Delprat

i'm late, sorry about my previous 4/ in post #2, the question was "why are you needing your monitor driver ?" The inf file contains only settings for 1280x1024 (like refresh rate), and a color profile. Is your monitor misfunctionning with default driver ? about your post #9, you're right : this will copy the ICM, providing you set the correct directory ID in txtsetup.sif (the first number "1" is correct, that's the others), the ICM should go in system32\spool\drivers\color. about your post #4 (monitor_test.inf), as far as i can tell, it seems correct, except for this section : [ControlFlags] ExcludeFromSelect=Monitor\Default_Monitor You should not modify it : it's meant to disallow the user to install the "Default Monitor" using the hardware wizard. If you set your monitor instead of the default one, it will be hidden from the "database" But i may have misunderstood your goad : you may want to hide your LG monitor, in this case, just add it after the default one, using a coma : ExcludeFromSelect=Monitor\Default_Monitor, Monitor\GSM4A90 (nota: the ".nt" is pointless 'coz you're modifying a NT-specific file) Also, check if LGL1915s.icm is not a copy of the default sRGB.icm. You can check the other monitor*.inf, maybe one of them include ICM files copying ? (if so, do not modify "monitor.inf" but the "monitor*.inf") ++ edit: the root ? wtf ? 8/

Hopefully for you, it seems like you're connected now. Sadly for everyone else, your caps lock key has prefered to lock your keyboard and go away ('coz it did saw your registry being destroyed by your chainsaw... er, sorry, your "registry cleaner"). Since you did disabled windows' built-in id***-proofing, you must know how to revert by yourself. (when you disabled it, you were prompted "it may be dangerous, are you sure ?", doesn't you ?) With so few details, my only advices will be : 1. read the board rules 2. read again the board rules, you missed something because you're reading too fast 3. read the board rules NOW because you forgot to do both the first and the second advice before reaching this line thinking it's unuseful to read the board rules before going to the next advice 4. go to advice 1 until you know the board rules by heart 5. repair not working apps 6. now, you're reached the end. can you tell me how many times i wrote "the board rules" without reading again my post ? If not, you were not atentive enough. ++

Untested ideas : 1/ Remove all the monitor*.inf from c:\windows\inf after windows installation, then install your monitor. In theory, windows will no more be able to re-install its default monitor "driver" (but it may ask you for the CD ?) 2/ Let windows install its default monitor, then disable it in device manager, then install yours. hopefully, windows will remember this "disable" and no more bugs you (but it may then think you want no display and *really* disable the monitor ?) 3/ In monitor.inf, search for this : [1280VESA60] HKR,"MODES\640,480",Mode1,,"60.0,50.0,-,-" HKR,"MODES\1280,1024",Mode1,,"48.0,65.0,+,+" HKR,,PreferredMode,,"1280,1024,60" and replace the third line with the one from L1915s.inf : HKR,"MODES\1280,1024",Mode1,,"30.0-83.0,56.0-75.0,+,+" This will set the windows' default monitor settings to be those of your monitor. Then, you'll need to do something for the ICM profile if you really need if. Maybe by adding the line : HKR,,ICMprofile,0,"LGL1915S.ICM" in the above section and copying manually the file. 4/ I have the exact same problem with a IIyama LS902U, but i can live with both monitors "drivers" : windows switches them randomly on boot, but i don't see any difference (i had configured both times ago, and the only *real* change is the description shown). Why can't you ? ++

Before disabling everything, try disabling just the SMART feature for the IDE slots. I've had similar "hangs at winlogo" problems, and it was this dumb feature. (plus, the setup said something about mountmgr, which is related to drives) ++

Not at all, that's the ultimate evidence of Microsoft's lazyness. Take Word, or any MS Office app, select some text, copy, select another text and paste. Did you saw a security alert "You are overwriting your text ! [continue/abort]" ??? But in explorer (which is newer than word), MS devs did feel it is better not to include a real undo mecanism (beside the NTFS journalisation, the recycle bin and the system restore, the WFP/SFC cache, the hotfixes uninstallers, and so on : there's many "undo-like" features in windows that are not "centralized", wasting space and resources...), and to bug the user each time it does a filesystem operation... (and i do say "bug" for a reason : if you're pasting a folder over another, the explorer just says "some things can get replaced, continue/abort ?" even if there's nothing about to be replaced... and when that's the case, seeing precisely what would be much more helping) ..../me feels better now roy1984, you could try alternate explorers or alternate shells ++

