Hi Eric,

> On Jul 29, 2021, at 10:16 AM, Eric Auer <e.a...@jpberlin.de> wrote:
> [..]
> Whether you want to spend 10 of your precious 640 kB to
> speed up floppy I/O depends on whether you use floppies.

Not the only reason.. But, it is one of the reasons it is not
run automatically in the installed systems config files.

Also when incomes to the install media, it is only loaded when
booting from floppy.

>>> 22. This reminds me that your scripts should
>>> check whether a sufficiently smart command.com
> 
>> 1) Works fine on MS-DOS, PC-DOS and probably others.
> 
> Your BAT scripts use no FreeCOM-specific features?

You specifically asked about computed jump targets in BATCH files. That works 
fine under other OS like MS-DOS, PC-DOS and probably others… Not just FreeCOM. 

If you mean the installer by “your BAT scripts”, the answer is the DO use 
FreeCOM specific features. But, only after they are known to be running under 
the FreeCOM shell.

If you mean batch scripts that get installed when the OS is installed, then 
generally they do not require any FreeCOM specific magic. But, that is not 
relevant anyway… Because you are running FreeDOS.  Besides, they are for 
FreeDOS not for *-DOS. 

>> 2) Those are only done AFTER the installer makes sure it is 
>> running under FreeCOM. There are other things the installer
>> does do that would not work under MS-DOS. So, it already
>> has guaranteed that FreeCOM is being used by the time
>> it gets around to using computed jump targets.
> 
> Interesting, so you already check for that? :-)

Not precisely.

> How
> do you implement "make sure to run under FreeCOM”?

The install boot media passes an option to inform the installer what media was 
booted (Like floppy disk, live cd, etc). If this information is present, the 
installer can assume it was booted from the install media and running under 
FreeDOS. If this information is not passed to the installer, it assumes you 
booted some other way and re-spawns itself under FreeCOM. It’s not perfect. 
But, it works and is something you’d have to intentional try and break for it
not to do what is expected. 

>> At present, only I update the "shipping packages”.
> 
> Would it help you to have volunteers for the version checks?

Version checking might help a little, but not that much. Jim is very good at 
noticing new versions for things and posting “news” items about the big ones 
and mirroring raw versions to ibiblio. Also, there is some custom stuff running 
on my server that monitors original websites for the packages. When for changes 
occur it publishes that information to an RSS feed which is available as a 
channel in the Slack group. 

What helps a lot are when those updated programs are provided precompiled and 
in a “ready to go” format. Accordingly, volunteers who would take those raw 
updates, compile and organize them, update there metadata, check for conflicts, 
etc. would be a huge help.

On a side note, tomorrow I’m releasing a new multi-language open-source program 
that makes it easy to view information on installed packages from the command 
line. PKGINFO will show the metadata, disk usage, all files or just it’s 
executables and more. It also has an advanced wildcard making things like 
*XMS*, ED*LN, *X*86, F?X*, etc. all return logical results.  But what I have 
not mentioned until right now, included in the PKGTOOLS package along with 
PKGINFO will be PKGMAKER. For more info on that, you’ll have to wait until 
tomorrow and see for yourself. Until then, have fun speculating on what it 
actually does. :-)

>> a large number of still active projects are released as source
>> only, or with their own installers, or in big globs of files 
>> that need reorganized to prevent utter chaos when installed.
> 
> That is surprising in context of DOS.
> 
>> It is a very time consuming process.
> 
> Definitely!

Yup.

> 
> Eric :-)

:-)

Jerome

_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to