ◈ Debian Developer
Support Debian LTS: http://www.freexian.com/services/debian-lts.html
Learn to master Debian: http://debian-handbook.info/get/
--- End Message ---