Did you tried to replace : HKLM,"%RunOnceEx%\Setup_2",1,,"%1%\WPI\WPI.hta" by HKLM,"%RunOnceEx%\Setup_2",1,,"%01%\..\WPI\WPI.hta" The "magic numbers" you can define in [sourceDisksNames] are to be used only in [sourceDisksFiles]. The "magic numbers" enclosed in percent signs are windows' directory identifiers, so %1% (or %01%) always refers to the directory the .INF file is in. You can't redefine %...% numbers ; but you can define new ones by reading a "custom" value in the registry via "CustomDestination" sections. (i don't know precisely how to use such sections, and they're "unsafe" in this case since you need to provide a default value) ++

I just didn't knew that wildcards are accepted this way Why such things aren't said in "manpages" ? @quijibo > you're welcome. in fact, i answered you because seeing your question answered a problem of mine ++

There is multiple solutions... here is one : You need a %~dp one to change the current dir to the one of you file Then you go to its parent (CD ..) and assign the current dir to a var. And now you can do a classic set substitution to remove the parent part of the path. @echo off setlocal enableextensions enabledelayedexpansion rem the path contains spaces, so it MUST be quoted : set myfile="c:\test path\folder1\Last Directory\somefile.pdf" call :getLastDir %myfile% echo Last Dir is : %LastDir% goto :eof :getLastDir set mypath=%~dp1 pushd "%mypath%" cd .. set mypath=!mypath:%cd%=! set LastDir=%mypath:\=% popd goto :eof With some testing, you will find this solution buggy when the %~dp1 folder does not exist. Thus here is a shorter one (two lines, no need for delayedexpansion, and working even with nonexistent target) : You need to nest two %~ substitutions, second one inside a for loop. First one is a %~p (or %~dp) to extract the path (excluding the last part of it, in your case, the file name + extension) with its trailing backslash. Then you remove this backslash. (with classic set substitution) And now you can use for on the remaining part to do a %~n substitution. @echo off setlocal enableextensions rem the path contains spaces, so it MUST be quoted : set myfile="c:\test path\folder1\Last Directory\somefile.pdf" call :getLastDir %myfile% echo Last Dir is : %LastDir% goto :eof :getLastDir set mypath=%~p1 for %%v in ("%mypath:~0,-1%") do set LastDir=%%~nv goto :eof ++

Using components that generates less heat... i would have said that as clearly. The water cooling is maybe poorly spread, but i've seen air-conditionned rooms hosting servers... and that's a waste of energy when there is only a few computers. Inapropriate software can be easily replaced with the saved money in this case (and all good waterblocks comes with dedicated soft) ; and i'm not sure there is really more maintenance : it's not always easy to keep these rooms "dust clean" in offices, and with some thinking, you can ensure that no liquid escape can create short-circuit. I've also seen "chained" circuits between computers, it seems that, when one is "heavy loaded" and another less, cooling capacity is better used ; but i'm unsure that's realy useful in small setups. Nor mine : just try to sort that matter with some smoke, or better with tamised dust or flour (use a "false" case with false hard drives, cards, etc for that ) 'coz more dust will fall inside the case if the flow is inconstant, insufficient, or too turbulent. In brief, you can make a non-direct equivalence between the amount of extracted dust in a cold test case and the amount of extracted heat in a real case. If you want very accurate results, burn a tire and see how many rubber is not extracted and, more important, where it goes inside the case : as soon as you have turbulences, cleaner places appears, and they are frequently "hot spots" in a real case ; rarely they can be "cold spots" if an "ordered turbulence" appears (like hurricanes are at the world's scale -- sorry i don't know how to explain better) About the T.S.O., it's obvious that having two fans in a tunnel-like structure is a bad idea, since it needs an extremely precise calibration to keep air presure the same between fans (and this will be dependant of the airflow around the S.O., a thing thermaltake engineers has no control over) ++ edit: 500th post ! chanpagne !

