Re: netbsd-10, merging device-streams (xstreamtodev, rdbinfo)

Roc Vallès <[email protected]> Tue, 14 Mar 2023 21:20:56 +0100
Newsgroups gmane.os.netbsd.ports.amiga
Message-ID <CAPq5wtaUg2TgF22PVB=a4yzQ=bqpivyehbhWhiZoXATpkP_Wjw@mail.gmail.com>
Hi David,

Much appreciated.

Roc

On Tue, 14 Mar 2023 at 21:04, David Brownlee <[email protected]> wrote:
>
> Hi Roc,
>
> I'm dealing with a sudden burst of $dayjob related stuff, but I should ge=
t time to look at this in a day or so - thanks again for the work on this
>
> David
>
> On Mon, 13 Mar 2023 at 05:04, Roc Vall=C3=A8s <[email protected]> wrote=
:
>>
>> Hi Carlos,
>>
>> Yes, this updates xstreamtodev. This "device-streams" name comes from
>> what the uuencoded source code tarball in netbsd's CVS is called. Its
>> makefile builds: streamtodev, devtostream, xstreamtodev, xstreamtodev
>> and rdbinfo.
>>
>> My changes make it so that it builds with the most common AmigaOS
>> crossdev toolchain today (it used SAS/C back in the day). I also
>> cleaned it up and changed everything necessary for large disk support
>> when available (TD64 and NSD) and otherwise to be safe against 4GB+
>> access.
>>
>> This work is in the github I linked, including the full history of the
>> changes I made. The license is of course kept to what the original
>> author had.
>>
>> A developer with commit access would have to uuencode and update the
>> rdbinfo and xstreamtodev binaries. Using mine outright (github release
>> binaries or aminet, they are the same) is an option.
>>
>> As the version we currently have in the netbsd CVS repo will happily
>> clobber your data if asked to write past the 4GB mark, it would be
>> particularly nice for this to be included in the netbsd-10 release.
>>
>> Regards,
>> Roc
>>
>>
>> On Sun, 12 Mar 2023 at 21:03, Carlos Mil=C3=A1n Figueredo
>> <[email protected]> wrote:
>> >
>> > From: [email protected] <[email protected]> On Beh=
alf Of Roc Vall=C3=A8s
>> > Sent: jueves, 9 de marzo de 2023 06:05
>> >
>> > > A while back, I updated device-streams, motivated by awareness it
>> > > clobbered data if writes were done above 32bit offsets, which is ver=
y
>> > > easy to do as the program targets partitions, they can exist above 4=
GB
>> > > boundary and large disks are common today.
>> >
>> > Sorry for not being able to answer your question, but I had another on=
e: would this tool replace xstreamtodev?
>> >
>> > Regards,
>> > Carlos
>> >
>> > Carlos Mil=C3=A1n Figueredo | HispaMSX System Operator | http://www.hi=
spamsx.org | telnet://bbs.hispamsx.org | https://calnus.com