Menu
▾
▴
Re: [Freedos-devel] The List
|
From: Fritz M. <fri...@ma...> - 2026-04-12 12:37:11
|
<html><head></head><body><div style="font-family: Verdana;font-size: 12.0px;"><div>Hi Jerome,</div> <div>I only made a first step as I am not sure if it will help you. As sending zip files always causes problems, I send it as text. This is the result of Google AI with INTERRUP.A. Please tell me if it makes sense to continue this work and if another format would be</div> <div>better for you.</div> <div> </div> <div>Willi</div> <div> </div> <div> </div> <div> <div>Here is the analysis of the first section (approx. 1500 lines) of the Interrupt List update, formatted for an American developer.</div> <div> </div> <div>[TYPO] – Spelling & Grammar</div> <div> DISCLAIMER (Line 17):</div> <div> Error: "...and please let us know hether or not it works..."</div> <div> Correction: "...and please let us know whether or not it works..."</div> <div> TABLE 00010 (Video Modes Header):</div> <div> Error: "colors disply scrn system"</div> <div> Correction: "colors display scrn system"</div> <div> INT 01 (Debugging Exceptions, Line 108):</div> <div> Error: "...will return to following instruc" (Word cut off at end of line).</div> <div> Correction: "...will return to following instruction"</div> <div> Consistency Note (US English):</div> <div> Observation: Several entries (e.g., INT 03, Line 163) use British spelling: "initialises", "initialising".</div> <div> Suggestion: Since RBIL traditionally uses US English and the recipient is American, these should be changed to "initializes" and "initializing" for consistency.</div> <div> </div> <div>[TECH] – Technical Inconsistencies</div> <div> Terminology - "Real 86 Mode" (INT 0C/0D, Lines 426 & 536):</div> <div> Entry: "...collides with IRQ #4 default handler. in Real 86 Mode, the two may be distinguished..."</div> <div> Problem: "Real 86 Mode" is a non-standard term. It risks confusion with "Virtual 8086 (V86) Mode."</div> <div> Recommendation: Use the standard term "Real Mode."</div> <div> MS-DOS 2018 Release (INT 01/03 BUGS section):</div> <div> Entry: "MS-DOS Debug from the 2018 release of MS-DOS v2 fails to restore..."</div> <div> Note: This refers to the Microsoft Open Source release on GitHub. It might be helpful to specify "MS-DOS v2.0 (2018 GitHub release)" to distinguish it from the original 1983 binaries, as build environments can affect behavior.</div> <div> Watcom Debugger (INT 03, Line 206):</div> <div> Entry: "...define a public byte variable named ___WD_Present..."</div> <div> Check: The entry lists three underscores. While some Watcom versions used this, most C-style mangling uses two. It is worth double-checking if the third underscore is a typo or specific to this interface.</div> <div> </div> <div>[FORMAT] – Layout & Structure</div> <div> Table 00006 (Keyboard Scan Codes):</div> <div> The column containing scan codes 63h through 6Bh (F16 through F24) is slightly misaligned (offset by one or two spaces) compared to the columns to the left. This can cause issues with fixed-width text parsers used by some RBIL viewers.</div> <div> Table 00010 (Video Modes):</div> <div> The spacing in the "Colors" column for mode 0Dh and 0Eh is slightly inconsistent.</div> <div> </div> <div>[SUMMARY]</div> <div>The structure of the divider lines (the -------- markers) is perfect and adheres to the Release 61/62 standard. The metadata in the header is also correct.</div> <div> </div> <div>Here is the technical analysis of the second section (INT 10h Video Services) of the Interrupt List update, formatted for an American developer.