As of Revision 17175 Reactos 0.2.7 RC-2 has been built and is donwloadable at sf.net
Since RC1 a couple of blocking bugs have been solved.
We welcome you to test this pre-release on your hardware. And report problems.
Most interesting to us is the behavour of your keyboard. Lately there occoured some problems in detecting always a keyboard on any hardware or to correctly detect release codes of different keys. Please report problems in this area with details to your hardware.
RC2 went out with the fix for bug 659 still unmerged. I have reopened the bug, and once again I am asking for someone to merge it.
On 8/7/05, Robert Köpferl rob@koepferl.de wrote:
As of Revision 17175 Reactos 0.2.7 RC-2 has been built and is donwloadable at sf.net
Since RC1 a couple of blocking bugs have been solved.
We welcome you to test this pre-release on your hardware. And report problems.
Most interesting to us is the behavour of your keyboard. Lately there occoured some problems in detecting always a keyboard on any hardware or to correctly detect release codes of different keys. Please report problems in this area with details to your hardware. _______________________________________________ Ros-dev mailing list Ros-dev@reactos.com http://reactos.com:8080/mailman/listinfo/ros-dev
It was easier to get commit access that it was to convince someone to do this for me. ;0) I applied the fix in 17216.
On 8/8/05, WaxDragon waxdragon@gmail.com wrote:
RC2 went out with the fix for bug 659 still unmerged. I have reopened the bug, and once again I am asking for someone to merge it.
On 8/7/05, Robert Köpferl rob@koepferl.de wrote:
As of Revision 17175 Reactos 0.2.7 RC-2 has been built and is donwloadable at sf.net
Since RC1 a couple of blocking bugs have been solved.
We welcome you to test this pre-release on your hardware. And report problems.
Most interesting to us is the behavour of your keyboard. Lately there occoured some problems in detecting always a keyboard on any hardware or to correctly detect release codes of different keys. Please report problems in this area with details to your hardware. _______________________________________________ Ros-dev mailing list Ros-dev@reactos.com http://reactos.com:8080/mailman/listinfo/ros-dev
-- #irc.freenode.net #reactos 01:03PM <filip2307> i don't know about any bug 01:04PM <filip2307> none exist 01:04PM <filip2307> ReactOS is prefect
WaxDragon schreef:
RC2 went out with the fix for bug 659 still unmerged. I have reopened the bug, and once again I am asking for someone to merge it.
*RC2 installation cdrom requires 52MB of RAM to succesfully copy files to harddisk in VMware. This looks to me as a regression, as 0.25/0.26 was able to do this with 20MB of XMS. If using a lower amount of RAM than 52, the copy process hangs at a certain percentage. Please confirm, as ROS requirements are stated as 'Pentium, 32MB RAM' (IIRC)
*With above point (incomplete installation), FreeLDR bootsector gets installed before file copy finished (or attempts to finish). FreeLDR itself isn't installed yet it seems (until finishing ROS setup). When booting this bootsector, the 'press a key to reboot' gets pressed by me, but no reboot happening at all.
*How much harddisk space is needed, and is this checked by the installer? User can define after all to create very small partitions.
*How do I install FreeLDR to both harddisk and floppy (as backup boot method) at end of installation.
*If I recall correctly, a FreeLDR floppy cannot be converted into a disk image and be booted by Memdisk (syslinux.zytor.org) as that's a realmode ramdisk 'program'/'loader' and FreeLDR is going into protected mode before reading freeldr.ini on the same disk(-image)
*Can ReactOS only be installed by booting from the installation cdrom? reactos.exe reports to be a win32 program, while I access the cdrom by booting a freedos cdrom with downloaded cdromdriver. Any DOS binary (like 'winnt.exe') available?
*Can't the ReactOS SETUP cache/buffer the to-be-copied-ReactOS-files into memory as soon as startup starts? That would reduce the actual copy/transfer times as by the time the installer starts to copy, only diskwrites are needed, no more reads from cdrom (which are done while user answers the installer's questions about hardware and disk layout).
*At SETUP, when you select REPAIR, ESC key brings you back a screen. When you select L (for License), ENTER brings you back. Kind of 'counter-intuitive'
*Disk partition speaks about multiple extended partitions on a disk, which is impossible. Logical Partitions are meant?
*At the partitions screen, pressing ENTER to install ROS on the unpartitioned entire disk creates a partition of full size. Nothing wrong with this, except that it's too early: at the next screen, when pressing ESC to go back, you notice the fullsize partition being created already.
*At an installed ReactOS, the 2nd stage installer asks for a computer name, but does not accept ANY numeric character, only alphabetical. a computername ROS027TST isn't possible for example, but ROSTEST is.
*A note for the VMware screen: for the VMware's "install VMware tools" command to work (which mounts the ISO containing VMware drivers) the virtual machine has to be configured as 'Windows' (not 3.xx ofcourse). If not, user will have to manually mount the ISO file before the virtual machine will see it.
Out of feedback for now, don't let any comments stop you from releasing 0.27 (except the excessive RAM requirements at setup copying procedure)
regards,
Bernd Blaauw
Bernd Blaauw schreef:
WaxDragon schreef:
RC2 went out with the fix for bug 659 still unmerged. I have reopened the bug, and once again I am asking for someone to merge it.
additional tiny bug: *ReactOS SETUP formats the target partition, but still a bootsect.old is created upon installing FreeLDR. Totally unnecessary.
Bernd
Bernd Blaauw wrote:
*RC2 installation cdrom requires 52MB of RAM to succesfully copy files to harddisk in VMware. This looks to me as a regression, as 0.25/0.26 was able to do this with 20MB of XMS. If using a lower amount of RAM than 52, the copy process hangs at a certain percentage. Please confirm, as ROS requirements are stated as 'Pentium, 32MB RAM' (IIRC)
This is indeed a blocker. The release version should not require anywhere near that amount of RAM. Do we have a leak? Are our flushing algorithms broken? I thought that the installer would notice when memory is running out, and clean up the cache.
Best regards, Alex Ionescu