deltacloud-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Francesco Vollero <razor...@gmail.com>
Subject Re: libdeltacloud bug fix and enhancement requests
Date Sun, 26 Aug 2012 13:28:34 GMT
Hi Chris,

On Sun, Aug 26, 2012 at 3:01 PM, Chris Lalancette <clalancette@gmail.com> wrote:
> On Sun, Aug 26, 2012 at 4:34 AM, Francesco Vollero <razorinc@gmail.com> wrote:
>> Hi Todd,
>>
>> On Sun, Aug 26, 2012 at 6:17 AM, Todd Loeber <mloeber@gmail.com> wrote:
>>> Bug) In file curl_action.c, do_get_post_url uses strdup to copy
>>> chunk.data, truncating binary blob content during download.
>>>
>>> I confirmed that returning chunk.data and header_chunk.data (instead
>>> of strduping and freeing them) allows retrieval of binary data and
>>> incidentally avoids an unnecessary copy.  The caller is already
>>> responsible for freeing the output parameters.
>>>
>>> If libdeltacloud is still alive, it would be nice to have
>>> do_get_post_url fixed so binary blobs can be downloaded.  And also,
>>> some enhancements!
>>>
>>
>> the libdeltacloud project is not currently maintained, but we would be
>> more than pleased to see that libdeltacloud keep gaining some
>> interest, with some new "fresh" patch, so why not help us to keep this
>> project alive? :) :) :)
>
> That's actually not strictly true.  I do still maintain it, though I
> am not actively working on it.  Any patches to improve it would
> definitely be welcome.
>

Let me be more clear, the project libdeltacloud afaik is not supported
by the core team of deltacloud.

> As far as this particular issue is concerned, yeah, I see the problem
> with using strdup.  However, I don't think we can just remove the
> strdup().  I'm pretty sure I wrote it that way for a reason, though I
> honestly don't remember the reason.  My suggestion would be to change
> the strdup to a memcpy, which should fix this particular problem.  If
> you really want to remove the strdup, you would have to go through all
> of the callers of do_get_post_url and make sure that none of them
> require the duplicated data.
>
> Anyway, thanks for looking into it!
>
> Chris

Mime
View raw message