[fd-dev] [ANNOUNCE] FreeDOS kernel build 2028
Kernel build 2028 is now available at the sourceforge download page: http://sourceforge.net/project/showfiles.php?group_id=5109release_id=121405 Bart 2002 Dec 9 - Build 2028 Bart Oldeman ([EMAIL PROTECTED]) + Changes Tom * move clock driver initialization to inittext * new ELSE_ISFAT16 macro to save some bytes * use _TEXT for old HMA_TEXT, _LOWTEXT for old _TEXT to get a slightly less complicated memory model from the C compiler's point of view. * add support for int2f/4a01,4a02: query/allocate HMA space. * updated config.txt + Changes Arkady V. Belousov * slightly improved asm for print_hex and int2f clear_ax stuff * use #error for unknown compilers * remove hexDigits reference * correct comment in kernel.asm (pad to 05fdh) + Changes Bart * fix ss:sp init bug for int21/ax=4b01 (visible in DEBUG) * fix redirection problem for character output routines * fix press the any key message for floppy swaps * correct cluster value for BAD cluster * correct FAT12/16/32 cutoff magic values as per the MS spec * the BUFFERS structure is now the same as in MSDOS 5+ and uses a circular double-linked list using near pointers. * move all old RCS log messages into one seperate file * divide the fnode f_cluster_offset field by the size of a cluster: this gives a more efficient map_cluster(). * point the CharMapSrvc pointer to the DOS DS for AARD compliance. * nls.c: set CX to cooperate with load_unf.exe -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] GPL
On Fri, 6 Dec 2002, Andreas K. Foerster wrote: On Fri, Dec 06, 2002 at 05:45:08PM +0300, Arkady V.Belousov wrote: AKF To release a non-free program is always ethically tainted, but AKF legally there is no obstacle to your doing this. If you are the AKF copyright holder for the code, you can release it under various AKF different non-exclusive licenses at various times. Thank you. BTW, what mean non-exclusive licenses in sense of applying to different distributions (releases) of same software? This text wasn't written by me, it was a cite. When I read your question, I first thought, this is obvious. But the more I think about it, the less obvious it is... I read non-exclusive as may not contradict each other. If I write software under license A and B and then license A says: this code may not be released under any other license, then I would be contradicting myself (hence, a fairly technical issue, sue yourself!). And yes, ethically tainted comes from the FSF position that all non-free (as in speech) software is ethically wrong. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] GPL
On Fri, 6 Dec 2002, Arkady V.Belousov wrote: BO If I write software under license A and B and then license A says: this BO code may not be released under any other license, then I would be BO contradicting myself (hence, a fairly technical issue, sue yourself!). But any commercial license (which restricts distribution) contradicts to GPL! only for non-full-copyright holders, unless the software was dual (or more)-licensed from the start. AOL/Netscape is allowed to release closed-source versions of Netscape but a large part of Netscape is Mozilla, where Mozilla is triple-licensed under the NPL, MPL, and LGPL. Similar for Sun with StarOffice and OpenOffice (SISSL/(L)GPL). And Pat Villani had a commercial (better said, proprietary) predecessor of the FreeDOS kernel which he, if he wants, can still treat as such. However all of my, Tom's, ror4's, James Tabor, ... changes are under the GPL, and we, as individuals, are not allowed to change the license of the FreeDOS kernel as a whole. Neither can Pat take those changes and make them proprietary. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] Dosemu without FreeDOS
On Fri, 6 Dec 2002, Ralf Quint wrote: On Wednesday 04 December 2002 10:03, you wrote: Hi all, I want to know can I start another version of DOS than FreeDOS under Dosemu? If answer is yes where I can find documentation on how to do it? Why nobody don't want to help me about this matter? Have you considered the possibilty that you are asking this question in the wrong mailing list? You should better try to seek contact through either a dosemu-specific mailing list (if it exists, i don't know) or in a dosemu newsgroup on the usenet. The fact that FreeDOS is the choice for the dosemu folks does not mean that a lot of FreeDOS people are familiar with the inner workings of dosemu He got a reply from Eric. In any case the question is like asking Bill Gates how to install DRDOS on my PC -- it's not relevant in this forum indeed. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] OT: DRDOS 7.03
On Wed, 4 Dec 2002, Mark Olesen wrote: OT: FYI: I noticed DRDOS.COM has 7.03 image available for download (license)? Nothing new here (non-commercial or 90 days evaluation use only) BIG FAT AD See what Doctor DOS can't let you see ... Use FreeDOS and you get FAT32, LBA, source code and unlimited commercial use for free! Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] NANSI / NNANSI / XTERM: License, mode setting, escapecodes
On Mon, 2 Dec 2002, Eric Auer wrote: What FREEWARE ANSI would you currently recommend? nansi.sys 3.4 is GPLed. see www.kegel.com. For Linux console: you must be root and use SVGATextMode, I believe. or use the graphical fbdev console (vesafb, ...), fbset and setfont. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
RE: [fd-dev] RV: FreeDos with FAT32 Success!!!!!
