Drugwash

Well, I plan on releasing the series once I fix them properly once and for all. They were released long time ago here somewhere but the link must have died since and they've been buggy anyway. Wanted to make a CE series (Central European fonts) but fonts got mixed and not all elements are present and/or match with each-other. So it will take a while. In the mean time, would you be so kind as to play a while with this very rough alpha of the editor and tell me whatever you find buggy, incomplete or unintuitive? Any good ideas are welcome. It still got a long way to go before being actually useful for the purpose I started to build it but for now it can edit the values and save the outcome. Most of the menu items are not implemented. Please be careful what you change and where you save. Try not to overwrite existing files. It will only save an ini, no bitmaps or anything (yet). Argh, there's something missing in 1.0.0.6, icons won't show up in toolbar. Will upload a new version soon. Uploaded v1.0.0.7, should be OK.

Should've checked them myself, sorry.

Watercolor at 114 and Windows 7 at 141? Are you sure it's not a typo? Quite a big difference, both being superbar... Anyway, thank you for testing. I haven't yet reached the checking part - still reading and storing the changes. The API doesn't help much and neither does my lack of knowledge. Reading from an .ini file is tedious, the system opens and closes the file for each and every value being retrieved (as it's implemented in AHK) and that makes the application a bit slow. Also, the language variations in the Start Button section is something I have yet to take care of and still have no idea how to approach it.

Thanks! For now I'm using Tihiy's own comments on the UberSkin dedicated webpage, in the details box. Eventually I'll try to polish them a bit, but for now they should be enough. The variety of data types (string, bool, int, colorref) is giving me a hard time. I want it all to be intuitive and easy to use, which right now isn't. But I'll do my best.

The taskbar is essentially a standard Rebar control holding a few bands. Its size is always recalculated anytime one of the bands changes size. The threshold may depend on the Common Controls version installed on each system. I have started working on a theme editor but with all the possible options I'm having troubles finding an optimal GUI layout. It will take some time before getting to a usable public version. Here's a preview. (hopefully it shows - they got a weird system at CloudMe)

Just by a quick look at my Royale Mauve, I see the Start button is 90px tall. Each of the three usable images have a border of 2px vertically plus a 2px left border and a 1px right border. 2x2x3=12px total vertical border in one button image, making each individual image 26px tall. Your offending button, based on the same calculation, would be 40 px tall each image, as opposed to a value of 30 in skin.ini. There is probably a certain margin that - when overcome - triggers the taskbar height increase. I suppose the 2px (transparent) border is useful when taskbar is locked, while leaving a small room between button and taskbar background when it's unlocked. The value of 141px for the superbar theme gives an individual image height of 43px (not considering the border) vs the value of 40 in skin.ini. That is probably an accepted margin. Same goes for your new button, where each image would be a 32px (not considering the border). You may try to play with different image sizes (height+2+2)*3px and see where it breaks the limit and makes the taskbar grow.

There's a value called 'TaskBarSize' in the 'Global' section. May be a leftover from older skins, dunno, but it can be checked against the full height of the button bitmap divided by the value of 'Count' in the 'Button' section. If you find any other similar glitches, please let us know. If I ever find the time to build that tool, it should have all possible checks built-in. My Royale set has seven color variations (blue, cyan, green, mauve, orange, red, yellow) but the fonts are not wisely chosen and some elements still need tweaking. Someday, maybe...

Could other skin elements contribute as well to such bug? I'm thinking of writing (sometime, maybe) a tool that checks for such mismatched sizes between .ini declarations and actual bitmaps. Maybe even with a preview screen just like the one in the display control panel. Of course, that's just wishful thinking for now. I should also review and fix my Royale color variations set of themes I've done years ago. Maybe at some point they'll make it into your package. But for now I have big hardware and software issues that need to be fixed so it will take some time.

