Back to home page

DOS ain't dead

Forum index page

Log in | Register

Back to index page
Thread view  Board view
DOS386(R)

15.12.2011, 08:57
 

BOCHS 2.5 released 2011-11-29 (Announce)

http://sourceforge.net/projects/bochs/files/bochs/2.5/

What's new:

* 10% speedup (untested)
* many enhancements in CPU emulation (not that useful for DOS)
* some fixes in HD and VGA/VESA
* added ES1370 soundcard emulation (the old buggy one ??? MPXPLAY and Jeff's code support ES1371 only ...)
* some bugs fixed

NOT tested in DOS

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

15.01.2012, 04:38

@ DOS386
 

BOCHS 2.5.1 released 2012-01-06

fixed some bugs ;-) (untested)

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

16.01.2012, 06:38

@ DOS386
 

BOCHS 2.5.1 released 2012-01-06 | DOS

BOCHS 2.5.1 still works in DOS (this is not obvious, recent versions of FFMPEG, FFMPEG2DIRAC, MPLAYER, INNOUNP, ... don't) ... one missing import, no new problems found, speedup not tested, old problems not (?) fixed, ... :-)

http://sourceforge.net/projects/bochs/files/bochs/2.5.1/

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

25.09.2012, 10:52

@ DOS386
 

BOCHS 2.6 released 2012-09-02 | DOS

http://sourceforge.net/projects/bochs/files/bochs/2.6/

What's new (untested):

* another 10% speedup (if this happens another 18 times, BOCHS xx.yy.zz will one day be 6.7 times faster than 2.4 :-) )
* improved VGA
* added support for PC speaker (probably not useful in DOS)
* ...
* ???

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

14.04.2013, 16:46

@ DOS386
 

BOCHS 2.6.1 released 2013-04-09 | DOS

http://sourceforge.net/projects/bochs/files/bochs/2.6.1/

What's new (untested):

* Added and improved some fancy instructions (RDRAND/WRRAND, ...)
* some hardware fixes
* ...
* ???

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

17.04.2013, 12:31

@ DOS386
 

BOCHS 2.6.1 released 2013-04-09 | DOS | test

> What's new (untested)

"Bochs-2.6.1.exe" 4'633'096 | 146 Files in 8 Dirs 18.4 MiB
F022'3C66'56C5'6D8A'A0AE'7FFD'582E'6450

"bochs.exe" 3'401'216 2013-04-07 08:17
C4F7'C4AD'897E'604B'CCA0'6F74'8F66'8DF6

What's new in 2.6.1

* Still works (minimal test only) in DOS
* Works (minimal test only) with BIOS from 2.5.1 (note that the opposite effect is the rule)
* Added missing import "SetScrollInfo", 2.5.1 has only "MapDialogRect", 2.6.1 now has both
* Changed "vga" , must be adjusted or kicked from config file
* Deleted "i440fxsupport" , must be kicked from config file
* Deleted "pnic" , must be kicked from config file
* Warns about "keyboard_type" , "keyboard_serial_delay" , "keyboard_paste_delay" , "keyboard_mapping" , "user_shortcut" in config file

PS: Georg has an (unreleased) DGJPP port of BOCHS: #12428

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

05.06.2013, 03:56

@ DOS386
 

BOCHS 2.6.2 released

What's new: almost nothing (minor CPU XXX-64 fixes, minor docs fixes)

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

24.06.2013, 15:45

@ DOS386
 

BOCHS 2.6.3 awaited (not necessarily soon)

there are many crucial fixes in SVN or in the pipeline :-)

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

13.07.2014, 12:46

@ DOS386
 

BOCHS 2.6.6 is out - 2014-Jun-15

What's new:

++ skipped 2.6.3 and 2.6.4 (but there is 2.6.5)

+ emulation speedup (untested)
+ fixed PC speaker BUG
+ reduced BLOAT of the main EXE
+ other improvements (reportedly, see official history)

