Bug#756722: lintian: false positive for embedded-library error

[Jakub Wilk <jwilk@debian.org>]

Control: tags -1 + confirmed

* Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp>, 2014-08-02, 15:28:
> > > 'embedded-library usr/lib/texmacs/TeXmacs/bin/texmacs.bin: freetype', 
> > > automatically rejected package.
> > [...]
> > > I rebuild the (accepted) former version of texmacs but got the same 
> > > lintian error so I believe this should be false positive.
> >
> > Hmm, I can't reproduce it here. I rebuilt texmacs (1:1.0.7.18-1) in 
> > up-to-date unstable amd64 chroot, but the resulting packages don't 
> > trigger any such error.
>
> Sorry but I mean texmacs 1.99.1-1 (accepted one) and 1.99.1-2 (rejected 
> one) in experimental.

Oh, I didn't notice that there's a package in experimental. That was a 
bit of information I was missing.

Even the package from the archive triggers the Lintian error:

$ lintian -F texmacs_1.99.1-1_amd64.deb
E: texmacs: embedded-library usr/lib/texmacs/TeXmacs/bin/texmacs.bin: freetype

(The package was accepted to the archive, because at that point Lintian 
didn't have the check for freetype implemented yet.)

To detect embedded copies of freetype, Lintian looks for the string 
"FT_Get_CID_Is_Internally_CID_Keyed" in the binaries. But this is a name 
of a public function, and texmacs just happens to use it. So it's indeed 
a false positive.

I'm not sure why exactly "FT_Get_CID_Is_Internally_CID_Keyed" was 
chosen, and what would be a better string to use... Until we figure 
it out, please add an override for the tag.

> I'll upload the package to my homepage but my main machine is in my 
> office so it will happen after monday.

It's no longer necessary. :-)

--
Jakub Wilk