The batch : open notepad, type in the line i gave you, and save as "defrag.cmd" (type the quotes in the save as dialog to ensure it will not get .txt added). Lastly point to this file where you pointed to dfrg.msc in the scheduled task. Minimizing the number of installed applications has few impact on fragmentation., but installing/removing apps frequently is a bad idea. ++

You're right on all of this , and reading me back i don't see where i'd say you weren't (hey, we're maybe both right ? ) You're talking about somewhat extreme conditions... I'm just saying that if i were a harddrive (not the ones you're talking about for sure), i'd feel better if i could stay cool whith no fan blowing on me : that's the difference between a windy day and a vveerryywwiinnddyyddaayy, not between a fan next to me and a fan that blows on me. I must insist on the importance of eliminating the turbulences inside the case, they do nothing really good for cooling on the long term, and they increase the static electricity problems... switching to passive or liquid cooling is also a good idea (specifically for the so classic "servers in a room @ 30" problem) btw, i've nothing against you ++

Ever heard of the simpsons ? I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. I will not start trolling about the best defragger. Seriously, since you're normal, you sometimes need to sleep, so you don't *really* need to have a "super-stealth-and-transparent" background defragmenter. just put : defrag.exe c: -f in the batch ++

Long short story : Yes for any "normal" home user Diskeeper is not really much efficient, only faster (way faster) O&O is designed for maniacs of the "i don't want ANY free cluster between my files. Line up in order ! One Two ! One Two !" PerfectDisk features the most colorful interface and acheive similar results than Diskeeper. Obviously, "professionals" (like people running a huge database, or a file server, or ...) will see benefits of switching to one of those paid apps. ++

You're lucky, you're english ! Go there : http://hf.xable.net/index.php Or try out Ryan's update pack. ++

Noh, that's the opposite : 3rd-party apps can be used after. You did said you were looking for an easy solution : writing a specific "reloader" procedure for "after setup in case xpize is already slipstreamed" is not an easy solution when a generic "reloader" procedure can be written. (that's already the way xpero did, even if he did not thought about this case) Thus my first post in this thread : "log slipstreamed resources" and my second "add a hash to this log". This way of patching resources allow the exact same "reloader" code to be used, minimizing the amout of work needed both for writing and debugging. But it's not the simplest nor the easiest : why extracting resources to compare them when it's possible to compare whole files in less time ? (when you slipstream Xpize, no files are partially patched : either they are, either they are not) Having "less system files replacements" is not safer than anything. Sorry for being bold, but having safe files replacements is safe no matter how many replacement there are. Again, sorry for being bold, but extracting-then-compare resources increases the number of operations above the bare necessary which is increasing the risk of error... About the resources IDs, they can be "same", "different" or "partially different". The first case is aready handled. For the second and the third, you're suggesting to ignore and continue, making the xpize reloader no more a reloader whereas it's possible to handle correctly these rare cases and reload the right resources. ++

Place them all in the same folder, for example C:\Hotfixes\, then open a command prompt and type : for %i in (c:\hotfixes\*kb*.exe) do start /wait /low %~dspnxi /q /o /n /z you need to reboot manually at the end This method works only with Type 2 Hotfixes, those a with squared icon (blank inside with blue borders) ++

you should read carefully... "other programs" means not "windows hotfixes" "other programs" means "all the wide variety of programs you can find anywhere" don't be so stubborn... stupid me, you may have interest in MS ? 1/ that's only a detail, but CRC/CRC32 is an unreliable algo ; and MD4/MD5 are known to create collisions also. For safety, another algo needs to be used (SHA for example was never broken AFAIK). 2/ all you say here is right... You presume "there aren't localized resources images". That's true. But adding an image as a resource in an executable is not enough to display this image. There is also (localized or not) "code" (mainly DIALOG(EX) or UIFILE resources) that have references to those images (by their ID). One can easily change this reference, and XPize-like shell packs are blind to this (they update a resource ID that is no more used by the binary file). This is not a "possible (theorical) bug case", since some shell packs does this sort of modifications. Another example : i were talking about sysdm.cpl 'coz there's here on msfn a thread on how to have a "vista-like" one : if you use it, xpize is no more able to reload sysdm.cpl, with or without your resources integrity checks (you will get the dialog layout of the "vistaized" one, with images from both xpize and vista-ized one, and it looks weird, defeating the "reloader" purpose) ++