? "fixed" mouse wheel support in the BIOS but incompletely (still unusable, fixed in SVN again, retest needed)

- broke SB16 emulation (does it still work foe anyone?)
- IDE DMA / XDMA / IDECHECK not unfixed
- needs "IPHLPAPI.DLL" now (and there are still 2 missing imports)

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

Rugxulo(R)

Homepage E-mail

USA,
13.07.2014, 23:12

@ DOS386
 

BOCHS 2.6.6 is out - 2014-Jun-15

BTW, do you know of a packet driver that works with BOCHS?

DOS386(R)

15.07.2014, 10:37

@ Rugxulo
 

BOCHS 2.6.6 is out - 2014-Jun-15

> BTW, do you know of a packet driver that works with BOCHS?

NO. I didn't test the networkig yet. But the rat BUG seems finally fixed ... thanks to Volker Ruppert :-)

Does the sound (SB16 ISA) still work (host<>DOS) for you or anyone in 2.6.6 ?

What's new in 2.6.6:

# "upgraded" DLL dependency "WSOCK32.DLL" -> "WS2_32.DLL"

# added junk DLL dependency "IPHLPAPI.DLL"

# reduced main EXE bloat (effect also persists after ZIP'ping)

# broke SB16 ???

# IDE DMA / XDMA / IDECHECK problems not fixed (not "IDE DMA / XDMA / IDECHECK not unfixed" damn :-( )

Missing imports in 2.6.1 to 2.6.6:
"SetScrollInfo" and "MapDialogRect" in "USER32.dll"

DLL's 5 KiO ZIP

Fixed BIOS (32 KiO ZIP):

http://sourceforge.net/p/bochs/code/HEAD/tree/trunk/bochs/bios/BIOS-bochs-latest
http://users.freebasic-portal.de/dos386/b266bio.zip

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

DOS386(R)

10.11.2014, 04:14

@ DOS386
 

BOCHS 2.6.7 is out - 2014-Nov-02

What's new in 2.6.7 (NOT yet tested at all):

# Included the "mouse-BUG" fix (untested)
# huge register (512-bit and 1024-bit) enhancements in the CPU
# some fixes

The XDMA-BUG is probably NOT fixed. Fix for the SB16 regression not mentined either.

DLL's 5 KiO ZIP

http://sourceforge.net/p/bochs/bugs/1331 CLOSED: FIXED

http://sourceforge.net/projects/bochs/files/bochs/2.6.7

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

Ninho(R)

E-mail

10.11.2014, 22:19

@ DOS386
 

BOCHS 2.6.7 is out - 2014-Nov-02

Uh the "Avira antivirus" went mad about the download Bochs-2.6.7.exe
saying it is or contains "TR/Dropper.Gen"!
Though I would trust Sourceforge not to point to infected downloads,
I'd like second advices !

> http://sourceforge.net/projects/bochs/files/bochs/2.6.7

I got file "Bochs-2.6.7.exe" 4,988,462 bytes
md5sum : 6bd870625f8346021c0e0c620dd55980

Opinions ?

---
Ninho

Rugxulo(R)

Homepage E-mail

USA,
11.11.2014, 00:44

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

> Uh the "Avira antivirus" went mad about the download Bochs-2.6.7.exe
> saying it is or contains "TR/Dropper.Gen"!
> Though I would trust Sourceforge not to point to infected downloads,
> I'd like second advices !
>
> > http://sourceforge.net/projects/bochs/files/bochs/2.6.7
>
> I got file "Bochs-2.6.7.exe" 4,988,462 bytes
> md5sum : 6bd870625f8346021c0e0c620dd55980
>
> Opinions ?

Did you forget that most antiviruses have tons of false positives due to bad heuristics? See here.

Also, did you forget about Virus Total? A quick (re)scan online shows only one out of 55 complaining (guess which one).


