Syslinux - Mar 2016 - Module Versioning... and other things

On 3/8/2016 07:23, Pete Batard via Syslinux wrote:
> [...] the merging of ldlinux.sys and isolinux.bin, which I would very 
> much like to rally people on this list into seeing as beneficial. [...]
My understanding is that isolinux.bin and ldlinux.bin are pretty much 
twins[1].  If I understand correctly, you'd like the latter to be tacked 
onto the former because people are too stupid to put both on the ISOs 
that they are too stupid to anticipate might have USB disks derived 
from.  Is that a rough understanding of the cause that you are rallying 
for?  - Shao

[1] There are obviously differences, including that the former has extra 
optical-disc-boot goodness

On 2016.03.08 18:28, Shao Miller via Syslinux wrote:
> My understanding is that isolinux.bin and ldlinux.bin are pretty much
> twins[1].  If I understand correctly, you'd like the latter to be
tacked
> onto the former because people are too stupid to put both on the ISOs
> that they are too stupid to anticipate might have USB disks derived
> from.  Is that a rough understanding of the cause that you are rallying
> for?  - Shao
Well, I wouldn't call people stupid for doing something that probably 
used to make some sense when the best you had was the capacity of one 
measly CD where every last KB counted, and then historically continued 
to use the same set of script (because why change something that works?).

But in essence, yes, that's pretty much it.

In the current landscape, people will be trying to use a Syslinux 
bootable ISO (which may not be ISOHybrid, or which they may prefer not 
to use as an ISOHybrid, so that they can still access their drive after 
applying it) to boot an USB, and, in my project-external opinion, I 
think it should fall on Syslinux to both acknowledge and commit to 
facilitate that, especially as, with the uber-reliance some maitainers 
seem to have on ISOHybrid (Looking at you openSUSE), I don't believe you 
can or should rely on distro maintainers to do that themselves.

Furthermore, while the complexity of the effort of doing so in Syslinux 
can be estimated, the effort of contacting all the Linux (and non-Linux) 
Isolinux based communities out there to ask them to do so is hard to 
estimate in itself. Yet, from my current understanding of the project, I 
suspect it will be a much larger amount of effort to contact all those 
people and get them to agree, than modifying Syslinux as I advocate. So, 
as far as I'm concerned, if the general answer from Syslinux is "We 
don't really want to do it ourselves - You should just contact distro 
maintainers to request they include ldlinux.sys", you bet that the first 
thing I'll be doing, when I find the time, is craft a patch that does 
just what I propose, rather than bother trying to e-mail hundreds of 
projects.

Finally I'm pretty sure that, were I to ask on Linux distros mailing 
lists to ensure that they also include 'ldlinux.sys', I'd probably
get
some replies in the vein of:

"You want us to include a Syslinux file, that isn't explicitly needed, 
because some Windows users are too stupid to understand what happens to 
their flash drive when they write an ISOHybrid in DD-mode?"

So, while I appreciate the honest opinion, I think we may want to try to 
consider this from the standpoint of what people might legitimately 
_choose_ to do, without resorting to calling them stupid for doing so.

Regards,

/Pete

On 3/8/2016 14:26, Pete Batard via Syslinux wrote:
> [...]
> But in essence, yes, that's pretty much it.
So then some thoughts:

 1. Did you mean a truly "merged" file, where the very same file could
    be used for ISOLINUX and for disk-based Syslinux?  Or did you mean
    "tacked on," where ldlinux.bin would be appended to isolinux.bin
so
    that you could later split them apart?
 2. What about people who are too stupid to anticipate their ISOs being
    converted to PXELINUX?  Do we tack on PXELINUX stuff to
    isolinux.bin?  There have been _plenty_ of times when I wanted a
    simple way to go from ISO to PXE.  Unfortunately, few tools that
    I've found cover that scenario as well as Rufus covers the ISO ->
    disk scenario.
 3. In your experience, are people picky about which .c32 modules they
    include, or do they just dump the whole lot in?  I ask because
    perhaps we could sneak ldlinux.bin into a .c32 module, which would
    [thankfully] leave isolinux.bin alone.  Such a module could be a
    trivial xtrafile.c32 program that simply lists the names of its
    embedded files, for example.
 4. If is was tremendously easy to find a disk-based Syslinux to work
    with .c32s you find in an ISO, would you still advocate merging
    isolinux.bin and ldlinux.bin, or is that the only reason to care
    about having an ldlinux.bin?

> [...]
>
> So, while I appreciate the honest opinion, I think we may want to try 
> to consider this from the standpoint of what people might legitimately 
> _choose_ to do, without resorting to calling them stupid for doing so.
I'm sorry that it wasn't obvious that "stupid" was used for
brevity's
sake.  No opinion about people actually being stupid was intended.  The 
e-mail was simply asking for confirmation about a rough understanding.

- Shao Miller

P. S. I hope you've observed another Rufus-specific e-mail regarding 
Rufus and Syslinux version concerns. :)