Hey Doug,
Once I get FreeBSD working I don’t mind doing most of the grunt work in getting
net and openbsd back in and updated. Then maybe someone who knows the factor
far better can review it. I’m trying to move off Linux to BSD but I really
want to take factor with me.
Jack
On Tue, Feb 5, 2019 at 11:10 AM, Doug Coleman <doug.cole...@gmail.com> wrote:
> Right! ``{free,open,net}bsd`` can be bsd io-backend again.
>
> I don't mind setting up test machines for freebsd and openbsd if someone else
> does most of the work getting Factor them ported again.
>
> Doug
>
> On Tue, Feb 5, 2019 at 10:46 AM John Benediktsson <mrj...@gmail.com> wrote:
>
>> You might also find this informative:
>>
>> https://github.com/factor/factor/commit/8cf18d1a82f08d1e9edf20f38b42eb1699ca0e67
>>
>> On Feb 5, 2019, at 7:28 AM, Jack Lucas via Factor-talk
>> <factor-talk@lists.sourceforge.net> wrote:
>>
>>> That’s perfect. I think this’ll set me back on track. Thanks Doug!
>>>
>>> Jack
>>>
>>> On Tue, Feb 5, 2019 at 9:37 AM, Doug Coleman <doug.cole...@gmail.com> wrote:
>>>
>>>> Hi Jack,
>>>>
>>>> You can try a more minimal bootstrap like this:
>>>>
>>>> ./factor -i=boot.unix-x86.64.image -include=math
>>>> ./factor -i=boot.unix-x86.64.image -include="math compiler"
>>>> ./factor -i=boot.unix-x86.64.image -exclude="io"
>>>>
>>>> After a quicker bootstrap you can do:
>>>> rlwrap ./factor
>>>> "io" require
>>>>
>>>> The code at work lives here:
>>>>
>>>> ! basis/bootstrap/stage2.factor
>>>> : load-component ( name -- )
>>>> dup "* Loading the " write write " component" print
>>>> "bootstrap." prepend require ;
>>>>
>>>> : load-components ( -- )
>>>> "include" "exclude" [ get-global " " split harvest ] bi@ diff
>>>> [ load-component ] each ;
>>>>
>>>> That might help you debug it yourself. I'm not sure what is happening
>>>> besides obviously not having an implementation for that
>>>> ``remove-input-callbacks``. You might check which io-backend is set:
>>>>
>>>> io-backend get .
>>>> macosx
>>>>
>>>> You can also set one with ``set-io-backend``.
>>>>
>>>> The io-backend in the vm is ``c-io-backend``. You might need to make a
>>>> ``unix-io-backend`` or something...maybe see how the old freebsd port did
>>>> it?
>>>>
>>>> http://downloads.factorcode.org/freebsd-x86-64/
>>>>
>>>> I hope this helps a bit,
>>>> Doug
>>>>
>>>> On Tue, Feb 5, 2019 at 7:35 AM Jack Lucas via Factor-talk
>>>> <factor-talk@lists.sourceforge.net> wrote:
>>>>
>>>>> Sure thing.
>>>>>
>>>>> https://pastebin.com/sWL9P3ex
>>>>>
>>>>> This includes all 3 stacks. Factor VM was compiled in debug mode if it
>>>>> helps.
>>>>>
>>>>> Jack
>>>>>
>>>>> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
>>>>> On Tuesday, February 5, 2019 3:55 AM, Jon Harper <jon.harpe...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi Jack,
>>>>>> Can you post the full log somewhere on the internet?
>>>>>> Jon
>>>>>>
>>>>>> Le mar. 5 févr. 2019 à 01:05, Jack Lucas via Factor-talk
>>>>>> <factor-talk@lists.sourceforge.net> a écrit :
>>>>>>
>>>>>>> Hello all. I'm trying to rekludge together freebsd support for factor.
>>>>>>> This has mostly involved trying to integrate some of the older commits
>>>>>>> that have it and using Kernigh's OpenBSD factor fork as a guideline.
>>>>>>>
>>>>>>> I don't know if it'll be good enough for a pull request but it has been
>>>>>>> very helpful for a focused tour of the factor source code. One issue
>>>>>>> I'm struggling with right now is trying to fix a problem that keeps
>>>>>>> popping up in the stage 2 bootstrapping. It compiles everything
>>>>>>> seemingly fine until it reaches the "io" component. Then it gets to
>>>>>>> loading the "basis/unix/utilities/utilities.factor" before loading
>>>>>>> "bootstrap-error" and throwing the "die" word.
>>>>>>>
>>>>>>> Now I'm not asking for anyone to take time out of their busy day to fix
>>>>>>> the issue for me, but I am wondering how the general process should
>>>>>>> go. In the data stack from this crash the only issue I can see in my
>>>>>>> novice eyes is a "source-files.errors:source-file-error T{
>>>>>>> generic.single:no-method f
>>>>>>> io.backend.unix.multiplexers:remove-input-callbacks }"
>>>>>>>
>>>>>>> But then why is failing at "utilities.factor" which doesn't include a
>>>>>>> reference to that word? The method is definitely defined in
>>>>>>> multiplexers.factor so my first thought is the specialization of it in
>>>>>>> the freebsd multiplexers/kqueue.factor is what's screwing up.
>>>>>>>
>>>>>>> My question seems to be thus, if utilities.factor is the last file it
>>>>>>> has to load for the io group then it must be doing some computation
>>>>>>> after it loads all these files that's causing it to fail. Is there any
>>>>>>> way to see what it does with these files? And likewise is there
>>>>>>> anyway to tell if a file somehow silently failed to load but the
>>>>>>> bootstrap kept going anyway; mostly just for the sake of trying to
>>>>>>> figure out how to get more helpful information out of this process.
>>>>>>>
>>>>>>> If these questions seem to strike you as the ramblings of a feeble
>>>>>>> newbie than please don't waste any time in passing this by; I'll keep
>>>>>>> bashing my head against the keyboard until I "get it". :P
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Factor-talk mailing list
>>>>>>> Factor-talk@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/factor-talk
>>>>>
>>>>> _______________________________________________
>>>>> Factor-talk mailing list
>>>>> Factor-talk@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/factor-talk
>>
>>>
>>
>>> _______________________________________________
>>> Factor-talk mailing list
>>> Factor-talk@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/factor-talk_______________________________________________
Factor-talk mailing list
Factor-talk@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/factor-talk