SHA256:         0bc168b5e3dfd7733ed793fd8b7aff1ee8059ba8f4f57fff2237927ead7127a9
File name:      Bochs-2.6.7.exe
Detection ratio:        1 / 55
Analysis date:  2014-11-10 23:27:26 UTC ( 0 minutes ago )



(Mon Nov 10, 05:26 PM) /tmp/doydoy # ls -log
total 4884
-rw-r--r-- 1 4988462 2014-11-10 17:26 Bochs-2.6.7.exe
(Mon Nov 10, 05:26 PM) /tmp/doydoy # md5sum *
6bd870625f8346021c0e0c620dd55980  Bochs-2.6.7.exe

DOS386(R)

11.11.2014, 18:37

@ Rugxulo
 

VIRUS 2.6.7 is out - 2014-Nov-02

> Uh the "Avira antivirus" went mad about the download Bochs-2.6.7.exe
> saying it is or contains "TR/Dropper.Gen"! I'd like second advices !

Delete Avira ;-)

PS: virii
PPS: virii

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

Ninho(R)

E-mail

11.11.2014, 19:52

@ DOS386
 

VIRUS 2.6.7 is out - 2014-Nov-02

>> Uh the "Avira antivirus" went mad about the download Bochs-2.6.7.exe
>> saying it is or contains "TR/Dropper.Gen"! I'd like second advices !


@Rugxulo wrote :
> Did you forget that most antiviruses have tons of false positives due to bad heuristics?

I didn't, that's why I asked for second opinions !
Although rare, an infection affecting Sourceforge's approved download centers is not totally out of the question, imho.

@DOS386 wrote :
> Delete Avira ;-)

Why would I ? It is non obnoxious and it is very very infrequently flagging downloads that I do.

> virii[/url]
> PPS: virii

Nope, that is an urban legend. /Virus/ is a latin word belonging to gender "neutral" and without a plural (in latin). If its plural existed, it woud NOT resemble viri, nor virii ! Ergo, in English, we should write viruses.

@Both :
Now please, have you downloaded the windows prebuilt executable in question ? If you have, has it matched the size and MD5 which I posted ?

---
Ninho

Rugxulo(R)

Homepage E-mail

USA,
12.11.2014, 01:28

@ Ninho
 

VIRUS 2.6.7 is out - 2014-Nov-02

> @Rugxulo wrote :
> > Did you forget that most antiviruses have tons of false positives due to
> bad heuristics?
>
> I didn't, that's why I asked for second opinions !

http://www.virustotal.com/ gives a better answer than we ever could. I don't know nearly enough (practically nil) about Win32/installers/.DLLs/APIs.

> Although rare, an infection affecting Sourceforge's approved download
> centers is not totally out of the question, imho.

You'd have to personally ask BOCHS devs or SF admins whether they scan stuff in advance before offering it for download.

If really worried, grab sources, compile it yourself, and re-scan it.

But actually, don't. Just from my own limited experience, many many compilers' output mistakenly gets flagged. It's really annoying, but there is no good solution.

The "only" (tm) solutions are:

1). Constantly (x 11, eat roast beef) send false positives to AV makers (and/or disable heuristics entirely and/or locally whitelist any specific programs and/or disable realtime scanning entirely).
2). Only download/install stuff for Windows under VMware/VBox/Hyper-V.
3). Don't use AVs at all. Reformat/reinstall when needed.
4). Just use Linux/OS X/etc.

> @DOS386 wrote :
> > Delete Avira ;-)
>
> Why would I ? It is non obnoxious and it is very very infrequently flagging
> downloads that I do.

I'm just unlucky by birth, I guess, stuff always breaks for me. Months back, even Avira stopped working/updating, and I couldn't even successfully reinstall. So I switched to Avast. And yes I still get false positives from time to time, usually whenever I try to compile something ultra trivial with an obscure compiler.