</div> <div><br/> [TYPO] – Spelling & Grammar</div> <div> INT 10/AH=0Dh (Line 1152):</div> <div> Error: "...specified pixel in grahics modes"</div> <div> Correction: "...specified pixel in graphics modes"</div> <div> INT 10/AX=1101h (Line 1317):</div> <div> Error: "...the "monochrome" patters are 8x14 pixels..."</div> <div> Correction: "...the "monochrome" patterns are 8x14 pixels..."</div> <div> INT 10/AX=1111h (Line 1391):</div> <div> Error: "...the "monochrome" patters are 8x14 pixels..." (Repeated error from 1101h).</div> <div> Correction: "...the "monochrome" patterns are 8x14 pixels..."</div> <div> Table 00011 (Everex Modes, Line 1060):</div> <div> Observation: Mode 62h uses the abbreviation "Vwpt TC" while modes 70h-77h use the full word "Viewpoint TC".</div> <div> Suggestion: Use "Viewpoint" for mode 62h to maintain consistency within the table.</div> <div> Spelling Consistency (US vs. UK):</div> <div> Observation: Labels 1.13 and 9.3 in your previous file check mentioned "Driver" (which is correct), but within the RBIL text (e.g., Line 1314), terms like "initialised" and "specifiers" appear.</div> <div> Recommendation: Since the target is an American developer, ensure "initialized" (with a 'z') is used throughout.</div> <div>[TECH] – Technical Inconsistencies & Verifications</div> <div> INT 10/AH=0Eh (Teletype Output, Line 1167):</div> <div> Note: The entry adds a 2025 SvarDOS bug report regarding AH corruption on original IBM PCs.</div> <div> Check: This is a highly specific addition. Verify if the "AH corruption" occurs only when scrolling or on every character output, as this significantly impacts TTY drivers.</div> <div> INT 10/AH=12h/BL=30h (Reference in Note 17, Line 1017):</div> <div> Verify: Note 17 mentions "VGA may also require using INT 10/AH=12h/BL=30h". This is the "Select Scan Lines" function. It might be helpful to explicitly state "to select 200/350/400 scan lines" for clarity.</div> <div> INT 10/AX=1110h (Recalculate Parameters, Line 1373):</div> <div> Note: "R14 = bytes/char ; underline loc (*** BUG: should be 1 less ***)"</div> <div> Check: This is a well-known BIOS bug. It might be worth adding if this bug is present in FreeDOS's VGA BIOS implementation or if it has been patched there.</div> <div>[FORMAT] – Layout & Structure</div> <div> Mode Table 00010 (Continuation):</div> <div> Mode 61h (Line 847): "...ATI VGA Wonder,VGA Wonder+"</div> <div> Correction: Add a space after the comma for readability ("...VGA Wonder, VGA Wonder+").</div> <div> Table Alignment:</div> <div> In the 256-color mode sections (Mode 5Fh through 67h), the "Colors" column (e.g., 256/256K) is sometimes slightly offset depending on the length of the "Resolution" string. Ensure the start of the "System" column (e.g., A000 or B800) remains strictly aligned for fixed-width viewers.</div> <div>[SUMMARY]</div> <div>The technical depth regarding the 2020 FreeDOS kernel commits and 2025 SvarDOS bug reports shows excellent attention to modern DOS development. Fixing the "patters" and "grahics" typos will bring the document up to professional standards.</div> <div> </div> <div>Here is the technical analysis of the third section (INT 10h continued, including Font Information, Flat-Panel, and XGA services) of the Interrupt List update.<br/> [TYPO] – Spelling & Grammar</div> <div> Table 00053 (Flat-Panel Information, Lines 1509–1511):</div> <div> Error: "...(0000h if nonochrome)" — appears three times in the Green, Red, and Blue levels descriptions.</div> <div> Correction: "...(0000h if monochrome)"</div> <div> Table 00058 & 00059 (Brightness/Contrast Settings, Lines 1530 & 1533):</div> <div> Error: "...set operation was succesful..."</div> <div> Correction: "...set operation was successful..." (Needs two 's' in the middle).</div> <div> INT 10/AH=1Bh (Functionality/State Info, Lines 1572–1573):</div> <div> Error: "...return funtionality/state information" (in the register description) and "...address of static funtionality table" (in Table 00040).</div> <div> Correction: "...return functionality/state information" and "...address of static functionality table".</div> <div> INT 10/AH=12h/BL=32h (Line 1438):</div> <div> Error: "...display adapters I/O registers"</div> <div> Correction: "...display adapter's I/O registers" (Possessive apostrophe required).</div> <div> INT 10/AX=1E03h (SeeAlso section, Line 1515):</div> <div> Error: "SeeAlso: AX=1E00h,AX=1E02h**.**AX=1E04h"</div> <div> Correction: "SeeAlso: AX=1E00h,AX=1E02h**,**AX=1E04h" (Dot instead of comma).</div> <div>[TECH] – Technical Inconsistencies & Verifications</div> <div> INT 10/AH=1Ch (Save/Restore State, Line 1618):</div> <div> Verification: The note stating "this function is not supported when DOS/V is running" is an excellent addition for modern compatibility checking, as DOS/V hooks many video interrupts for DBCS (Double Byte Character Set) handling.</div> <div> INT 10/AH=1Dh (Tseng Bug, Line 1640):</div> <div> Check: The entry mentions a "fencepost error." For an American developer, it might be clearer to specify if this results in an Infinite Loop or an Illegal Instruction exception, as "random location" is technically broad.</div> <div> Table 00060 (XGA DMQS, Line 1673):</div> <div> Verify: "paragraph--multiply by 10h to get physical address". Usually, segment/paragraph math is straightforward, but in XGA context, these often refer to absolute 32-bit addresses in extended memory. Ensure the terminology "paragraph" isn't misleading for a protected-mode developer.</div> <div>[FORMAT] – Layout & Structure</div> <div> Table 00040 (State Information, Line 1573+):</div> <div> Offset 27h and 29h notes: The entry correctly flags a Trident BIOS bug where these fields are not set. This is high-quality documentation.</div> <div> Table 00028 (ATI Mode Table, Line 1409):</div> <div> The indentation for the bitfields of CRTC registers (offsets 00h through 11h) uses a mix of tabs and spaces. To ensure alignment in standard text viewers (like LIST.COM or INTER61), I recommend converting all to fixed spaces.</div> <div> Table 00033 (Cirrus Frequency, Line 1494):</div> <div> The "43i Hz" (interlaced) notation is standard, but some entries in this block use "int" (Line 1501). It is better to use either "i" or "-int" consistently throughout the file.</div> <div>[SUMMARY]</div> <div>The addition of the IBM SurePath BIOS "Private" functions and the Diamond Stealth64 installation check (AX=1DAAh) provides significant value to the 2026 update. Correcting the "nonochrome" and "funtionality" typos will ensure the list maintains its historical reputation for accuracy.</div> <div> </div> <div> </div> <div>Here is the analysis of the fourth section (3270 PC through VESA VBE/AI Services) of the Interrupt List update.<br/> [TYPO] – Spelling & Grammar</div> <div> INT 10/AH=4Bh (Hercules GRAFIX, Line 1832):</div> <div> Error: "...the character is writting using the drawing function..."</div> <div> Correction: "...the character is written using the drawing function..."</div> <div> INT 10/AH=4Bh (FRIEZE API, Line 1814):</div> <div> Error: "...vector of screen->printer color correspondnces"</div> <div> Correction: "...vector of screen->printer color correspondences"</div> <div> INT 10/AX=4F01h (VBE Mode Info, Line 1928):</div> <div> Error: "...bit 11: user user-specified CRTC refresh rate..."</div> <div> Correction: "...bit 11: user-specified CRTC refresh rate..." (Remove duplicate word).</div> <div> Table 00092 (VBE/AF SetBackColor, Line 2073):</div> <div> Error: "Call VBE/AF v1.0P function 'SetbackColor' with:"</div> <div> Correction: "Call VBE/AF v1.0P function 'SetBackColor' with:" (Capitalize the 'B' for consistency with 'SetForeColor' in Table 00091).</div> <div> INT 10/AX=1E07h & AX=1E08h (Flat-Panel Settings, Lines 2235 & 2252):</div> <div> Error: "...set operation was succesful..."</div> <div> Correction: "...set operation was successful..." (Ensure double 's' in the middle).</div> <div>[TECH] – Technical Inconsistencies & Verifications</div> <div> Table 00080 (VBE Mode Attributes, Line 1965):</div> <div> Note: "application must call EnableDirectAccess before calling bank-switching functions."</div> <div> Check: This is a critical requirement for VBE/AF (Accelerator Functions). It might be worth adding a cross-reference here to Table 00108, where the function is defined, to help developers find the calling convention.</div> <div> INT 10/AX=4F0Bh (VBE/AF Context Buffer, Line 2049):</div> <div> Note: "application must never pass X coordinate greater than this value to the SetClipRect function (values are NOT range-checked!)"</div> <div> Verification: This is an important warning for high-performance drivers. Ensure the exclamation mark and capitalization remain, as it highlights a potential system crash risk.</div> <div> INT 10/AX=4F13h (VBE Audio Interface, Line 2197):</div> <div> Note: "BX = version (0010h for 1.00 -- note nonstandard value!)"</div> <div> Observation: This is a correct and vital observation for VBE/AI, as most VESA versioning uses BCD (e.g., 0100h for 1.00). Keep this note prominent.</div> <div>[FORMAT] – Layout & Structure</div> <div> Table 00114 (Wave Feature Bits, Lines 2225–2249):</div> <div> The list of frequencies (8000hz, 11025hz, etc.) is very long.</div> <div> Suggestion: Historically, RBIL uses "Hz" (capital H) rather than "hz". Changing this would improve the professional look of the technical data.</div> <div> Table 00113 (WAVE Info Structure, Line 2217):</div> <div> The field "unused data" at offset 6Dh shows "3 BYTEs". Check if this aligns with the structure length at offset 04h (0000007Eh).</div> <div> Calculation: 6Dh + 3 = 70h. Offset 70h is the next field. The alignment is correct.</div> <div>[SUMMARY]</div> <div>The VBE/AF and VBE/AI sections are exceptionally well-documented in this update, including the "draft documentation" caveats which are standard for RBIL. Fixing the "writting" and "correspondnces" typos will finalize this section nicely.</div> <div> </div> <div><br/> Here is the technical analysis of the fifth section (3270 PC Tables through Chips & Technologies Extensions) of the Interrupt List update, prepared for an American developer.<br/> [TYPO] – Spelling & Grammar</div> <div> LIR Configuration (Line 2445):</div> <div> Error: "...codepage support package develoved by LABAS u.a.b."</div> <div> Correction: "...codepage support package developed by LABAS u.a.b."</div> <div> Matrox PINS Data (Table 00126, Offset 26h):</div> <div> Error: "maximum RAMDAc frequency???"</div> <div> Correction: "maximum RAMDAC frequency???" (Capitalization consistency).</div> <div> VBE/AI Error Codes (Table 00116, Offset 44h):</div> <div> Observation: "...06h don't understand the PCM size/format"</div> <div> Correction: "...06h does not understand the PCM size/format" (RBIL usually avoids contractions in technical descriptions).</div> <div> LIR Alphabet Flags (Line 2447):</div> <div> Error: "...alphabet flags(see #04120)"</div> <div> Correction: "...alphabet flags (see #04120)" (Missing space before parenthesis).</div> <div>[TECH] – Technical Inconsistencies & Verifications</div> <div> Hexadecimal Notation Consistency (Table 00116 & 00124):</div> <div> Entry: "...vendors can add DevChks above 0x80" (Line 2276) vs "...vendors can add DevChks above 0080h" (Line 2383).</div> <div> Recommendation: Use the traditional RBIL suffix notation "0080h" consistently throughout. The C-style 0x notation is rare in the original list.</div> <div> EDID Manufacturer ID (Table 00136):</div> <div> Observation: The entry for "LEN" (Lenovo Group Ltd.) is a modern addition.</div> <div> Check: Verify that the "Big-endian manufacturer ID" logic in Offset 08h (Line 2516) correctly generates "LEN" based on the VESA compressed ID formula. (Calculation: 'L'=12, 'E'=5, 'N'=14. Binary: 01100 00101 01110 -> 30AEh). If the BIOS returns 30AEh, the entry is correct.</div> <div> Matrox Millenium URL (Line 2432):</div> <div> Error: http:\\grafi.ii.pw.edu.pl\gbm\matrox</div> <div> Correction: http://grafi.ii.pw.edu.pl/gbm/matrox (Standard URL syntax uses forward slashes).</div> <div> Volume Services (Table 00124, Offset 1Ch):</div> <div> Check: The entry point at 1Ch says "User master volume" and "application master volume".</div> <div> Recommendation: Check if these should both be capitalized ("User" / "Application") for consistency.</div> <div>[FORMAT] – Layout & Structure</div> <div> 3270 PC Function Flags (Table 00064):</div> <div> The bitfield descriptions are clean and follow the standard 7-0 order. This is correct.</div> <div> EDID Timings (Table 00130 & 00131):</div> <div> The use of the "@" symbol (e.g., 720x400 @ 70 Hz) is consistent and helpful.</div> <div> Alignment in Table 00136 (Manufacturer IDs):</div> <div> The spacing between the three-letter ID and the company name is mostly consistent, but double-check the "Lenovo" entry (Line 2611) as it seems to have one less space than the surrounding entries.</div> <div>[SUMMARY]</div> <div>The inclusion of Lenovo and the Matrox PINS v3.0 data adds excellent modern and legacy value. Correcting the "develoved" typo and the mixed hex notations (0x80 vs 80h) will ensure the list remains the definitive technical resource.</div> <div> </div> <div><br/> Here is the technical analysis of the sixth section (Chips & Technologies continuation through Compaq Extensions) of the Interrupt List update, prepared for an American developer.<br/> [TYPO] – Spelling & Grammar</div> <div> INT 10/AX=5F02h (C&T 65530 BIOS, Line 2750):</div> <div> Error: "...new vlaue for control register..."</div> <div> Correction: "...new value for control register..."</div> <div> INT 10/AX=5F05h (C&T 64300 BIOS, Line 2831):</div> <div> Error: "...program new clock and CRT parametes"</div> <div> Correction: "...program new clock and CRT parameters"</div> <div> INT 10/AX=5F53h (C&T Extended BIOS, Line 2911):</div> <div> Error: "...enable buffer/acelerator"</div> <div> Correction: "...enable buffer/accelerator"</div> <div> Table 00201 (SOLLEX Info Structure, Line 3097):</div> <div> Error: "...Seiko Epson Corporation intended to provided a standardized..."</div> <div> Correction: "...Seiko Epson Corporation intended to provide a standardized..."</div> <div> Table 00208 (SOLLEX Displays Attached, Line 3230):</div> <div> Error: "...attached to digital outpt"</div> <div> Correction: "...attached to digital output"</div> <div> Table 00212 (Compaq VDU Support, Line 3433):</div> <div> Error: "...132-column support availble"</div> <div> Correction: "...132-column support available"</div> <div> INT 10/AX=BF04h (Compaq Mode Switch, Line 3462):</div> <div> Error: "...1993/04/08 sytem ROM..."</div> <div> Correction: "...1993/04/08 system ROM..."</div> <div>[TECH] – Technical Inconsistencies & Verifications</div> <div> INT 10/AX=5F10h (C&T Linear Memory Info, Line 2854):</div> <div> Error: "SI;DI = linear memory size"</div> <div> Correction: "SI:DI = linear memory size" (RBIL standard uses a colon for register pairs).