Coda File System

Re: pkgsrc patches

From: Greg Troxel <>
Date: Tue, 27 Jan 2015 16:11:38 -0500
Jan Harkes <> writes:

> Yes someone is here.

Glad to hear it.

> Haven't checked yet if the patch is applied to CVS, but I have a couple
> of emails with Coda related patches that were sent to codalist and me
> privately tagged to be applied.
> I figured I'd apply them after the CVS -> Git conversion was complete
> and then release a new Coda version with the state of things. I've
> scripted the whole conversion process with reposurgeon and stopped
> trying to 'perfect' the conversion mostly because I seemed to be getting
> stuck on reposurgeon related bugs.

If that's soon, that's fine.  I see getting buffixes applied and patch
release out as more important than converting.  I'm all for switching to
git, but it's not coda's biggest problem.

> The only remaining step is to validate that the source of the released
> versions of Coda actually match the tagged commits. I tried to automate
> the check, but due to subtle differences because of CVS/RCS tag
> expansions and such it looks like it might have to be done manually by
> unpacking a released tar and then diffing against a checked out release
> and reading the diffs to see if any of the differences are significant.

It sounds encouraging that this is the level of issue you are down to

>> +#if defined(__NetBSD__) && __NetBSD_Version__ >= 499002400   /* 4.99.24 */
>> +	if (error < 0)
>> +	    error = mount("coda", venusRoot, 0, (void *)kernDevice, 256);
>> +	if (error < 0)
>> +	    error = mount("cfs", venusRoot, 0, (void *)kernDevice, 256);
>> +#else
> I don't like the inline #ifdef style, but I can see that any other
> solution would just push the ifdef into a worse place. btw. what does
> the 256 stand for? shouldn't it include a header and use the symbolic
> name?

A fair comment.  Basically, mount's signature is

     mount(const char *type, const char *dir, int flags, void *data,
         size_t data_len);

So it's the size of the kernDevice object.  Probably "sizeof(*
kernDevice)" is apppropriate.  I can check that soon.

Also, there are solaris patches just arrived in pkgsrc.  I am unclear on
whether that means someone has the kernel code working, or if it's just
part of a "make all packages build on $OS" rampage.  But they look like
cleanups that ought to be applied anyway.  One issue is being careful to
use the stdint types consistently.

With the change to git, will there be a bugtracker?  It seems we don't
have one now, or at least I'm unaware.

Received on 2015-01-27 16:11:46