« some things are not seen » ? like what ? noise ? heat ? or debiliting rays emitted by plasma screens imported from Beta-Centauri ? (just kidding...) What i said is from my experience, that seems to differ than yours. i suppose you know that the heat a hard drive, a CPU or whatever, generates does not depend on how it's cooled, thus having it hitted by the airflow to get it a few degrees lower is useless if its temp remains stable under the limits (that is, having a drive at 35 or 40 will not change anything since different constructors are placing the sensor at different places ; but having it taking one degree per hour is a problem). on the opposite, the global cooling offered by the airflow dramatically depends on the constance of the flow. a turbulent flow will provide poor thermal dissipation and let hot points gets even hotter. If you have a correct flow from the front, you don't need those side inducts for CPUs (except for overclocking, or hard gaming when it's 45 C outdoor). moreover, those side inducts are bad for the airflow, they provide better CPU cooling, but the rest of the case gets warmer (like a disc burner for example) ++

You don't need to explain me again, i understood the first time, and you still don't. This "feature" suggestion will lead to a bug. -> adding a tag doesn't mean other programs are able to see it, thus they can update resources leaving the tag intact, letting "xpize_modded_with_your_idea" think its xpized when it is not (same issue if using the fileversion) -> comparing the resources themselves can be done only by predefined IDs, since xpize updates only the graphics and not the dialogs codes (for example), one can modify a dialog to make it use another resource ID, letting xpize updates a blind resource. This issue is present in the actual xpize, and in some other shell packs, your idea will just slow things without solving that (thus my "will also work"). (if you don't understand this point, just look around for sysdm.cpl customisations, for run box customisations and for logonui.exe customisation -- but all dialog boxes are concerned) i'm talking about a script i'm writing, it's not xpize-related, but resources-customizing-related. i just want more flexibility than what xpize do. (and use other resources than just graphics) ++

stop brainstorming and implement your idea in a real program of your choice (or find someone to do it). you should be aware that i've already wrote a bug report for it it was referring to "Log slipstreamed resources, slipstream the log, read back this log after windows setup." It does some other thing too, like actually slipstreaming the resources or extracting them back (it's in fact @T-13 before any hotfix can be applied, not after setup). i'm still debugging the "backup before hack" and "reloader" parts. ++

That isn't a good idea because there are some MCE hotfixes that should be applied after windows installation. It is a good idea(*1), and this error can be avoided with one of these three little tricks : - hotfixes can not be applied before T-13 (svcpack.inf), and the order things happen from this file is easy to control - md5 hashes or sha-1 hashes or whatever algo you want hashes embedded in the log - files' versions embedded in the log Like you said in the other thread, adding "tags" inside the files(*2), or extracting-then-compare resources will also work, but checking each binary is slower than reading a single log. (*2): this should be considered as a bad technique, since the updated resources themselves can act like "tags" (*1): i have a working batch that does it flawlessly, sorry for this egocentrism ++

You just can't under windows (switch to a linux-like filesystem !) ++ edit: but you can always try to specify an ARC path, like in boot.ini...

Un, dos, tres : Log slipstreamed resources, slipstream the log, read back this log after windows setup. ++

You just can't under windows (switch to a linux-like filesystem !) You just need to choose the right "UnattendedMode" value, there a tab with only this setting in nlite, click the "?" icon for the description (i'm unsure, it may be "default hide") ++

Mp3Tag is really weak compared to The GodFather : http://users.otenet.gr/~jtcliper/tgf/ Many formats are supported and the interface is marvellous (as feature-full than colorful). You can update files one by one, or by folders. You can make it guess tag values from file names, folder names, or from the online database you prefer. There's also a playlist editor and a player to listen music while your updating your tags (or to convert between formats). And a full-featured "library", a database of all your songs/tags to organize them the way you want without messing your folder structure (uh, oh, i forgot, it can also create this folder structure and move/copy/rename your files). And if that's not enough (i would hardly beleive that), you can make little scripts for any of the above functions (errm.. not the "library" of course). You just can't live without TGF's green parrot icon on your desktop... ++