Sir Woody Hackswell wrote:
>On Tue, 25 Mar 2003, Garrett Rooney wrote:
>
>
>
>>Sir Woody Hackswell wrote:
>>
>>
>>
>>>+ user = (char *) malloc(16 * sizeof(char));
>>>+
>>>
>>>
>>>
>>don't do that.
>>
>>we use memory pools, not malloc, unless the malloc is necessary to
>>interoperate with another library (neon or berkeley db come to mind).
>>also, what's with the hard coded 16? and is user even freed anywhere?
>>
>>
>
>It was a quick hack. Sorry. I'm just starting out with svn, and this patch
>is more proof-of-concept than a polished version. I threw it together in an
>hour or so. :/
>
>Should I work on a more "proper" solution/patch? Where do I read up on
>memory pools, or what one or two functions can I look at for examples?
>
well, a more "proper" patch would at least be a good intro to the svn
source, so sure, why not ;-)
for info on how to use memory pools, there's a good section on it in the
svn book... check out http://svnbook.red-bean.com/html-chunk/ch07s05.html
>>as for the actual functionality, i'll leave it to ghuson to comment, as
>>i think he's got plans for ra_svn's authentication.
>>
>>-garrett
>>
>>
>
>I'd be willing to work on a better version. :) But we kind of needed
>something now to prove to management that this is a doable solution. ;)
>
oh, i completely understand.
>And as I said, this isn't really "authentication"... It's passing a username
>into svnserve. ;)
>
i realize that, but if greg has some kind of scheme for how
svnserve/ra_svn will do this sort of thing in the future, then it might
be better to keep this as a local patch until that solution
materializes. we'll just have to wait and see what he says.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Mar 25 19:30:44 2003