Re: ash.c history loading and ENABLE_FEATURE_EDITING_SAVEHISTORY

[Ceriel Jacobs]

The use case is not so "very rare" and is used within the 
Relax-and-Recover (rear) project. Albeit bash is used there.


In case ENABLE_FEATURE_EDITING_HISTORY would be split in:
ENABLE_FEATURE_EDITING_LOADHISTORY
ENABLE_FEATURE_EDITING_SAVEHISTORY

then there is more purpose in having
ENABLE_FEATURE_EDITING_SAVEHISTORY depend upon 
ENABLE_FEATURE_EDITING_LOADHISTORY

then LOAD depends on SAVE.

Thank you for the re-written shell function hint.


Where is the user interface to quickly issue a pre-written shell 
function? And how to see the available functions?



Op 24-12-17 om 02:26 schreef Kang-Che Sung:



2017年12月24日 09:17，"Ceriel Jacobs" <mailto:linux-...@crashplan.pro>>寫道：


There is no need to remember what was just typed.
There is a need to use commands that were typed a long time (years) ago.


Disasters often happen a long time after the backup (script/system)
was made.


The point is that you don't have to:
1. remember the commands
2. type the commands
3. correct typing mistakes


I know that I can save .ash_history
but it seems I can't load it without feature
ENABLE_FEATURE_EDITING_SAVEHISTORY set.


And convincing Arch Linux developers to turn on the
ENABLE_FEATURE_EDITING_SAVEHISTORY feature will be difficult.
Because in their initramfs environment they don't want to have a
SAVEHISTORY to file functionality.


I'm not convinced by this. Since your use case is a very rare one 
(read-only "history" somehow defeats the purpose of the history - and it 
overlaps what pre-written shell function is for). While I'm not one who 
can make decisions on this, I'd suggest you just keep your local patch 
to do this.



___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox

Re: ash.c history loading and ENABLE_FEATURE_EDITING_SAVEHISTORY

[Ceriel Jacobs]

There is no need to remember what was just typed.
There is a need to use commands that were typed a long time (years) ago.


Disasters often happen a long time after the backup (script/system) was 
made.



The point is that you don't have to:
1. remember the commands
2. type the commands
3. correct typing mistakes


I know that I can save .ash_history
but it seems I can't load it without feature 
ENABLE_FEATURE_EDITING_SAVEHISTORY set.



And convincing Arch Linux developers to turn on the 
ENABLE_FEATURE_EDITING_SAVEHISTORY feature will be difficult. Because in 
their initramfs environment they don't want to have a SAVEHISTORY to 
file functionality.



On 24-12-17 at 01:10 Kang-Che Sung wrote:



2017-12-24 07:55，"Ceriel Jacobs" <mailto:linux-...@crashplan.pro>>寫道：


It seems that BusyBox v1.27.2 doesn't load history from a history
file when ENABLE_FEATURE_EDITING_SAVEHISTORY is not set.

I think that the corresponding source code logic is here:
<https://git.busybox.net/busybox/tree/shell/ash.c#n13950
<https://git.busybox.net/busybox/tree/shell/ash.c#n13950>>

Such logic doesn't match the feature name, which is ...SAVEHISTORY
and not ...LOADHISTORY.

For example:
Busybox is often used in (emergency/disaster) recovery environments.
There is no need for any saving (readonly media). However it comes
handy when frequently used commands can be pre-loaded into the
initramfs (for interactive logins).

Is such a loadhistory-without-savehistory scenario tackled when
changing:

#if MAX_HISTORY > 0 && ENABLE_FEATURE_EDITING_SAVEHISTORY

to

#if MAX_HISTORY > 0

in ash.c?

In case it is not, what more to change to allow loading a HISTFILE
when ENABLE_FEATURE_EDITING_SAVEHISTORY is not set?


What's the point of loading frequently used commands if you don't save 
what you have typed *just now*? You can have .ash_history in a temp 
filesystem if you like, but it still counts as saving the history.



___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox

ash.c history loading and ENABLE_FEATURE_EDITING_SAVEHISTORY

[Ceriel Jacobs]

It seems that BusyBox v1.27.2 doesn't load history from a history file 
when ENABLE_FEATURE_EDITING_SAVEHISTORY is not set.


I think that the corresponding source code logic is here:


Such logic doesn't match the feature name, which is ...SAVEHISTORY and 
not ...LOADHISTORY.


For example:
Busybox is often used in (emergency/disaster) recovery environments. 
There is no need for any saving (readonly media). However it comes handy 
when frequently used commands can be pre-loaded into the initramfs (for 
interactive logins).


Is such a loadhistory-without-savehistory scenario tackled when changing:

#if MAX_HISTORY > 0 && ENABLE_FEATURE_EDITING_SAVEHISTORY

to

#if MAX_HISTORY > 0

in ash.c?

In case it is not, what more to change to allow loading a HISTFILE when 
ENABLE_FEATURE_EDITING_SAVEHISTORY is not set?


Is such a change a candidate for an upstream patch?
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox