Spheris

Aaron, I'll have some ideas - but it'll have to be later today when I get them in. We have a little bit of a mess to deal with here this morning

{deleted}

Very welcome Just don't crucify me for what I put in the hotfix roll-up thread

{deleted}

Not quite following. But i'll try to shore it up. Maybe its just the whole forum typing thing getting in the way. the =000 are examples of the ordering for the folder names, you can attach the 0's to the beginning of the folder names or swap them for the 000 paradigm. Either way will work fine OemPnPDriversPath="Drivers\000;Drivers\001;Drivers\002\2KXP_INF;Drivers\002;Drivers\003;Drivers\004;Drivers\005;Drivers\006" that is absolutely correct as to the syntax of what we're doing here The missing numbered folders. left out for clarity. I can give you a mirror listing of my own driver order if you want, but went with what you had given me as to not confuse the issue. I only set them by the order they would load in the mouse and key are the last drivers inserted after reboot as they replace the setup gui runtime drivers.

Drivers\intel_chipset; = 000 Drivers\onboard_gigabit =001 Drivers\catalyst3_7\2KXP_INF; =002 Drivers\tv_out\WDM_XP; = 003 Drivers\mitsubishi_monitor; =004 Drivers\logitech_mouse; =008 Drivers\logitech_key; (keyboard) =009 should look something like this

TRM, generally the same as RTM but usually used for non public ISO builds (i.e. OEM builders and vendors only) Tactical Release to Manufacturing as in marketing damage control

The release will be the 24th to RTW and TRM to OEM Distributors. There will be a slipstreamed ISO for OEMs, there will be no slipstream for the RTW package But the RTW will support slip through webmedics method as I have already tried it within that framework last evening One catch this addresses only the "laziness" updates and does not address known hotfix flaws in the OS operation itself. So do not expect any performance related improvements or differences

Aaro, you're on the right track but try it with the folder renamed in successive order Drivers\intel_chipset; = 000 Drivers\logitech_key; (keyboard) =009 Drivers\logitech_mouse; =008 Drivers\catalyst3_7\2KXP_INF; =002 Drivers\tv_out\WDM_XP; = 003 Drivers\mitsubishi_monitor; =004 Drivers\onboard_gigabit =001 take a sharp look at the alpha order and you get the idea

Incorrect but only by a margin I wish to hell I could get the knowledge base to put this sort of thing out there Install order MUST follow as this $OEM$<DriverFolder> order of install follow as listed base chipset Scsi I/O usb/firewire ata etc cpu video sound accessory base hal/keyboard/video/mouse are handled by the setup runtime drivers during gui setup and upon first boot are replaced with the updated drivers best practice scenario is that the seek for bundled & updated drivers is by practical alpha/numeric sort - I'd recommend using 000-009 for sort order on drivers with base chipset at folder 000 and so on Those set the stage for best result on install loader no running afoul and setting out of sequence resource registrations one caveat to the intel and via system drivers there was a problem with driver sets previous to 1003 for the intel and 2.46 for via recognizing their smbus modules. A problem supposedly fixed as 1003 has hql'd and I am going to guess that via fixed their between 2.46&48 (I tend to ignore via for obvious reasons - great idea, crap manufacturing but that only what the guys at hql tell me. I've never owned one to know.

Try it with the 3.7 cats, there is a logged problem with the monitors and use the cumulative installer which will put all the components necessary in their drivers are becoming much more tied to the control panel than they should be, and are risking their hql status for extra "bounty" points in the race against nvidia. but from there, install order for them should be only after install of the intel or via chipset driver, THERE IS THE RISK OF IMPROPER install otherwise from this as the agp and pci/smbus cannot set resource correctly and to be honest. I do not believe it will install correctly unless the chipset is installed and booted first. Side effects can go from blue screens to impaired performance but again its by individual system and these are just practice recommendations. Your SIF is formated correctly but cannot address the cat drivers as the SIF is really only directly affecting the internal drivers and not external items. a second monitor in a multimonitor setup so be prepared to go to the panel to set it up in any case One last note, if the monitor is not DDC and PNP compliant (and believe it alot of them aren't even if they are stickered to be) It cannot determin res and refresh and will have to be set manually in any case Let me know if any of those things factor in. Thank you for calling Microsoft. This Thread has been recorded for quality assurance purposes

Bachus, I'll need the cat version the install order you set from the cmd or sif and the absolutes of what went wrong. get them in the thread when you can and we'll work it out

Medic, the tool is for in-house use, true it does generate an MS header but is NOT the tool used for serialization of the CDs themselves, that is located elsewhere in the disc structure.

Aaron, Shortcut or any extraneous files with commented alterations either OEM or ANSI should be Placed Last to the compile dir, Compile with the J1 or U1 switch only and ignore the secondary switches as the Joilet format overrides in any case. I use the OEM to ANSI for clarifying purposes in house as the vendors provide some of these files the USB file does not reproduce the error here (went ahead and tested just to be sure)

Right on the money, roll with it medic

CDIMAGE 2.47 CD-ROM Premastering Utility Copyright © Microsoft, 1993-2000. All rights reserved. For Microsoft internal use only. Usage: CDIMAGE [options] sourceroot targetfile -l volume label, no spaces (e.g. -lMYLABEL) -t time stamp for all files and directories, no spaces, any delimiter (e.g. -t12/31/91,15:01:00) -g encode GMT time for files rather than local time -h include hidden files and directories -n allow long filenames (longer than DOS 8.3 names) -nt allow long filenames, restricted to NT 3.51 compatibility (-nt and -d cannot be used together) -d don't force lowercase filenames to uppercase -c use ANSI filenames versus OEM filenames from source -j1 encode Joliet Unicode filenames AND generate DOS-compatible 8.3 filenames in the ISO-9660 name space (can be read by either Joliet systems or conventional ISO-9660 systems, but some of the filenames in the ISO-9660 name space might be changed to comply with DOS 8.3 and/or ISO-9660 naming restrictions) -j2 encode Joliet Unicode filenames without standard ISO-9660 names (requires a Joliet operating system to read files from the CD) When using the -j1 or -j2 options, the -n, -nt, and -d options do not apply and cannot be used. -js non-Joliet "readme.txt" file for images encoded with -j2 option (e.g. -jsc:\location\readme.txt). This file will be visible as the only file in the root directory of the disc on systems that do not support the Joliet format (Windows 3.1, NT 3.x, etc). -u1 encode "UDF" file system along with mirror ISO-9660 file system (-n, -nt, -d, -c, or -j1, -j2 options apply to ISO-9660 portion) -u2 encode "UDF" file system without a mirror ISO-9660 file system (requires a UDF capable operating system to read the files) -us non-UDF "readme.txt" file for images encoded with -u2 option (e.g. -usc:\location\readme.txt). This file will be visible as the only file in the root directory of the disc on systems that do not support the UDF format. -b "El Torito" boot sector file, no spaces (e.g. -bc:\location\cdboot.bin) -s sign image file with digital signature (no spaces, provide RPC server and endpoint name like -sServerName:EndPointName) -x compute and encode "AutoCRC" values in image -o optimize storage by encoding duplicate files only once -oc slower duplicate file detection using binary comparisons rather than MD5 hash values -oi ignore diamond compression timestamps when comparing files -os show duplicate files while creating image (-o options can be combined like -ocis) -w warning level followed by number (e.g. -w4) 1 report non-ISO or non-Joliet compliant filenames or depth 2 report non-DOS compliant filenames 3 report zero-length files 4 report each file name copied to image -y test option followed by number (e.g. -y1), used to generate non-standard variations of ISO-9660 for testing purposes: 1 encode trailing version number ';1' on filenames (7.5.1) 2 round directory sizes to multiples of 2K (6.8.1.3) 5 write \i386 directory files first, in reverse sort order 6 allow directory records to be exactly aligned at ends of sectors (ISO-9660 6.8.1.1 conformant but breaks MSCDEX) 7 warn about generated shortnames for 16-bit apps under NT 4.0 b blocksize 512 bytes rather than 2048 bytes d suppress warning for non-identical files with same initial 64K -k (keep) create image even if fail to open some of the source files -m ignore maximum image size of 681,984,000 bytes -a allocation summary shows file and directory sizes -q scan source files only, don't create an image file NOTE: Many of these options allow you to create CD images that are NOT compliant with ISO-9660 and may also NOT be compatibile with one or more operating systems. If you want strict ISO and DOS compliance, use the -w2 warning level and correct any discrepencies reported. YOU are responsible for insuring that any generated CDs are compatible with all appropriate operating systems. Also note that Microsoft company information is placed in the image volume header, so don't use this program to generate CDs for companies other than Microsoft. For those wondering what Aaron and I are on about

Side note I usually do a run through compile of source files with the non joliet extensions and then dump the files from the resulting image because it removes certain errata from the oem and ansi structures in house. the usb components I'll check on in the next few minutes, we don't build them so there might be a name structure problem (intel provided)

make sure you test it thoroughly though, part of the dx reg structure is nested in the hives on the iso so you can and will run across some cross writes with it..make sure sequence is before reboot but post component register

Here's an example syntax for you since I don't see a reply yet C:\CDIMAGE.EXE -lXXPSD2_DI -t04/24/2003,12:00:00 -c -h -j1 (-u1 in the case of cumulative dvr testing) -x -bC:\BOOTIMG.IMG C:\BUILDDIR C:\XXPSD2_DI.ISO

Can you give me the full command line syntax? I can get you an answer straight back.

Will have a pack installer (orca based) for you to download and work with to slipstream later on today

We use CDIMAGE 2.47 and use CDBURN/DVDBURN to write and finalize no runs, no drips, no coasters

Medic, looks like it might be possible. I'll get back to you later today after I get through here and have a talk with one of the Athena developers. I also have a question about the inf used for the GUI-setup portion of the install so if you can pitch me a clue about it. I'd like to take a look at it before tomorrow while I have this guys ear

Sounds good I'll shoot you a PM if I can get anyone to take a few and give me an alternate on .net.

There might be an alternate, but I'd have to talk to someone in the server group tomorrow to know for sure. But as I understand it, its rolled into 2k3 in the same way as the svcpack method right now, just using an inf, same as the inf your using for addon apps