On Mon, 2 Dec 2002, Steffen Kaiser wrote: However, the POSIX.1 rationale includes statements like this: The C Standard [aka ANSI-C] includes functions fgetpos() and fsetpos(), which work on very large files by use of a special positioning type. -and- An invalid file offset [...] may be both implementation defined and device dependent [...]. A negative file offset may be valid for some devices in some implementations. -and- See B.6.5.2 for a explanation of the use of signed and unsigned offsets with lseek(). But I can't find any references in there. So, to deal with files beyond 2GB from within C is an implementor defined issue. The API function DOS-42 is a straight forward implementation of this definition with sizeof(off_t) == 4. it's a little different from lseek(2). The difference is in the return value. POSIX: error: (off_t)-1 returned and errno set DOS: error: CF set and error in AX. This makes it possible for DOS-42 to use unsigned off_t's with wrap-around semantics. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: Fw: [fd-dev] Dosemu - X windows
On Mon, 2 Dec 2002, Eric Auer wrote: Hi, according to Andreas, if you use the LZ algorithm or something similar and just compress your text mode screen buffer and send it (or send a diff from the version 1/10 second ago, even), you get a rather fast program-independend copy of your screen which you can send for example through a serial cable to your Linux or anything else that decodes the data. I think text mode is okay for many things, and as said, it is fast. Well you can certainly run dosemu (and hence DOS programs) remotely. The way it works is that * DOS programs write to the text mode buffer at 0xB or 0xB8000 * n times per second, (n=5 by default), DOSEMU examines the text buffer and tells the contents to the Slang library. * the Slang library can then compare the screen contents with the old state, and generates terminal escape sequences to update the screen. It depends on your CPU how long this takes. It's quick enough for me. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
[fd-dev] [hands on!] COUNTRY= and load_unf
Hi, here's an approach that actually works and isn't one of the maybe we should implement esoteric feature X emails ;) Steffen claimed that even though country= doesn't work yet in our config.sys, I could use load_unf instead. So I tried and failed but with some adjustments succeeded. Here's how you can have 24-hour times in DIR and dd-mm-yy for instance: * download unf_bin_0.3_2000-07-01.zip from ftp://ftp-fd.inf.fh-rhein-sieg.de/pub/freedos/local/unf_bin_0.3_2000-07-01.zip * patch recompile the kernel: nls.c needs this change: @@ -131,6 +131,7 @@ /* 0x14FF == installed */ r.BX = 0; /* make sure the NLSFUNC ID is updated */ + r.CX = NLS_FREEDOS_NLSFUNC_ID; if (muxGo(0, r) != 0x14ff) return DE_FILENOTFND; /* No NLSFUNC -- no load */ if (r.BX != NLS_FREEDOS_NLSFUNC_ID) /* FreeDOS NLSFUNC will return */ * from the unf_bin zip, extract load_unf.exe and 049-850.unf * run: load_unf 049-850.unf * run: command * run: dir * watch it (German-style dates). Notes: - the new country info only affects secondary shells, not the freecom that you run load_unf.exe from - I also tried something that makes sense for the Netherlands (and it works) --- 001-437.unf Mon Dec 2 21:44:30 2002 +++ 031-437.unf Mon Dec 2 21:52:31 2002 @@ -1,5 +1,5 @@ -## Universal NLS data Format file: 001-437.unf -$country 1 +## Universal NLS data Format file: 031-437.unf +$country 31 $codepage 437 # Global options @@ -8,16 +8,16 @@ [table 1] # Extended Country Information DOS-65-01 / DOS-38 $make_size - = W1 W437# Country ID Codepage - = W0 # Date format: 0/1/2: U.S.A./Europe/Japan - = 36 0 0 0 0 # $# Currency string - = 44 0 # ,. # Thousand's separator - = 46 0 # .. # Decimal point + = W31 W437 # Country ID Codepage + = W1 # Date format: 0/1/2: U.S.A./Europe/Japan + = 69 85 82 0 0 # EUR..# Currency string + = 46 0 # .. # Thousand's separator + = 44 0 # ,. # Decimal point = 45 0 # -. # Date separator = 58 0 # :. # Time separator = 0 # Currency format (bit array) = 2 # Currency precision - = 0 # time format: 0/1: 12/24 houres + = 1 # time format: 0/1: 12/24 houres $go_forward 4 # Far address of upcase function (to be calc at runtime) = 44 0 # ,. # Data separator - Contributions (unf files) for other countries are welcome. This doesn't involve difficult kernel hacking or anything. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] [hands on!] COUNTRY= and load_unf
