Re: [PATCH 1/4] scripts/run-qemu.sh: better error checks & formatting

[Demi Marie Obenour <demiobenour@gmail.com>]

[-- Attachment #1.1.1: Type: text/plain, Size: 2751 bytes --]

On 11/13/25 11:56, Alyssa Ross wrote:
> Demi Marie Obenour <demiobenour@gmail.com> writes:
> 
>> On 11/13/25 07:23, Alyssa Ross wrote:
>>> Demi Marie Obenour <demiobenour@gmail.com> writes:
>>>
>>>> No other functional change intended.
>>>>
>>>> Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
>>>> ---
>>>>  scripts/run-qemu.sh | 24 ++++++++++++++----------
>>>>  1 file changed, 14 insertions(+), 10 deletions(-)
>>>
>>> This is a lot of changes all in one, without individual explanation.
>>> I'd appreciate it if they could be broken up and explained, because I'm
>>> going to have to ask what the purpose of each change is.
>>>
>>>> diff --git a/scripts/run-qemu.sh b/scripts/run-qemu.sh
>>>> index 64fd29259ab108bc547cb7c74623ae9dc288b3b7..9c6c8193bbeba5916038c82d8f76992051719c19 100755
>>>> --- a/scripts/run-qemu.sh
>>>> +++ b/scripts/run-qemu.sh
>>>> @@ -1,11 +1,15 @@
>>>> -#!/bin/sh -ue
>>>> +#!/bin/sh --
>>>
>>> The idea here is to be robust against the script being invoked with
>>> argv[0] set to "-c" or something?
>>
>> Correct.
> 
> Seems a little paranoid but I suppose it doesn't hurt.  I'd take a
> treewide change that did this for every /bin/sh shebang.

Will do later.

>>>> +if [ -n ${ARCH+test} ]; then
>>>> +	ARCH=$(uname -m)
>>>> +fi
>>>>  
>>>> -case "${ARCH:="$(uname -m)"}" in
>>>> +case $ARCH in
>>>>  	aarch64)
>>>>  		machine=virt,accel=kvm:tcg,gic-version=3,iommu=smmuv3
>>>>  		;;
>>>
>>> Why is this better?
>>
>> If uname exits with a non-zero status, the script will exit rather
>> than continuing.
> 
> I see — makes sense.  And why is the check not just the following?
> 
> 	if [ -z "${ARCH-}" ]; then

That works just as well.

>>>> -			set -- "$@" -append "${append:+$append }$1"
>>>> +			set -- "$@" -append ${append:+"$append "}"$1"
>>>>  			i=$((i + 2))
>>>>  			shift
>>>>  			continue
>>>
>>> Don't understand this one.  We've gone from one set of quotes to two
>>> sequential ones.
>>
>> I remember reading that the first version might not conform to POSIX.
>> I'm not sure if it matters though.
> 
> POSIX says[1]:
> 
>> If a parameter expansion occurs inside double-quotes:
>>
>> * Pathname expansion shall not be performed on the results of the expansion.
>>
>> * Field splitting shall not be performed on the results of the expansion.
> 
> Seems desirable to me.
> 
> [1]: https://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_02

I recall reading about shells not properly handling
${append:+"$append"} properly when the whole expansion is itself quoted
The double quotes around "$append" prevent it from being expanded.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)

[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 7253 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]