General Testing for 0.4.2

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

71248 works
71502 fails

I will work towards narrowing it more tomorrow.

C1

decodp
Posts: 7
Joined: Sun Apr 24, 2016 2:42 pm

Re: General Testing for 0.4.2

Post by decodp »

cirrus one wrote:As 41 is ok if you tell me which incremental build number
in the 7xxx format corresponds to 41 I will test ISO files
starting there in increments of say 25 until I find it then in
increments of 1 until it is narrowed downed.

Is there a way to access source code?

C1
Source code is in http://svn.reactos.org/svn/reactos

I hope it helps you

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I have found it. First:

42 is broken on my PC.

I get the following error:

Error reading NLS file
Multi(0)
Disk(0)
Cdrom(111)\
Reactos\system32\l_intl.nls

The PC is an ASUS K53E I5 legacy BIOS with a built in DVD reader.

41 boots just fine.

==============

The last build that works properly is 71365.

Build 71366 and all later builds HAVE THE PROBLEM on my PC.

I tried to look at the comments in your repository for that build
but it will not let me go back that far.

C1

User avatar
Black_Fox
Posts: 1584
Joined: Fri Feb 15, 2008 9:44 pm
Location: Czechia

Re: General Testing for 0.4.2

Post by Black_Fox »


User avatar
DougLyons
Posts: 21
Joined: Sat Oct 25, 2014 9:10 pm
Location: Pensacola, Florida, USA

Re: General Testing for 0.4.2

Post by DougLyons »

You can go here http://code.reactos.org/ and then at the top right search for your failed revision of 71366 to see what changed.

There was only one change in pcmem.c as follows:
Old: #define MAX_BIOS_DESCRIPTORS 40
New: #define MAX_BIOS_DESCRIPTORS 80

Also on this page is a link to the issue that it involved here http://jira.reactos.org/browse/CORE-11247.
If you follow it you can see that this fixed a users problem with booting.

You should log this problem in JIRA here https://jira.reactos.org/.
Then the developers can have a look at it and hopefully create a fix for this.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I have logged the problem.

Is there a way to track it.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I dug a very old PC out of my old stuff basket.
It is a DELL INSPIRON 9200 and it runs any
REACTOS without the problems.

The parameter that was changed was #defined.

If you can comment that out and give me a
special DEBUG version for me only where
I can set this parameter from the keyboard
or possibly a hex editor I could modify the
ISO I can determine for you how high I can
go on this ASUS pc without encountering
problems.

C1

User avatar
DougLyons
Posts: 21
Joined: Sat Oct 25, 2014 9:10 pm
Location: Pensacola, Florida, USA

Re: General Testing for 0.4.2

Post by DougLyons »

You can track this by going to your particular entry in JIRA which is https://jira.reactos.org/browse/CORE-11811.
Also you can just open up the JIRA main page here https://jira.reactos.org/secure/Dashboard.jspa and at the top right search enter you issue number of 11811.

To try and test this on your system you can try building ReactOS on your system by going to the wiki site here https://www.reactos.org/wiki.
Look at the "Building" section and you can see how to build your own copy of ReactOS that you can modify in any way you want.
Last edited by DougLyons on Mon Aug 08, 2016 5:39 am, edited 2 times in total.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I will look at that.

What needs to be done is to take the current build's ISO,

Take the current source revision, modify the value back
to 40. and re-build. Make sure the number of bytes in
the ISO is the same.

Either keep doing this, or find the suspect value with
a binary file compare and keep modifying it with a
hex editor.

How hard will it be for me to do the builds?

Quicker for you guys to do it one time down to 40,
let your auto system rebuild it, then restore it
right back to 80 and auto rebuild again.

And then I would do the hex editor procedure.

I will look at doing the builds myself.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

In https://jira.reactos.org/secure/Dashboard.jspa
there is no "building" section I can find.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I am going to download both the 65 and 66 ISO files.

Assuming there is only one byte value changed, the
total bytes should be unchanged.

I will hack at them with a binary file compare and a
hex editor, attempting to find a max value that still
works.

Then you guys can decide.

C1

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

I have no idea how to forward this
message to the guys on JIRA.

Please help.

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

Well that didn't quite work.

The byte counts between livecd65 iso and livecd66 iso are identical
but there are several hundred differences.
'
If it is 40 and 80 decimal to compare that is 28 and 50 hex.
There are 4 possible byte positions that match that.

Same with the bootcd iso files.

because there are only 4 differences I may try gambeling.

User avatar
DougLyons
Posts: 21
Joined: Sat Oct 25, 2014 9:10 pm
Location: Pensacola, Florida, USA

Re: General Testing for 0.4.2

Post by DougLyons »

Look at the branch for 0.4.2 here http://code.reactos.org/browse/reactos/ ... .c?hb=true.
Do a search on "MAX_BIOS_DESCRIPTORS" and you will see that this is used to allocate arrays so changing single bytes does not work well for these instances.

Also, we have somewhat gotten away from the original subject and maybe this thread should be moved to its own entry.
There is a chance that you can just replace the "loader" directory of the ISO with the old files and this will work fine, but this is just a guess.
My Build System is broken at the moment or I would give a try at creating a couple of files for you to compare.

Sorry for the previous wiki link mistake. Please review above where it has been corrected now as follows:
https://www.reactos.org/wiki

cirrus one
Posts: 68
Joined: Sat Jul 30, 2016 7:32 am

Re: General Testing for 0.4.2

Post by cirrus one »

We have NOT gotten away from the subject of this thread.
I am testing 42 and it does not boot on my ASUS machine.

Testing 42 is the thread.

Post Reply

Who is online

Users browsing this forum: Bing [Bot], Google [Bot], Semrush [Bot] and 3 guests