On Mon, 2 Dec 2002, Bart Oldeman wrote: * from the unf_bin zip, extract load_unf.exe and 049-850.unf * run: load_unf 049-850.unf i forgot: use the /a switch to activate it too: load_unf /a 049-850.unf instead of just load_unf 049-850.unf. * run: command * run: dir * watch it (German-style dates). Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
RE: [fd-dev] RV: FreeDos with FAT32 Success!!!!!
On Fri, 29 Nov 2002, Ralf Quint wrote: Natasha Portillo wrote: Sorry, I made a mistake: Nevertheless the operating system, the FAT file system is limited to a 2^32 bytes (minus one) file size. This applies on Linux, MacOS (Classic and X), Windows 9x, Windows NT, DOS, etc. Sorry, but this generalized answer is definitely wrong. At least for Windows 2000 and Windows XP, the maximum filesize on a FAT32 volume is 4GB, and on NTFS it would be limited only by the available space on the NTFS volume I thought that 2^32 bytes==4GB (more exactly, GiB). In any case under FAT32 DOSes you have to open using int21/ah=6c with bit 4 of BH set, otherwise the limit is 2GB and lseek (int21/ah=42) might complain. FreeDOS might have problems with files 2GB; it still ignores this bit4. Also many DOS programs assume that the program size is a signed long int (C) or longint (TP) so the 2GB--4GB region is a bit risky. I've just never tested it so I don't know. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
[fd-dev] Harddisk corruption w. FreeDOS
Hi, I haven't used the FreeDOS install program for quite some time, but I think this is strange: he seems to have only used FreeDOS in r/o operation (dir) so filesystem corruption should be impossible. The one thing I'd like to ask the install people (Jeremy, Jim) is if the install program between step 2 and 4 (see below) copies anything to the hard disk? Bart From: Bogislav Rauschert ([EMAIL PROTECTED]) Subject: Harddisk corruption w. FreeDOS View: Complete Thread (3 articles) Original Format Newsgroups: comp.os.msdos.misc Date: 2002-11-28 09:45:57 PST Hi! I am currently using linux on my desktop and had no experience with DOS, but I want to install a small OS on a very old laptop, so I tried FreeDOS (I know this is not a FreeDOS newsgroup but may be there is someone who can help?!?). I downloaded the latest disk image release and just started it from CD on my desktop (not the laptop); there is (or were!!!) two Fat32 partitions, one with Windows 98 on it (and one should be for FreeDos testing in the future). I just booted the FreeDOS kernel and it recognized the two partitions. Then I only made a dir c:, and wondered why there were only rubbish symbols (like !_ ? and so son). NOW THE WHOLE PARTITION IS LOST, which means that there are still files on the partition, but the filenames are rubbish and I can not find any directories. Again, here is what I did: 1. Starting the computer with FreeDOS image CD. 2. Booting the CD. 3. Selecting DOS promt with Kernel 32. 4. Typing dir c: 5. Swithcing off computer. Starting Linux again, Linux was able to read from that partition, but it showed only the corrupted files. Please tell me if there is ANY way to get the data on my /dev/hda1 back?!?!?! I use this partition also for storage with Linux, so this is sort of important to me. Does anyone know this bug or know a tool to fix the harddisk? Thank you, Bogislav -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
Re: [fd-dev] codepage IDs
On Tue, 26 Nov 2002, Axel C. Frinke wrote: On Sat, 23 Nov 2002 01:39:39 -0500 (EST), Bart Oldeman wrote: BO This is what the Linux console does: its console driver first translates BO from the used character set to Unicode, and then maps the Unicode to BO the font used. How does this work in full screen text mode? At the end, this requires to display more than 256 different chars at the same time! (Tell me if I'm completely wrong with this.) This is full screen text mode what I was talking about. You see funny little square boxes for each unicode character for which the font has no equivalent. I saw these square boxes for instance in Matthias mails because he used 264 180 B4 ยด ACUTE ACCENT instead of the ASCII ' and I was using cp437 on the console. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^^=== This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^^===
