The pain of Alpha
Moderator: Moderator Team
The pain of Alpha
Alright, I do know and realize this is an Alpha version, and although I am going to point out flaws that many of the developers probably already know, I just want to document my experience with the current SVN build as of 03/29.
Jaix's tutorial on how to set up the Trunk and RosBe were fantastic. You rule my brother.
So - I go and build my ISO, burn my CD, then go to my old DELL XPS R400 (PII / 512 MB RAM / 10 GB HD / 16 MB NVIDIA), and go for a new install.
First hurdle: Freeloader. The math is just not right. If there are existing partitions in the drive and I delete them in order to create a single partition, the available partition size ends up not working properly. I had to use my old Win98 cd to wipe out the partition and the MBR before being able to install.
Second Hurdle: My "Flushing Cache" screen never actually rebooted. It says it might take a couple of minutes, but 15 minutes exceeds a couple, so I had to manually reboot. Not a big problem, but one that the developer might wanna look into (I would, but I can't code [yet]).
Third Hurdle: Installation screen automatically set itself to 800x600. Now, that in itself won't be a problem for most people, since most modern monitors support that right of the bat. But I have the dell attached to an old black and white that only does 640x480, and the video went all floozy. I had to go unplug the spare monitor I use for my BeOS machine to plug into this one, just so I could complete the setup. I know its not a big deal, but frankly, all versions of windows easily overcome this by asking me if I want to keep the 800x600 setting (to which I just hit cancel and it reverts to 640x480 automatically). This is the type of idiot-proofing that the project needs to strive for, I think. Anyhow, Installation was remarkably similiar to windows.
Not sure that I like that. I mean, I know the aim is to have a Windows-clone, but I don't think it's right to copy the look-and-feel of it. I think that as long as the functionality is cloned, ReactOS should be free to evolve into its own look. Just my two cents, don't slam me for it.
Fourth Hurdle: After the install is complete (which required another manual restart), it comes up and the skeleton for plug and play automatically kicked in. Obviously, it didn't work as there are no drivers present, but it was exciting to watch detect the components, similiar to the way windows does it. It made me realize the DAUNTING task that lies ahead, which in turn, inspired to write my little review, which was obviously done on real hardware. Problem is, on the fifth item it was detecting, the computer simply turned itself off. Not a shut down, or a restart. Sucker just went off. Don't know why.
Fifth and last hurdle: After turning it on again, after the spontaneous turn-off, the Freeloader gave me some kind of CodePage error, and it wouldn't start the OS. I had to reinstall one more time, and it was basically a reprise of the experiences above.
In closing, I gotta say this project is extremely exciting, and I'm looking forward to becoming involved. There's a lot of crap I need to learn first. hahahah. In any event, I'll probably build a couple of scrap machines I got in my storage area, using an old Pc-Chip 598 MB, and I think I also got a Celeron 566 on an MSI MB. I'll try it on those as well and see how it fares.
Jaix's tutorial on how to set up the Trunk and RosBe were fantastic. You rule my brother.
So - I go and build my ISO, burn my CD, then go to my old DELL XPS R400 (PII / 512 MB RAM / 10 GB HD / 16 MB NVIDIA), and go for a new install.
First hurdle: Freeloader. The math is just not right. If there are existing partitions in the drive and I delete them in order to create a single partition, the available partition size ends up not working properly. I had to use my old Win98 cd to wipe out the partition and the MBR before being able to install.
Second Hurdle: My "Flushing Cache" screen never actually rebooted. It says it might take a couple of minutes, but 15 minutes exceeds a couple, so I had to manually reboot. Not a big problem, but one that the developer might wanna look into (I would, but I can't code [yet]).
Third Hurdle: Installation screen automatically set itself to 800x600. Now, that in itself won't be a problem for most people, since most modern monitors support that right of the bat. But I have the dell attached to an old black and white that only does 640x480, and the video went all floozy. I had to go unplug the spare monitor I use for my BeOS machine to plug into this one, just so I could complete the setup. I know its not a big deal, but frankly, all versions of windows easily overcome this by asking me if I want to keep the 800x600 setting (to which I just hit cancel and it reverts to 640x480 automatically). This is the type of idiot-proofing that the project needs to strive for, I think. Anyhow, Installation was remarkably similiar to windows.
Not sure that I like that. I mean, I know the aim is to have a Windows-clone, but I don't think it's right to copy the look-and-feel of it. I think that as long as the functionality is cloned, ReactOS should be free to evolve into its own look. Just my two cents, don't slam me for it.
Fourth Hurdle: After the install is complete (which required another manual restart), it comes up and the skeleton for plug and play automatically kicked in. Obviously, it didn't work as there are no drivers present, but it was exciting to watch detect the components, similiar to the way windows does it. It made me realize the DAUNTING task that lies ahead, which in turn, inspired to write my little review, which was obviously done on real hardware. Problem is, on the fifth item it was detecting, the computer simply turned itself off. Not a shut down, or a restart. Sucker just went off. Don't know why.
Fifth and last hurdle: After turning it on again, after the spontaneous turn-off, the Freeloader gave me some kind of CodePage error, and it wouldn't start the OS. I had to reinstall one more time, and it was basically a reprise of the experiences above.
In closing, I gotta say this project is extremely exciting, and I'm looking forward to becoming involved. There's a lot of crap I need to learn first. hahahah. In any event, I'll probably build a couple of scrap machines I got in my storage area, using an old Pc-Chip 598 MB, and I think I also got a Celeron 566 on an MSI MB. I'll try it on those as well and see how it fares.
Re: The pain of Alpha
This problem occur only on some hardware, connected probably with HDDs.vperez69 wrote:Second Hurdle: My "Flushing Cache" screen never actually rebooted. It says it might take a couple of minutes, but 15 minutes exceeds a couple, so I had to manually reboot. Not a big problem, but one that the developer might wanna look into (I would, but I can't code [yet]).
You can already choose the default screen-resolution in the first step setup (text mode), in the same screen where you can select your keyboard layout, etc.vperez69 wrote: Third Hurdle: Installation screen automatically set itself to 800x600. Now, that in itself won't be a problem for most people, since most modern monitors support that right of the bat. [...] all versions of windows easily overcome this by asking me if I want to keep the 800x600 setting [..]
It seems like that your registry got corrupted ...vperez69 wrote: Fifth and last hurdle: After turning it on again, after the spontaneous turn-off, the Freeloader gave me some kind of CodePage error, and it wouldn't start the OS.
Did you exit ReactOS with the "log-off" function?
Awesome
Frik85, my firend, you are absolutely right! I was able to change the resolution during the text-phase of the install. On the DELL its still a little faulty (like the spiontaneous shutdown). To answer your question, no I did not shut down with the button, the machine turned itself off spontanously during the stub-detection screens.
I put together a second test machine. Pentium II/233 on a QDI Legend MB with 128 MB RAM. Its even older hardware than the DELL has, and the install went beautiful. It was even able to read the CD-ROM this time, without a hitch, and the automatic restarts worked fine on this system. I even tried to install Acrobat Reader 5.05. The install gave me a couple of operation errors, and after the restart, registry got corrupted again.
I actually have to get going to work now - (I work for PeoplePC) but tonight, I'm gonna try installing it on a PCChip 598 LMR with an AMD K6-2/500 Mhz processor with 256 MB RAM on another 10 GB Drive, see how that rolls out. Its very exciting to see win32 software running on a non-windows box. Very exciting.
(/me runs away laughing like a giddy school boy)
I put together a second test machine. Pentium II/233 on a QDI Legend MB with 128 MB RAM. Its even older hardware than the DELL has, and the install went beautiful. It was even able to read the CD-ROM this time, without a hitch, and the automatic restarts worked fine on this system. I even tried to install Acrobat Reader 5.05. The install gave me a couple of operation errors, and after the restart, registry got corrupted again.
I actually have to get going to work now - (I work for PeoplePC) but tonight, I'm gonna try installing it on a PCChip 598 LMR with an AMD K6-2/500 Mhz processor with 256 MB RAM on another 10 GB Drive, see how that rolls out. Its very exciting to see win32 software running on a non-windows box. Very exciting.