Install VC6 and set it as default debugger. In the error box you should get an additional button called 'Debug' which - when clicked - will open VC6 in debug mode and offer details on the error. To set VC6 as default debugger, add the following to C:\WINDOWS\WIN.INI: [AeDebug]Auto=0Debugger="C:\Program Files\Microsoft Visual Studio\Common\MSDev98\Bin\msdev.exe" -p %ld -e %ldIf for some reason you already have another debugger set, just disable it by placing a semicolon in front of it. For example: [AeDebug]Auto=0;Debugger="C:\Utils\FineSSE29.exe" -p %u -e %u -vDebugger="C:\Program Files\Microsoft Visual Studio\Common\MSDev98\Bin\msdev.exe" -p %ld -e %ld

Of course not, rest assured. Thank you, just wanted to be on the safe side. One never knows when their actions (read: words) may be misinterpreted. NO NO continue please continue. I'm guessing that this program isn't normally supported with KernelEx... Interesting. If there are a lot more programs like this then UPX should get its own thread . Indeed, trying to run that application straight after installing it would result in crash, even on a KernelEx system. My 98SE machine has survived uncounted crashes since Sept.2006 and is still going strong. But nowadays it seems we have to delve deeper. I've posted the essentials above. Waiting for anybody interested to confirm my findings or post adjacent issues. Ideally we should find out what exactly UPX does to prevent crashes. At first I thought it might me OS version/Subsystem version/Linker version in the exe/dll to make trouble, but then I read about .text sections above and I'm not sure now what the correct answer would be.

Going to 7-zip's SourceForge project page and the discussions, one would see that 7-zip has been in alpha/beta stage for years already. Sadly, 7.31 codebase is the last to work in a Win9x environment, according to the discussion topics. After tinkering for a a while, I've been able to build the 7-zip plug-in for Total Commander from the old 0.7.2.1 sources, right before it got switched to Unicode only. I used the 7-zip 9.20 sources for building it and had to emulate a few newer API to get it to compile under VC6. The only compatibility problem I suspect is the LZMA2 compression method, which is not present in older 4.x 7-zip versions (reason why I compiled the new plug-in version, as the old 0.5.5 was also based on 7-zip 4.65). Therefore using a newer - and perfectly 9x-compatible standalone 7-zip setup - would be preferrable. And a software being in alpha/beta stage isn't necessarily bad or buggy - it may just contain incomplete features.

Apologies for breaking the VLC trend here, but a few days ago, using the same UPX trick discovered by loblo, I managed to use the CloudMe application (v1.8.0), uploading files from my Win98SE machine. Synchronisation also seems to work correctly. There are still a few quirks, probably unfixable: - the systray menu won't show up so the application must be killed when finished working with it - file/folder navigation is buggy and requires copy/paste for getting to the desired file/folder It is mandatory to upgrade the root certificates in order for the application to connect to the cloud server. The certificates can be obtained from uSP3 or downloaded directly from M$ (validation required). The CloudMe application is built around Qt5 and uses the VC10 and VC11 runtimes as well as a recent SSL library. All of the DLLs and EXEs must be packed with UPX, including those in subfolders. I used no commandline parameters when packing. That's about all. If you want a free 3GB cloud space, get it, fix it, run it! (I sincerely hope the above won't be taken as advertising, since I got no truck whatsoever with that cloud storage company or whatever)

