Hello Luchezar,

>> or call the 'optimized' kernel keUNSTABLExxx or keARxxx, as the main
>> stream kernel should concentrate on FIXING bugs, rather then introducing
>> new ones.

> 100% agreed. Since I use "unstable" kernel every day in practice, I think
> it has no more bugs than the "stable" one.

that you use it doesn't mean there are no news bugs.
e.q. the complete ioctl was 100% mixed/changed/...
how many ioctl's are you using - not every day, but at least once per
month?
do you use MS-LANMAN networking  (at least once per month)?
etc...

just because YOU don't find bugs (old or new) doesn't mean there are
no (old or new) bugs.

in particular (by your own finding) the new SYS is buggy ( if loadseg
!= 60), but you only found out after using SYS (at least once).

this shall not imply the arkady introduces new bugs every other day
(although I've seem a couple of OOPS's); it's simply that I can't
verify (perr review) what he changed simply because his stupid 'change
everything' attitude.

and so I won't touch the arkady branch with a 3 meter long stick.


> But during the last few weeks I
> noticed several more bugs and incompatibilities present in both "stable"
> and "unstable" branches, most of them extremely difficult to fix, namely:

> 1. While copying a directory tree containing many LFNs *without* DOSLFN
> loaded and *with* delayed write of SMARTDRV in a FAT32 directory, many
> invalid directory entries are created, which consist of parts of the
> contents of the copied files. Turning delayed write off makes the bug
> vanish. If FreeDOS ir replaced by MS-DOS or if copying to a FAT16 
> directory under FreeDOS, there is no bug.

if you are using (and distributing) SMARTDRV, you can also distribute
msdos xx.

smartdrv is an entirely different beast, that knows a lot about the
underlying DOS; expecting compatibility with SMARTDRV is probably
beyond the requitrements of this project (wouldn't hurt for sure, but
not *required* )

you didn't report, if the same problem happens with LBACACHE.

> 3. If load segment (/L option) was different than 60h, bootstrap causes a
> read error during the loading of the kernel if it's longer than about 62
> KB. The exact value of the load segment doesn't matter, and the bug 
> reveals for *any* value other than 60h.

*very* easy to fix if you have softice loaded.

boot normal.
softice>bpint 13 ah=2
softice>bpr 0:0 a000:0 t
softice>boot             (which reboots the machine, but leaves s-ice
                         softice>active, and you can step booting !!)

> 4. SYS hangs up or causes an invalid opcode at the time of reading the
> boot sector of my ATA flash cards, if compiled by anything than Watcom.
> Non-Watcom versions of SYS work on all other drives.
again - softice should be your friend.

> So, as a prospective user of the kernel, after contributing to it for more
> than an year, I can conclude than it's good enough for simpler tasks not
> involving writing a lot of long named files on a FAT32 partition. For more
> complex tasks, however, MS-DOS 7.1, PC-DOS 7.1 and ROM-DOS 7.1 are more
> suitable.
not 'complex tasks'. for tasks involving LFNDOS and creating new
files; creating a lot of them just makes it more probable to happen.

> Thank you for your attention. It's time for me to concentrate on other
> tasks. It's been a pleasure for me to work with you guys. I hope that
> FreeDOS will be THE DOS one day. But hardly "real soon now"...

It's been a pleasure for me to work with Bart, you and the nearly
forgotten victor who gave us FAT32 in a very ompressive way.
For me, it's THE dos today - and I can affort to
distribute a junk DOS even less then Steve Gibson.

It's good enough for what I do, and I take care that it remains that
way.

tom




-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
_______________________________________________
Freedos-kernel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/freedos-kernel

Reply via email to