(/me runs away laughing like a giddy school boy)
-
- Developer
- Posts: 508
- Joined: Mon Apr 25, 2005 12:46 pm
Wow, you seem to have a lot of computers 
I've had the problem with the messed up registry, too. It seems it happens, when ROs doesn't shut down correctly and the cache isn't flushed. It would be great if the cache was written to disk correctly when you get that BSOD, wich is in most cases the way ROS crashes. It says it is shutting down the system to prevent damage to the computer, but in fact this causes the damage to the filesystem

I've had the problem with the messed up registry, too. It seems it happens, when ROs doesn't shut down correctly and the cache isn't flushed. It would be great if the cache was written to disk correctly when you get that BSOD, wich is in most cases the way ROS crashes. It says it is shutting down the system to prevent damage to the computer, but in fact this causes the damage to the filesystem

NT does that, don't it?
I think NT does that, the moment you log on successfully, it backs up the current registry settings into Last Knwn Good Configuration, such that if the Current Configuration craps out, the next time you restart, you can press F8, use Select Last Known Good, and continue working.Switchboy wrote:Maybe there should be a backup of the registry wich is made every time ROS starts. So when the original regestry gets damnaged there is a more or less up-to-date backup. A function like this will prevent alot of reïnstalls because of registry damage.
Gosh I wish I could code C or Assembler. My books won't be here for another 2 weeks!
-
- Developer
- Posts: 508
- Joined: Mon Apr 25, 2005 12:46 pm
Re: NT does that, don't it?
Take some time, read a C tutorial, read a Windows programming tutorial and start codingvperez69 wrote:Gosh I wish I could code C or Assembler. My books won't be here for another 2 weeks!

Several month ago, I didn't know much about C. In fact I always hated the C syntax, I always prefered Pascal. But ROS aroused my ambition and just recently my first patch was included into ROS.
Now I feel ready for something bigger, than a 5 lines patch

I was under the impression windows runs the active settings in the registry in a striped set, holding the active contents in current configuration, configuration.xxx (with xxx being the mirror of the active copy), and then one paged in memory but in binary form. It makes sense to have it in a striped set where its able to continuously check itself for corruption and to rebuild itself on the fly when found to be corrupted.
*************************************
Go Huskers!
Go Huskers!
Who is online
Users browsing this forum: Bing [Bot], Yandex [Bot] and 3 guests