On Tue, 7 Dec 2010 23:47:47 +0100 Andreas Tille <ti...@debian.org> wrote:
> Hi Neil, > > thanks for your effort in solving this. > > On Tue, Dec 07, 2010 at 09:44:17PM +0000, Neil Williams wrote: > > As no other solution has actually been uploaded, I propose that the > > two packages involved should both be changed to use sub-directories > > just like the one other library containing the relevant file. Hence, > > splitting this bug report into two to allow two uploads. > > > > It's fairly trivial to adapt relevant code to use a subdirectory, > > with surrounding #ifdef where appropriate. > > ... > > I'll make the upload to delayed/2 tonight. (This bug has been > > hanging around without a resolution for long enough.) > > Probably you missed my remark when I deleted the pending tag [1]. No, I saw that. The two files *are* different, albeit still similar, hence every reason to have *both* files in different locations. There isn't necessarily a need for a common interface within the packages - just as long as any code including these headers can uniquely identify the *right* header consistently and reproducibly. > While in principle no real harm is done by your solution it would have > been IMHO perfectly correct to not fix the bug until both projects > really issued a reasonable common header file which would be > technically the best solution. Really? If the files are different but can be used in similar ways that argues for separation, not integration IMHO. > The fact that there is now a > workaround seems to decrease the presure onto this clean solution. I'm not at all sure why the solution has to be in one or other library as if code including the header (none of which is in Debian currently) cannot be trivially modified with an #ifdef or two. > I > wonder whether you simply was hunting for open bugs or whether you > really need one of those libraries. If it would be libctapimkt I > hope you did not missed my hint that there are other issues which > bring the package into a state which makes it unfit for release. It was just a hint, no detail, no proof or way for others to test. This bug isn't enough to prevent either package being in Squeeze - if there are other issues, please file appropriate bugs and give people a chance to verify and possibly fix. If you believe this to be the case, please file the appropriate RM bug with the reasoning. -- Neil Williams ============= http://www.linux.codehelp.co.uk/
pgpPIZQ0argdl.pgp
Description: PGP signature