> > virii[/url]
> > PPS: virii
>
> Nope, that is an urban legend. /Virus/ is a latin word belonging to gender
> "neutral" and without a plural (in latin). If its plural existed, it woud
> NOT resemble viri, nor virii ! Ergo, in English, we should write viruses.

Lingua latina, which is derided more widely than DOS! Yes, I took that in high school, I liked ancient things even then. Don't forget that "agricola" is "m." but "virtus" (manliness, courage) is "f."!

> @Both :
> Now please, have you downloaded the windows prebuilt executable in question
> ? If you have, has it matched the size and MD5 which I posted ?

Yes, it was the same, see my quote from PuppyLinux above.

Rugxulo(R)

Homepage E-mail

USA,
11.11.2014, 00:53

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

Well, I've gone ahead and uploaded this .EXE to their website's form for false positives, so I hope they'll correct the error (doubt it but who knows).

https://analysis.avira.com/en/submit

Rugxulo(R)

Homepage E-mail

USA,
11.11.2014, 23:52

@ Rugxulo
 

BOCHS 2.6.7 is out - 2014-Nov-02

> Well, I've gone ahead and uploaded this .EXE to their website's form for
> false positives, so I hope they'll correct the error (doubt it but who
> knows).
>
> https://analysis.avira.com/en/submit

> (email): "The file 'Bochs-2.6.7.exe' has been determined to be
> 'FALSE POSITIVE'. In particular this means that this file
> is not malicious but a false alarm. Detection will be
> removed from our virus definition file (VDF) with one of the
> next updates."

Ninho(R)

E-mail

12.11.2014, 00:08

@ Rugxulo
 

BOCHS 2.6.7 is out - 2014-Nov-02

Thanks !
This has been the first time ever that Avira gave Bochs a (false)positive.

>> Well, I've gone ahead and uploaded this .EXE to their website's form for
>> false positives, so I hope they'll correct the error (doubt it but who
>> knows).

>> https://analysis.avira.com/en/submit

>> (email): "The file 'Bochs-2.6.7.exe' has been determined to be
>> 'FALSE POSITIVE'. In particular this means that this file
>> is not malicious but a false alarm. Detection will be
>> removed from our virus definition file (VDF) with one of the
>> next updates."

---
Ninho

Rugxulo(R)

Homepage E-mail

USA,
12.11.2014, 01:33

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

> Thanks !
> This has been the first time ever that Avira gave Bochs a (false)positive.

First for you, maybe, not everybody else. Get used to it. It's a sad fact of life.

Anyways, a rescan on Virus Total shows this:


SHA256: 0bc168b5e3dfd7733ed793fd8b7aff1ee8059ba8f4f57fff2237927ead7127a9
File name:      Bochs-2.6.7.exe
Detection ratio:        0 / 54
Analysis date:  2014-11-12 00:31:15 UTC ( 0 minutes ago )

DOS386(R)

12.11.2014, 09:08

@ Rugxulo
 

BOCHS 2.6.7 is out - 2014-Nov-02

> > Delete Avira ;-)
> Why would I ?

# Avira doesn't work in DOS
# not a single virus (less than 15 years old) works in DOS
# a good DOSser knows how to use a PC
# a good DOSser knows what's going on eir system, a virus has no chance to hide
# a good DOSser downloads stuff only from trusted pages and authors that don't contain virii
# false positives

> If you have, has it matched the size and MD5 which I posted ?

YES

"Bochs-2.6.7.exe" 4'988'462 | Installer
6BD870625F8346021C0E0C620DD55980
"http://sourceforge.net/projects/bochs/files/boc
hs/2.6.7/Bochs-2.6.7.exe/download"

"bochs.exe" 3'509'248 2014-11-02 08:29 | 2.6.7 main EXE
5E655943E1D8D3675701D85C7E039B4E


> (email): "The file 'Bochs-2.6.7.exe' has been determined to be
> 'FALSE POSITIVE'. In particular this means that this file
> is not malicious but a false alarm.