I have a board which is now completely dead (internal short on ATX12V socket which I can't pinpoint). Previously, on two occasions, I had replaced a set of capacitors on it, 14 in all. Not that badly leaking as the ones in your pictures but bad enough to cause the board malfunction. It worked just fine until last week. Therefore a good electronics technician could attempt to revive the board if so desired. The hard drive, however, I believe to be dead. The grinding noise and the clicking are definite signs of internal damage. I've had a couple drives go clicking, in time, which brought me near a heart attack each time as I had no backup of the current work. I lost everything then. So you may wanna prepare your friend for the bad news. Chances are slim that a specialized technician would recover the data, but if it's that important then you/your friend may try to go to a local repairs/recovery store. Attempting to recover the data yourself in this situation by repeatedly connecting and powering up the drive might destroy it even further, taking the recovery chances down rapidly, so I'd strongly advise against it. EDIT: Well, replies have added up since I started typing. If the fans are running, that means the +12V is present. However the PSU could lack any of the other important voltages, such as the +5V or the +3.3V. Or they could all be lower than minimum accepted, due to... dry/leaking/gonflated capacitors inside the PSU. It's worth a check. A multimeter would be enough to prove the PSU voltages. Find the ATX pinout at http://pinouts.ru . Make sure there is the +5V SB (the so-called 'power good' voltage) - no board would start up without it. Another thing drawing attention is the apparently non-functional optical drive. I'd disconnect it completely (IDE cable and power cable), at least to eliminate the possibility of a faulty drive taking down the PSU.

Dunno about others, but in my eyes you already are a legend. Thank you for all your efforts through the years! Unfortunately my hardware is going down along with my brain, reason why I haven't been active in a long time. I'm sorry I promised a lot of things and couldn't fulfill them or even completely forgot about those things. I wish I could've helped you (and others) more. Personally I'm happy with my eight-years old Win98SE installation and will always consider that as my main system. It's also the only one still up and running with no major repairs inspite of its 24/7 operation. For a brief period of time I've had Millenium installed, years ago and the lack of true DOS apart I was rather pleased with it; maybe it was just luck, dunno. Right now I don't have a ME setup and no spare hardware, since my most powerful machine died (mobo fried, short on ATX 12V socket) and two other machines have bad sectors on their HDDs. Therefore I won't be much of a help, unless you need a specific tool built or something. Thank you again and good luck with all your projects!

There's been a long discussion here about that. NUSB 3.6 introduced the sysdm.cpl (Device Manager control panel) from Windows ME in order to take advantage of unattended detection of USB devices. That brought the issue of misleading version information along with an icon bug and unneeded System Restore options. Some of us tried to edit those resources and fix the bugs and I think there's a usable version floating around, maybe in the uSP3 or attached in the respective discussion thread. I do have my own modified version here but I'm not sure it's bug free - it's been too long since I worked on it and just forgot the details.

They probably get the drive handle the wrong way. In Win9x it should be: CreateFile("\\.\vwin32", 0, 0, NULL, 0, FILE_FLAG_DELETE_ON_CLOSE, NULL)while the NT family uses: CreateFile("\\.\C:", 0, FILE_SHARE_READ, NULL, OPEN_ALWAYS, 0, NULL) // C: is the actual drive letterThere may be additional issues with DeviceIoControl() parameters though. For an experienced programmer, this should be a walk in the park. They're either all amateurs or just afraid of the big bad wolf bullying anyone that dares to support Win9x. I'm not a programmer but I'm still building my own software on and for Win9x. Open-source. Anyway, media file conversion is always done through external encoders/decoders/filters so if the tool doesn't come with its own, it has to find and use the system ones, if existing. AnyBurn comes with Lame encoder, libFLAC, libvorbis and MonkeyAudio, which should be usable with any other tool, provided it can handle its API.

I have a feeling it's not the SQLite library driver that breaks those functions but some missing Unicode API or ANSI/Unicode mismatch in names/paths. That wrapper wouldn't be of much use, as it's meant to be only a testing tool for the library. Much like the command line tools that come with the library in the official page, only with a graphical interface for ease of use. There are a lot of internal functions in the SQLite library and there's always changes and additions in newer official versions, which make things harder. I doubt jumper would know AHK but in any case I can provide the sources since it's all open-source with me.

Unfortunately, no. As always, I've been sidetracked by something else. Which reminds me I once promised you to check and improve an AHK script related to MPClassic - again, it took me so long to make that player work somehow, that I completely forgot why I needed it. My apologies! Dunno what's going on but I'm slowly losing it - too much on my mind and can't focus. Will try to catch up sometimes soon but please do remind me about this by mail, after a while.