Re: [fd-dev] codepage IDs
On Thu, 21 Nov 2002, Axel C. Frinke wrote: BO with cp850 vs. cp437 you already lose the mixed double/single box drawing BO characters, but these aren't used as often as the non-mixed ones. I don't think so! Consider a programm which displays a table with different box styles for 'inner' and 'outer' lines. Under cp850, the 'junctions' will already look weird. Sure, that's exactly what I said. Let me rephrase: cp850 does not have the mixed double/single box drawing characters, but cp437 has them. On the other hand there are many DOS programs who avoid those characters and only use the single-only or double-only boxes. And all of the DOS style codepages have at least those characters, at the expected places. BO So it is not so much the nature of text mode, but more an artifact of the BO direct-video approach. If all text mode apps would write using, say, ANSI I don't think this has to do with direct-video approach, but with the assumption that cp437 ist active instead. BO sequences (fast enough with NANSI.SYS) then the ANSI driver could do the BO translation. But the MSDOS ANSI.SYS was slow and not loaded by default so I don't think so. In text mode, every display of characters ends up in a direct-video approach, regardless whether a program does this by itself or uses INT 21h function calls. In text mode there is no way to display mixed box chars under cp850! Sure, NANSI itself uses direct video access, but (N)ANSI.SYS could do a translation. This is what the Linux console does: its console driver first translates from the used character set to Unicode, and then maps the Unicode to the font used. So if you boot Linux using the codepage 437 font, then you can still use iso8859-1 as the character set without changing the font in the video hardware -- as long as there is an equivalent position of the iso8859-1 character in cp437. Boxes can also be drawn using VT100 sequences. This is impractical under DOS *because* so many programs use direct video access instead of the console driver. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
RE: [fd-dev] Codepage IDs
On Fri, 22 Nov 2002, Matthias Paul wrote: On 2002-11-20, Arkady V.Belousov wrote: This program is not required. It required only if you wish to switch codepages on the fly, but if you work only with one codepage, you may (should) initialize it with COUNTR= statement. Of course, in MS-DOS MODE and KEYB without NLSFUNC loaded will fail to load fonts/layouts other than pointed in COUNTRY=. This is correct, but still, a COUNTRY.SYS file parser is needed not only for NLSFUNC, but also for FreeDOS' DOS BIOS. Well FreeDOS doesn't have a DOS BIOS as such. There's kernel.sys which consists (presently) of a) PGROUP 256 bytes at 60:0, used for startup and the init code's PSP b) TGROUP ~1500 bytes at 70:0: small device drivers (CON, AUX, PRN), assembly interface code, intxx trampolines, XMS callers to enable and disable the HMA. c) DGROUP ~5000 bytes now at 00eb:0. The DOS DS: LOL, SDA and so on, COUNTRY tables, constant strings and the low deblock BUFFER (this buffer is actually part of our SDA). d) HGROUP ~4-43000 bytes: main DOS code + block and clock device drivers. e) I_GROUP: ~18000 bytes init code and data: config.sys parser, block device driver init, main init. I_GROUP feels much like a normal .COM DOS program: after setting int vectors it simply calls int21 to open and read config.sys. Discarded after init. But I guess you mean the roughly same thing when I say that once nlsload.c gets fixed up it will be part of I_GROUP. Bart -- list options/archives/etc.: http://www.topica.com/lists/fd-dev unsubscribe: send blank email to: [EMAIL PROTECTED] ==^ This email was sent to: archive@mail-archive.com EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Rv5.bbRv4l.YXJjaGl2 Or send an email to: [EMAIL PROTECTED] T O P I C A -- Register now to manage your mail! http://www.topica.com/partner/tag02/register ==^
