DFU for X86

[saprykin]

Is there anything like DFU for X86? It is a pain to deal with code repositories which include files with special characters in their names. On Alpha and IA64, DFU nicely handles these cases and allows to remove directories full of such files. However, I did not find DFU port for x86.

[arne_v]

I have a note saying that DFU will never be ported to x86-64.

But I cannot find the original source for that claim.

Anyway - for renaming files to valid ODS-2 names or similar then I would think any decent scripting language could do it.

Added in 1 minute 28 seconds:
I frequently have to rename jar files with lots of periods (version numbers) and I use the little Groovy script:

Code: Select all

curdir = new File(".")
for(f in curdir.listFiles()) {
    fnam = f.name
    if(fnam.endsWith(".jar")) {
        fnewnam = fnam.replace(".", "_").replace("_jar", ".jar")
        println("${fnam} -> ${fnewnam}")
        f.renameTo(new File(fnewnam))
    }
}

[sms]

Code: Select all

> [...] It is a pain to deal with code repositories which include files
> with special characters in their names. [...]

   That apparent complaint tells me approximately nothing about what you
wish to achieve, how you've been trying to achieve it, and what went
wrong when you did whatever you did to whatever you did it to.  Or how
you think DFU would help.

   None of that is a useful problem description.  It does not say what
you did.  It does not say what happened when you did it.  As usual,
showing actual actions (commands) with their actual results (error
messages, ...) can be more helpful than vague descriptions or
interpretations.

[craigberry]

Is there anything like DFU for X86? It is a pain to deal with code repositories which include files with special characters in their names. On Alpha and IA64, DFU nicely handles these cases and allows to remove directories full of such files. However, I did not find DFU port for x86.

DELETE/TREE works ok for me, though it's a lot slower than DFU's version.

[hein]

Arne wrote: "I have a note saying that DFU will never be ported to x86-64."

If that's true, that's madness! It is an essential tool.

Not porting early on is ok, one has to pick pones battles with limited resources.
Not ever porting it shows disrespect for OpenVMS users and system managers.

Just my opinion, Hein.

[reinhardtjh]

There was a discussion of this a year ago in April 2023 on C.O.V.

On 4/6/2023 12:57 AM, Jan-Erik Söderholm wrote:

We (and the source where I heard this) only uses DFU for non-update
tasks. There are many nice search and report features in DFU. And
at least my source has a strict ban against using DFU for anything
that updates disk data.

My source is a major VMS user in Sweden that is in the work of their
IA64 to x86-64 migration. They goes through all their 3-party tools
and DFU has been "red-flagged".

My source wrote:

"DFU is unfortunately flagged red. Its a freeware and the maintainer
is not interested in porting this to x86-64. We have asked VSI but
have not got any positive answer. They are probably fully occupied
with their own software, and there might be parts of DFU that VSI
might not want to take responsability for, like DEFRAGMENT, MODIFY,
UNDELETE and VERIFY".

I have no idea if they actually have been in contact with the DFU
maintainer, or if that is hearsay. Or if they have got any other
reply from VSI apart from a "positive" one.

Probably not a show-stopper for the x86-64 migration (I think they
definitely will migrate), but might need rewrite of a few routines.

So, I guess the question is if there is someone else that would like
to pick up the DFU sources and try a x86 build...

The problem is that the sources are held by Jur van der Burg and he is disinclined to port to x86 for whatever reason.

On 4/6/2023 6:48 AM, Carl Friedberg wrote:
I once wrote to Jurg, and he replied quickly with a suggestion for me,

I am not in a position to support this tool; but if I were, I would
rename it for x86 functions and only support read-only mode (perhaps
DFURO, although that sounds wonky). This is a great opportunity for
someone interested in the internals of the file system.

Carl

And the sources are not publicly available.

https://www.digiater.nl/dfu.html

License:

DFU is freeware. It may be freely downloaded and used. The sources are not available, and support is limited.

Maybe if someone can convince Mr van der Burg to let them (and only them?) see the sources for porting purposes? But I don't know who that would be.

[hein]

Indeed read-only would serve 99% of my needs.
Basic fragmentation report, search by name, search by fragments, search by size and such.

Hein.

[duncanmorris]

I would echo Hein's opinion.

We make extensive use of the various search options.
I would love to see the DELETE/DIR function retained, but would happily settle for DFURO.

[imiller]

DFU is a very useful utility for any VMS system and I would like to see it on x86.

Ideally

• the sources would go to VSI and DFU would be ported and officially supported by VSI

or if that can't be done,

• could it be ported by VSI and not officially supported

or

• open sourced so whoever has the time and expertise could work on it ( there may be copywrite issues preventing this )

[joukj]

I have a note saying that DFU will never be ported to x86-64.

But I cannot find the original source for that claim.

I'v seen someone claim that too, but I never saw an "official" stament form Jur (the digiater). Did anyone contact him directly?

Jouk