Thanks, but I always had known this. I don't need the monstrous "virus detection" crap.

> You'd have to personally ask BOCHS devs or SF admins
> whether they scan stuff in advance before offering it for download.

I would NOT do this.

> If really worried, grab sources, compile it yourself, and re-scan it.

Virus distributed as source code? Virus distributed inside compiler library?

> It's really annoying

NO

> but there is no good solution.

There is: delete Avira & Co :-)

> The "only" (tm) solutions are:
> 1). Constantly (x 11, eat roast beef) send false positives to AV makers

No need.

> 2). Only download/install stuff for Windows under VMware/VBox/Hyper-V

No need.

3). Don't use AVs at all. Reformat/reinstall when needed.

YES :-) Use DOS :-)

4). Just use Linux/OS X/etc.

If everybody uses Linux one day, everybody will have virii one day - security of minority ;-) (PS: Linux can't get faked)

> And yes I still get false positives from time to time, usually
> whenever I try to compile something ultra trivial with an obscure compiler.

Does this happen in DOS too?

> Nope, that is an urban legend. /Virus/ is a latin word

OK, BOCHS is full of virae and viroa

Test results ... what's new in 2.6.7:

# it doesn't work anymore
# increased bloat of main EXE
# compiled for SSE2 or higher (maybe this is the chief reason why it doesn't work and crashes on a fancy MOVNTQ-like instruction)

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

Ninho(R)

E-mail

12.11.2014, 10:16

@ DOS386
 

BOCHS 2.6.7 is out - 2014-Nov-02

>>> Delete Avira ;-)
>> Why would I ?
> # Avira doesn't work in DOS

Not a reason to call Avira names : there is not a single antivirus product (with updated threat database) for DOS anymore. F-Prot was (probably) the last one to yield, and it's been many years now since it quit the DOS scene.

> # not a single virus (less than 15 years old) works in DOS
> # a good DOSser knows how to use a PC
> # a good DOSser knows what's going on eir system, a virus has no chance to
> hide

Hardly anybody uses DOS _exclusively_ (do you ?)
In actual fact nowadays, hardly anyone uses DOS intensively ( as _we_ both do).

> # a good DOSser downloads stuff only from trusted pages and authors that
> don't contain virii

You'll have your wings burnt someday by adhering to this credence. Even assuming the software authors, and their hosting sites, are trustworthy, you can't rule out compromised sites - the more so as scripting is so pervasive nowadays it is often difficult to download using only plain HTML or FTP, etc.

> # false positives
False positives are much less annoying than false negatives !


>> If you have, has it matched the size and MD5 which I posted ?
> YES

Danke sehr !


> ... I don't need the monstrous "virus
> detection" crap.

With that attitude, you'll be burnt eventually, imho. While virus detection is far from perfect - rather, I concede, near to imperfect - it's still useful precaution, in a dangerous virtual world full of malicious script injection and other attacks.

>> You'd have to personally ask BOCHS devs or SF admins
>> whether they scan stuff in advance before offering it for download.
> I would NOT do this.

Even though a file could have been checked by their authors prior to uploading it, still subsequent downloads could be compromised if the site was victim of some type of attacks.

>> If really worried, grab sources, compile it yourself, and re-scan it.

I like to have the "official" compiles provided by the Bochs team, for reference and in order to do reports. I do compile personal, lighter & faster versions lacking the features I don't usually need (AMD64, e.g.). However :

> # compiled for SSE2 or higher (maybe this is the chief reason why it
> doesn't work and crashes on a fancy MOVNTQ-like instruction)

Awww ! If this is the case, the official build is no use here :=)

---
Ninho

Rugxulo(R)

Homepage E-mail

USA,
23.11.2014, 22:47

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

> > # compiled for SSE2 or higher (maybe this is the chief reason why it
> > doesn't work and crashes on a fancy MOVNTQ-like instruction)
>
> Awww ! If this is the case, the official build is no use here :=)

The only sources besides .tar.gz or .src.rpm are apparently msvc-src.zip. IIRC, latest MSVC (not that I tried) enables SSE2 by default these days (though maybe you can disable that??). Presumably they think everyone has a new enough machine since Pentium 4 (2000) and Athlon64 (2003). What cpu are you using: PIII? Athlon XP? (Even antiX Mepis Linux 13.x says this: "32-bit edition uses older Flash for PII/PIII processors".)

