Re: [Emutos-devel] [hatari-devel] EmuTOS version number display

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 03/10/2020 à 14:00, Vincent Rivière a écrit :
> On 03/10/2020 at 10:09, Thomas Huth wrote:
> > 1) It's slightly confusing to sometimes have one dot in the version 
> > string, and sometimes not.
>
>  From a programming view, yes.
>
>  From the user view, on the contrary, it is supposed to be simpler. Use 
> only major.minor for normal releases, then append a third number for 
> rare bugfix releases (a thing which only happened once in 19 years of 
> EmuTOS existence). So we expect to see only 2 numbers most of the time, 
> or 3 numbers in very rare occasions.

Hi

for previous version (before 1.0(.0) was released), you had version 
0.9.1 or 0.9.12 for example.
What is major.minor for you in that case ? Major=0 and minor=9 ? But 
then it doesn't fit the rule you're describing above, how do you call .1 
or .12 endpart ?

Or you can say that rule changed since v1.0 was released because you 
think it makes the display nicer when not showing final .0, but then as 
thomas said it was not advertized anywhere.

> > 3) The code for displaying the version is more complicated in every 
> > emulator that supports it
>
> I definitely agree with you. This is why I generally prefer using the 
> "string version" provided by EmuTOS, when there is room to display it 
> (it can be rather long for manual builds and snapshots). The "string 
> version" is what is displayed on the EmuTOS welcome screen and about 
> box, it can't be wrong.

An easy way to fix this display problem could be that emutos also 
provides a "short version string", that would just be the corresponding 
text of the numerical version that we parse in Hatari at the moment.
So numerical version could mean 1.0.0 but short version could then be "1.0".

Emulators would use the numerical version for internal needs but would 
print the "short version string" to show it to the user (instead of 
interpreting the numerical version)

> > 6) That rule isn't even mentioned in doc/version.txt of the EmuTOS 
> > sources as far as I can see.
>
> You are right, this has been forgotten, and should be fixed.
> Unless I'm plainly wrong and I invented that rule? By looking a the 
> official name of EmuTOS 1.0, it seems that such rule actually exists.

Maybe that rule exists only since emutos 1.0 was released ;)

Nicolas