On Fri, September 28, 2018 13:25, Sebastien Marie wrote:
> On Fri, Sep 28, 2018 at 11:53:48AM +0200, Antoine Jacoutot wrote:
>>
>> It's the kqueue implementation in it that is half broken and unmaintained
>> (and
>> that was written for BSD, so definitely *not* NIH). If someone can have a
>> look
>> at it and make sure it's working fine, then we can open the usage to a
>> broader
>> amount of ports.
>
> Maybe starting by updating the port ?
>
> We have 20170711 and the latest is 20180201.
>
> https://github.com/libinotify-kqueue/libinotify-kqueue/releases/tag/20180201
>
> 1. Directory diffing has been rewritten. No more memory hungry shallow
> lists allocations/deallocations
>
> 2. Fix compilation on platforms where kevent.data size is bigger than
> void * (issue #4)
>
> 3. Workaround sporadic crashes caused by reversing of kevent order
> (issue #5)
>
>
> $ make test
> ...
> --------------------
> Run: 125
> Passed: 109
> Failed: 0
> Skipped: 16
>
>
> The diff below is just a base for more testing. I have *not* tested the
> reverse-dependencies:
>
> devel/hs-hinotify
> devel/py-pyinotify
> devel/py-pyinotify,python3
> graphics/sxiv
> net/owncloudclient
> net/seafile/seafile
> sysutils/incron
> sysutils/inotify-tools
> sysutils/lsyncd
> x11/kde4/libs,-main
> x11/xmobar
>
Hi!
Ive tested simmilar diff some time ago with owncloudclient and seafile
and met no issuses. Not using other consumers so can't said if the're
working fine but at least all of theb built fine for me earlier.
No comments:
Post a Comment