Rugxulo(R)

Homepage E-mail

USA,
26.11.2014, 22:04

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

> >> You'd have to personally ask BOCHS devs or SF admins
> >> whether they scan stuff in advance before offering it for download.
> > I would NOT do this.
>
> Even though a file could have been checked by their authors prior to
> uploading it, still subsequent downloads could be compromised if the site
> was victim of some type of attacks.

I just noticed that some projects on SourceForge (and even Bochs) now have a big logo that says "Verified virus free!". See for yourself: http://sourceforge.net/projects/bochs/

I have no idea what that means exactly, not sure if that's something new behind the scenes or not.

A quick search only shows some random comments from the Pale Moon web browser peeps.

Rugxulo(R)

Homepage E-mail

USA,
27.11.2014, 01:24

@ Ninho
 

yet another false positive

> >> If really worried, grab sources, compile it yourself, and re-scan it.
>
> I like to have the "official" compiles provided by the Bochs team, for
> reference and in order to do reports. I do compile personal, lighter &
> faster versions lacking the features I don't usually need (AMD64, e.g.).

BTW, I just now upgraded Regina Rexx (from SourceForge) on this machine. So I downloaded the (manual) install .ZIPs (32-bit and 64-bit) and source .ZIP, but Chrome browser (in its infinite wisdom) flagged two of the three files. Not warned, like they often do, but actually said it's malicious! Though it did (barely) let me "recover" the files ("hurt me plenty" ... sic!).

This is ridiculous. First of all, SourceForge, as mentioned, declares it is "verified virus free!". Secondly, how the hell do you have a virus in a source-only .ZIP file?? Sure, it's possible but extremely unlikely! Thirdly, I just uploaded both problematic files to http://www.virustotal.com , and they found literally nothing!

It's like they want to force us to sign all binaries (and webpages), which is just ridiculous. I'm not saying all protection is bad. I reluctantly use an anti-virus myself (and often have false positives). Maybe it helps (sometimes). The problem is that the cure is often worse than the disease!

DOS386(R)

01.12.2014, 05:00

@ Ninho
 

BOCHS 2.6.7 is out - 2014-Nov-02

> You'll have your wings burnt someday by adhering to this credence

No. Why? Because I

1. don't store any private / valuable / sensitive data into Windaube

2. if I use Windaube, I just restore it regularly from a clean backup - virii can play, but they will never succeed to cause a notable damage

3. if I process private / valuable / sensitive data, my Internet cable is physically unplugged, and if I access the Internet, all my storage media with private / valuable / sensitive data are unplugged, so YES, until virii learn to fly (to travel between physically unplugged components), I'm 100 % safe :-) :-) :-) :-) :-) :-) :-) :-) :-) :-) :-) :-)

> > # compiled for SSE2 or higher (maybe this is the chief reason why it
> > doesn't work and crashes on a fancy MOVNTQ-like instruction)

> Awww ! If this is the case, the official build is no use here :=)

http://sf.net/p/bochs/bugs/1363/ CLOSED: NOT a BUG :-(

---
This is a LOGITECH mouse driver, but some software expect here
the following string:*** This is Copyright 1983 Microsoft ***

Back to index page
Thread view  Board view
13907 Postings in 1226 Threads, 206 registered users, 10 users online (0 registered, 10 guests)
DOS ain't dead | Admin contact
RSS Feed
powered by my little forum