Re: [Qemu-devel] QEMU desired libiscsi.so clashes with libiscsi.so from iscsi-initiator-utils

[ronnie sahlberg]

Sorry about this.

First, libiscsi is a really good name for a general purpose
multiplatform library, like libiscsi.
Second,  a generic name like this is a horribly poor idea for a single
distribution/ single use / obscure private library.


I want to solve a problem  to make it available on all platforms.
I dont like to chose a suboptimal name for this reason,  but I thought
I had no choice.


I am not really excited with the concept of "obscure single use
private library polluting the namespace like this" but what are my
options ?
I can live with renaming my library if that is what it takes.


What do you want me to do ?
Tell me and I will do it.



regards
ronnie sahlberg


On Tue, Mar 6, 2012 at 9:38 PM, Michael Tokarev <address@hidden> wrote:
> On 06.03.2012 14:07, Daniel P. Berrange wrote:
>> On Sat, Mar 03, 2012 at 03:54:19PM +0400, Michael Tokarev wrote:
>>> It looks like iscsi-initiator-utils package in fedora is built using
>>> shared libiscsi which appears to be package-specific, e.g.
>>
>> Yes, it is the iscsi-initiator-utils package
>>
>>> I'd rename _that_ library to be libiscsi-initiator-utils.so, to
>>> reflect the name of the package in question, since it - again -
>>> _appears_ to be package-specific.
>>
>> That's not really practical at this time, since it has been shipped this
>> way for years now,
>
> Why it is not practical for something which is not used for
> anything but just the internal library?  I'd say it was a
> mistake to name this internal library this way to start
> with, and now it is as good time to do so as any other time,
> because, again, it is just some internal-to-the-package thing.
>
>>          and in any case ronnie has already been good enough
>> to rename to libiscsiclient to avoid the potential ambiguity.
>
> Now the resulting libiscsiclient has alot of _real_ ambiguity,
> unfortunately.
>
> /mjt