app-arch/lzop: Fix building with GCC-6 and purge deprecated 'dohtml'

[Peter-Levine]

Bug: https://bugs.gentoo.org/show_bug.cgi?id=594472
Package-Manager: Portage-2.3.5, Repoman-2.3.2

For GCC-6, signed overflow in a declared array's size is treated as a non-constant (see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69746).

This causes a conftest failure at configure time. In particular:

src/miniacc.h:4475:24: warning: result of '1 << 31' requires 33 bits to represent, but 'int' only has 32 bits [-Wshift-overflow=]
     ACCCHK_ASSERT((1   << (8*SIZEOF_INT-1)) < 0)
                        ^
src/miniacc.h:1712:75: note: in definition of macro 'ACC_COMPILE_TIME_ASSERT_HEADER'
 \#    define ACC_COMPILE_TIME_ASSERT_HEADER(e)  extern int __acc_cta[1-2*!(e)];
                                                                           ^
src/miniacc.h:4475:5: note: in expansion of macro 'ACCCHK_ASSERT'
     ACCCHK_ASSERT((1   << (8*SIZEOF_INT-1)) < 0)
     ^~~~~~~~~~~~~
src/miniacc.h:1712:59: error: variably modified '__acc_cta' at file scope
 #    define ACC_COMPILE_TIME_ASSERT_HEADER(e)  extern int __acc_cta[1-2*!(e)];
                                                           ^
tst.c:134:33: note: in expansion of macro 'ACC_COMPILE_TIME_ASSERT_HEADER'
 #define ACCCHK_ASSERT(expr)     ACC_COMPILE_TIME_ASSERT_HEADER(expr)
                                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
src/miniacc.h:4475:5: note: in expansion of macro 'ACCCHK_ASSERT'
     ACCCHK_ASSERT((1   << (8*SIZEOF_INT-1)) < 0)
     ^~~~~~~~~~~~~
src/miniacc.h: In function 'test_acc_run_time_assert':
src/miniacc.h:4475:24: warning: result of '1 << 31' requires 33 bits to represent, but 'int' only has 32 bits [-Wshift-overflow=]
     ACCCHK_ASSERT((1   << (8*SIZEOF_INT-1)) < 0)
                        ^
src/miniacc.h:4475:5: note: in expansion of macro 'ACCCHK_ASSERT'
     ACCCHK_ASSERT((1   << (8*SIZEOF_INT-1)) < 0)
     ^~~~~~~~~~~~~
src/miniacc.h:4480:24: warning: result of '1l << 63' requires 65 bits to represent, but 'long int' only has 64 bits [-Wshift-overflow=]
     ACCCHK_ASSERT((1l  << (8*SIZEOF_LONG-1)) < 0)
                        ^
src/miniacc.h:4480:5: note: in expansion of macro 'ACCCHK_ASSERT'
     ACCCHK_ASSERT((1l  << (8*SIZEOF_LONG-1)) < 0)
     ^~~~~~~~~~~~~
src/miniacc.h:4709:42: warning: result of '128 << 24' requires 33 bits to represent, but 'int' only has 32 bits [-Wshift-overflow=]
     ACCCHK_ASSERT( (((unsigned char)128) << (int)(8*sizeof(int)-8)) < 0)
                                          ^
src/miniacc.h:4709:5: note: in expansion of macro 'ACCCHK_ASSERT'
     ACCCHK_ASSERT( (((unsigned char)128) << (int)(8*sizeof(int)-8)) < 0)
     ^~~~~~~~~~~~~

There exist two patches. One from Debian and one from Patrick McLean on Gentoo Bug-Tracker (via OpenSUSE et al.). The patch from Debian uses C11's _Static_assert(). The patch from Gentoo Bug-Tracker enables the originally intended bitwise-shifted expression while avoiding overflow. The latter is chosen.

[gentoo-repo-qa-bot]

Pull Request assignment

Areas affected: ebuilds
Packages affected: app-arch/lzop

app-arch/lzop: @mgorny

[gentoo-repo-qa-bot]

😞 The QA check for this pull request has found the following issues:

Issues inherited from Gentoo (may be modified by PR):
https://qa-reports.gentoo.org/output/gentoo-ci/61b23577b/output.html#sci-biology/mammoth
https://qa-reports.gentoo.org/output/gentoo-ci/61b23577b/output.html#sci-chemistry/ambertools
https://qa-reports.gentoo.org/output/gentoo-ci/61b23577b/output.html#sci-chemistry/gamess
https://qa-reports.gentoo.org/output/gentoo-ci/61b23577b/output.html#sci-chemistry/mead