Comments for Coding Clarity

Comment on Hacking JCIFS to support setting share permissions by Ronny Bremer

Ronny Bremer — Mon, 02 Apr 2012 07:38:24 +0000

Thank you Chris. I will take a look at it and let you know.

Comment on Hacking JCIFS to support setting share permissions by chris

chris — Fri, 30 Mar 2012 17:47:30 +0000

You are correct, the original patch does seem to be missing. The original patch had some issues with it. It had some formatting/whitespace changes in it making it difficult to apply without causing problems. I managed to take the pieces I needed.

I will attach the original patch to my post so it is available to you.

Comment on Hacking JCIFS to support setting share permissions by Ronny Bremer

Ronny Bremer — Fri, 30 Mar 2012 11:40:54 +0000

Chris, it is interesting to see that you have been stumbling across almost the same issue I have at hand for a customer project. In my case, however, I need to set the permissions on a single directory, so I was about to follow up on your approach with the path you mentioned (setSecurity on smbFile). However, this patch has been removed from the nabble.com website and I can’t seem to find it anywhere. Do you have a copy of it you could send me? I would love to enhance your patch to also include setSecurity().

Comment on Hacking JCIFS to support setting share permissions by Michael B Allen

Michael B Allen — Sun, 18 Mar 2012 06:05:16 +0000

Nice job. This is how people should contribute to JCIFS. You studied the right specs, wrote a little documentation about your mod and posted the whole package in a way that someone can pull down and use it without too much fuss (including me if I decide to add the NetShareSetInfo call to the stock JCIFS).

Although your statement about the midlc IDL being different from midl IDL is not true at all. For example if you study at the SamrOpenDomain call from JCIFS’s samr.idl and the same IDL in the [MS-SAMR] document you site, you should find that they are actually the same IDL. It’s just the types that *look* different because Microsoft uses a lot of extraneous typedefs whereas JCIFS normalized a lot of common types. For example, JCIFS has a “policy_handle” type that is used all over the place in different RPC interfaces whereas Microsoft uses a different typedef in each interface for policy handles (SAMPR_HANDLE* in the case of samr). MS IDL is very much based on C (because with attributes removed it actually is C and MS embeds this stuff in their C files) so no doubt the typedefs would be confusing for someone who hasn’t used C much. To really see how the types are actually the same, you would have to trace back the typedefs to their origins. But usually you can tell just by the name or by looking at what JCIFS uses for some parameter that is the same in a different call.

Comment on Subversion is not Dead by Working with Git against Subversion Repositories - Coding Clarity

Working with Git against Subversion Repositories - Coding Clarity — Fri, 15 Apr 2011 23:12:10 +0000

[...] my previous post on why Subversion is not Dead, I really do love Distributed Version Control Systems. I’ve wanted for some time to move our [...]

Comment on Subversion is not Dead by chris

chris — Tue, 22 Mar 2011 13:12:10 +0000

@Arne Yes, there are ways to bundle GPL software with a commercial offering. The issue is more with enterprise customers. Many big firms look at GPL as a liability and have blanket policies that they will not accept any GPL software. Their reasons may be unjustified but it does not change the fact that they will not purchase your software if you use GPL. You have no choice but to comply if you want their business.

This becomes even more of a problem for the OEM market. We have OEM partners who resell our software. The list of supported licenses for this is even more restrictive. Some of our OEM partners will not accept LGPL either. Their reasons are probably not justified but their legal departments refuse to budge on this.

GPL is bad for political reasons more so than for technical ones.

Comment on Subversion is not Dead by Arne Babenhauserheide

Arne Babenhauserheide — Tue, 22 Mar 2011 12:50:41 +0000

You can bundle GPL software in a commercial product, you just have to provide the sources of the free software program (not of your unfree software which you just distribute along with it) – which is dead simple in the case of Mercurial

You just can’t link against it with unfree software, but you *can* use the commandline API (both the main stable interface of Mercurial and of Git).

Besides: As far as I know, Bazaar is GPLv3.

And Bitkeeper is not GPL. Bitkeeper not being free (and actively restricting the kernel devs) was the reason, git and Mercurial were created.

Comment on Subversion is not Dead by chris

chris — Tue, 22 Mar 2011 12:28:38 +0000

@foobar I think you are correct on bitkeeper. I edited your initial comment for you.

Comment on Subversion is not Dead by foobar

foobar — Tue, 22 Mar 2011 12:26:17 +0000

I mean “thank”, not “think”. My bad.

Comment on Subversion is not Dead by foobar

foobar — Tue, 22 Mar 2011 12:25:38 +0000

I think we can thank bitkeeper for all the DVCS being GPL.