Maybe you're missing the bitmap pertaining to the Systray area, in the skin package - have you tried a different skin? Or open in Notepad the skin.ini corresponding to the current skin and check if the file mentioned by the Bitmap value in the TaskBarTray section exists in the current folder; it may be missing or its name could be wrong in the ini. You may fix the name or borrow the file from another skin (or build it yourself in an image editor - 101x28px, 24bpp bitmap in Royale/Windows XP/ZuneXP themes). As for KernelEx and Revolutions Pack, I've been running both on my main 98SE machine for years, so yeah, they do get along fine.

Which OS version you've installed it on (98 Gold/98SE/ME)? Is it a true install or in a virtual machine? Have you installed the required updates first, as mentioned in post #2 of this thread? Are there any other system enhancements installed (unofficial SP3/AutoPatcher/98SE2ME/etc)? Is there any application running that could mess with that taskbar area (TClock, for example)?

Yes, I've noticed certain settings are not ported over when migrating to ini, especially the WordWrap option. I may look into that sometime but considering my very bad memory, don't hold your breath. I haven't yet tested the UTF-8 capability. Since it's been a patch from someone esle - as noted on Alex's blog - it may well be incomplete. I'm so angry on myself for not having learned C/C++ in due time, otherwise it would've been so easy to fix/improve a lot of code offered freely by the good people. But I'll try, maybe even contact Alex for some tips. In the mean time I've managed to build metapad in VC6 but it needs extensive testing to make sure it works correctly. The new icon is a major pain because VC6 won't recognize it so I had to rename the original, delete a few images, save it, compile with the modified icon and then replace it in ResHacker (newer version, because the older one would bail out). The manifest needs to be added afterwards too in ResHacker. (edited for typo)

Thanks for the heads up, Charles, didn't know there was an update although it's a couple years old. Source code is available too and I see an attempt at making it Unicode-aware. Unfortunately I couldn't modify it to create a Unicode RichEdit control although I have built a similar application myself in AutoHotkey which can even load a RichEdit50W control if available in 9x. Anyway, I've been using Metapad as main editor (renamed as Notepad) for as long as I know, even in XP. Hopefully the new UTF-8 capability will alow correct file conversion, which I was missing.

There used to be a utility called Nopey that could flush the HDD cache using the command nopey sync X: (where X is the drive letter).

Shamefully it's been a long time since I last installed uSP3. Can't focus on anything anymore and lack of spare hardware for testing is also a problem. However I hope the starter is working correctly. Is Word 97 the same as Word that comes with MS Office 97? I'm just thinking of a possible confusion in case one also installs the full Office 97 suite. A shortcut anywhere on desktop or Start menu > Programs would be welcome, just to let users know it's installed; they can always move it to any desired location afterwards. I also sort my shortcuts in custom category folders under the Start menu. Used to have a completely clean desktop with no icons other than My Computer with a blank icon and name, but I "lost it" years ago. Remember Mel Gibson and Danny Glover in the "Lethal weapon" series: "I'm too old for this sxxt".

Due to my very bad memory I can't tell exactly which version it was or what the problem was, but I do recall very long time ago I had some NTFS driver installed on a 9x machine and at some point a major problem popped up. Since then I never trusted any such driver. That is why nowadays my current 98SE machine doesn't use either of the available NTFS for 98 drivers although my HDD does have such a partition along with a ext2 (or ext3?) partition available read-only through a Total Commander plug-in. I'm saying this just to back up previous posters that claim certain issues with the NTFS drivers under Win9x. One idea just popped to mind: is it possible that the files don't "stick" due to a delay write (drive cache) setting enabled for the respective (USB) drive? It's worth a check but I couldn't tell where it is, if available in Win9x. I've seen that in XP though so there may be some way to control this in 9x too.