</div> <div> Table 00184 (C&T F65520 Status, Line 2872):</div> <div> Inconsistency: "Bit(s) 15-0 reserved" is listed at the top, but then sub-bits like 6-5 and 4 are defined.</div> <div> Recommendation: Change the first line to "15-7 reserved" to avoid logical contradiction.</div> <div> Realtek RTVGA Bugs (Lines 2715, 2736, 2786, 2795):</div> <div> Note: The entry mentions "improper stack restoration code" in version v3.C10 causing the return value of AX to match the input value.</div> <div> Observation: This is a very helpful technical detail for developers debugging legacy drivers. It highlights why detection might fail on certain Realtek cards.</div> <div> INT 10/AX=7000h (Everex, Line 2977):</div> <div> Verify: The return AL = 70h is only for Trident-based Everex cards. It might be useful to clarify if non-Trident Everex cards return a different signature or simply don't support AX=7000h.</div> <div>[FORMAT] – Layout & Structure</div> <div> Table 00192 (Hunter 16 Control Block, Line 2963):</div> <div> Terminology: "center X co-ordinate"</div> <div> Suggestion: US English typically uses "coordinate" (no hyphen). While both are technically correct, "coordinate" is standard in American technical writing.</div> <div> INT 10/AX=5F5Ch (C&T Vertical Compensation, Line 2942):</div> <div> Consistency: "0Ah disable text stretching" uses a tab/space mix that looks slightly wider than the surrounding entries. Recommend checking the alignment for 8-character tab stop viewers.</div> <div>[SUMMARY]</div> <div>The addition of the Husky Hunter 16 laptop functions and the SOLLEX (Seiko Epson) extensions significantly broadens the hardware coverage of the list. Fixing the semicolon in SI;DI and the "acelerator" typo will maintain the high technical quality expected of RBIL.</div> <div> </div> <div> </div> <div> </div> <div> </div> </div> <div> </div> <div> </div> <div> </div> <div> </div> <div> </div> <div> </div> <div> </div> <div> <div> <div name="quote" style="margin:10px 5px 5px 10px; padding: 10px 0 10px 10px; border-left:2px solid #C3D9E5; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"> <div style="margin:0 0 10px 0;"><b>Sent:</b> Tuesday, April 07, 2026 at 8:48 PM<br/> <b>From:</b> "Jerome Shidel via Freedos-devel" <fre...@li...><br/> <b>To:</b> "FreeDOS Developers" <fre...@li...><br/> <b>Cc:</b> je...@sh...<br/> <b>Subject:</b> [Freedos-devel] The List</div> <div name="quoted-content">Hi all,<br/> <br/> Work continues on updating "The List". [1]<br/> <br/> There has been a lot of additional information from ECM, which has been added.<br/> <br/> As well as, a few things from other contributors like myself.<br/> <br/> I don’t know how soon it will be until we have our first Release (62).<br/> <br/> But, if you just cannot wait, you can fetch a nightly build of "The List" from my server [2].<br/> <br/> Again, this is a big long term project. If you have information you wish to share in The List, you are welcome to help contribute.<br/> <br/> You don’t need to be a tech wizard to help out. There is a lot of stuff in there. Just fixing typos, punctuation and grammar can be big help.<br/> <br/> :-)<br/> <br/> [1] <a href="https://github.com/LoopZ/TheList" target="_blank">https://github.com/LoopZ/TheList</a><br/> [2] <a href="https://fd.lod.bz/redist/devel/TheList/" target="_blank">https://fd.lod.bz/redist/devel/TheList/</a><br/> <br/> <br/> <br/> _______________________________________________<br/> Freedos-devel mailing list<br/> Fre...@li...<br/> <a href="https://lists.sourceforge.net/lists/listinfo/freedos-devel" target="_blank">https://lists.sourceforge.net/lists/listinfo/freedos-devel</a></div> </div> </div> </div></div></body></html> |
