Syslinux - Aug 2018 - initramfs_init() changes in 6.04

Commit?73fb321a5d56b ("lua: make initramfs structures full userdata
objects with methods") broke my boot script, requiring a workaround
along the lines of the patch below:

Is this an intentional user-visible change? I note the documentation at
https://www.syslinux.org/wiki/index.php?title=Lua.c32#Example:_Print_kernel.2Finitramfs_information_and_load_them
hasn't been updated and still shows what to do for 6.03 without any
hint that it won't be compatible with 6.04.

Also... if we are doing this, then why do we need to pass the first
'initramfs' argument into the initramfs.load() function? Surely it
should have access to 'self' without the caller having to explicitly
pass it in? Isn't that kind of the point?

--- a/boot/syslinux.lua
+++ b/boot/syslinux.lua
@@ -87,8 +87,14 @@
?-- Booting the kernel
?
?local kernel = sl.loadfile("/BZIMAGE")
-local initramfs = sl.initramfs_init()
-sl.initramfs_load_archive(initramfs, "/INITRD")
+local initramfs
+if sl.initramfs_init ~= nil then
+??initramfs = sl.initramfs_init()
+??sl.initramfs_load_archive(initramfs, "/INITRD")
+else
+??initramfs = sl.initramfs()
+??initramfs.load(initramfs, "/INITRD")
+end
?
?local cmdline =??get_content("/CMDLINE")
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 5213 bytes
Desc: not available
URL:
<http://www.zytor.com/pipermail/syslinux/attachments/20180802/030ad700/attachment.bin>

David Woodhouse <dwmw2 at infradead.org> writes:
> Commit?73fb321a5d56b ("lua: make initramfs structures full userdata
> objects with methods") broke my boot script, requiring a workaround
> along the lines of the patch below:
>
> Is this an intentional user-visible change? I note the documentation at
>
https://www.syslinux.org/wiki/index.php?title=Lua.c32#Example:_Print_kernel.2Finitramfs_information_and_load_them
> hasn't been updated and still shows what to do for 6.03 without any
> hint that it won't be compatible with 6.04.
Hi!

Please accept my apologies.  This was an intentional API change, but it
wasn't intended to be user visible on the basis that nobody used it yet.
Frankly, I didn't expect anybody come after me almost four years after I
committed it...
> Also... if we are doing this, then why do we need to pass the first
> 'initramfs' argument into the initramfs.load() function? Surely it
> should have access to 'self' without the caller having to
explicitly
> pass it in? Isn't that kind of the point?
Yes it is.  You have to use the Lua method syntax to achieve that as
shown in com32/lua/doc/syslinux.asc:

local sl = require "syslinux"

kernel = sl.loadfile "/SuSE-11.1/x86_64/linux"
print("File name: " .. kernel:name() .. " size: " ..
kernel:size())

initrd1 = "/SuSE-11.1/x86_64/initrd"

initrd = sl.initramfs()
initrd:load(initrd1)
print ("File name: " .. initrd1 .. " size: " ..
initrd:size())

sl.boot_it(kernel, initrd, "init=/bin/bash")

Please refer to the in-tree documentation in preference to the online
one.
-- 
Regards,
Feri