On Tue, Mar 05 2019, Brian Callahan <bcallah@devio.us> wrote:
> On 3/5/19 12:11 AM, Jeremie Courreges-Anglas wrote:
>> webkitgtk4 fails to build on sparc64 and powerpc.
>>
>> http://build-failures.rhaalovely.net//sparc64/2019-02-03/www/webkitgtk4.log
>> http://build-failures.rhaalovely.net/powerpc/2019-02-04/www/webkitgtk4.log
>>
>> With the diff below I can successfully build the port on sparc64.
>> Rationale for the diff:
>> - patches/patch-Source_WTF_wtf_Platform_h: sparc64 MD code doesn't
>> expose enough registers to use the HAVE_MACHINE_CONTEXT code path.
>> - patches/patch-Source_JavaScriptCore_runtime_MachineContext_h: see above
>> - patches/patch-Source_WTF_wtf_RAMSize_cpp: Linux-only method to get the
>> amount of ram, use sysctl(2) instead.
>> - patches/patch-Source_WebCore_platform_graphics_FontSelectionAlgorithm_h:
>> std::min/std::max from libestdc++ aren't annotated with "constexpr"
>> yet so using them in a "constexpr" function fails. This workaround
>> shouldn't be needed with lang/gcc/8.
>> - Makefile: I doubt that webkit's JIT works on sh... While here, switch
>> to https HOMEPAGE/MASTER_SITES.
>>
>> I haven't tested runtime yet but intend to do so in the next days. The
>> original motivation is to unlock a bunch of ports:
>>
>> http://build-failures.rhaalovely.net//sparc64/2019-02-03/summary.log
>>
>> --8<--
>> ritchie ~$ grep -c webkitgtk4 summary.log
>> 114
>> ritchie ~$ wc -l summary.log
>> 296 summary.log
>> -->8--
>>
>> The diff could also fix powerpc (cc Charlene).
>
> I don't have a sparc64 machine but a bit ago cwen and I came up with
> a similar diff to RAMSize.cpp to fix webkit on macppc (I think we used
> sysconf).
sysconf might be a better candidate indeed, at least it should be
reasonably portable on not-so-ancient unix-like systems out there.
> JavaScript failed at runtime but I think it's better to get
> a successful build than not.
Yeah www/surf doesn't seem very happy on sparc64...
> Same for the gcc-4.9 bits. Those bits are
> good.
Committed!
--
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
No comments:
Post a Comment