Hi,
Here is a quick dump of the patches that I have used to get around most of the
issues with the previous ISO image. I thought that it was something in my
chroot environment that was causing the issues, but scsijon is reproducing them
with the current ISO image.
These fixes were not yet meant to be final solutions, their main purpose was to
be able to continue building. I can send some annotation on these when I
return home, tomorrow night, but most of this should be self explanatory to
scsijon since you are hitting the same issues in the same order.
Jan
On Tuesday, January 21, 2020, 04:04:14 PM PST, scsijon
<[email protected]> wrote:
On 21/01/20 21:11, scsijon wrote: > > we will see what jan rovins comes up with; > > On 21/01/20 20:11, René Rebe wrote: >> Hi, >> >> >>> On 21. Jan 2020, at 07:08, scsijon <[email protected]> wrote: >>> >>> >>> more with ninja >>> >>> On 21/01/20 15:14, scsijon wrote: >>>> Next a Ninja build fail, but actually a Python problem:- >>>> Got as far as ninja, but couldn't build ninja, tracking showed me >>>> there was no python executable in the installed r49128 to start >>>> with, or anywhere in the TOOLCHAIN/cross.build/ directories, as well >>>> as a few other related problems. >>>> It seems with python3.7 and with no python2.x there is nothing for >>>> the link created to link into. I added a link from >>>> /usr/bin/python3.7 to create the /usr/bin/python link. >>>>> I added this and ninja built, but there may be other problems >>>>> later, i'm still building 0- packages at this stage. >>> >>> the python link seems to have worked it's magic, however ninja is >>> erroring after the line > >>> >>> creating >>> /t2-trunk-r49129/build/minimal-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/user<..> >>> >>> if required ... >>> fl_wrapper.so: write outside basedir (/t2-trunk-r49129) >>> /usr/lib/python3.7/encodings/__pycache__/__init__.cpython-37.pyc.140658706702768 >>> >>> >>> due to previous errors, no 0-ninja.log file! >>> >>> etc.... > > > yeah, python seems still to have a problem > > >>> >>> advice is wanted on this one, i'll try marking the package without 0 >>> for now to test. >>> > > done, -1---5---- for ninja for the moment, will get back to it when > lzma.py has appeared! > >>>> Also, the python.conf (for python3 )has a python_postmake(), but >>>> there is no /lib/python to link into ?so not sure what to do here, >>>> and python2's python2.conf had a python_postmake() had a different >>>> export PYTHON line that ?may be needing to add into the python.conf >>>> so both PYTHON and PYTHON3 have export lines. >>>>> this has yet to be sorted properly, however to test I have just >>>>> copied the python2.conf's export line into python.conf's under the >>>>> python3's one and will see what happens when i rebuild. >>> > > >>> and a new error with glib needing lzma for python3 (not in minimal by >>> default). > > no, not the total fix, apparently lzma is needed for python to create > lzma.py whicch is needed for the latest glib, but it's still not being > built at present, > > i'll chase this one up tomorrow afternoon (my time), split shift (Early am & > evening, yuck) > > looks like there needs a switch to be added in python3's python.conf. > > Worse problem is there are a number of different lzma.py files, i'll > have to first work out which one is needed, as a couple i've already > looked at want packages not in t2 yet like cpython > (https://github.com/python/cpython)which might be worth adding anyway, > and liblzma (https://github.com/kobolabs/liblzma) to start with. > >> well, with lzma installed the lzma.py IS being built and installed in the TOOLCHAIN and the correct directory, the problem actually seems to be with building glib, it's finding 'the' file, it's just not able to process it correctly, and the file is almost the same as my installed earlier 3.5 version which works ok, plus glib is now built with meson and supposedly ninja not ./configure and make, so i'd best go back to the ninja problem and try to sort that out first. Thought it might happen, but was hopeing a fallback build was still included in the packages . At least I now know some of others build sucessfully. >> I noticed the same live on YouTube >> (https://www.youtube.com/watch?v=VMyOrmmk7ok), >> IMHO really ehem, annoying upstream changes / regressions. It should >> not hard depend >> on something like lzma, I matched that away temporarily, need to >> commit some fix soon. > > ok, thanks Rene, > but now I don't know whether to try to fix it or just set a switch in > glib not to build it for now. I think i'll see if it's easy to have it > included in the python build, else i'll try setting it off in glib for > now, after all i'm working at getting it building through 2-$package at > present, a full t2-minimum build is still some time away before I start > adding Puppy-Quirky-EasyOS stuff into the mix and "causing headaches > online". :-)} (bearded smile) > >> >> Greetings, >> René >> > > thanks again for all your continued work and putting up with me, > scsijon > > ps, do you have a list of your 'still active' youtube videos somewhere > (please if possible), I think i've said before that as i'm on satellite > here they are just too costly to run locally, but I will have a couple > of hours in the next two weeks when elsewhere for meetings, and have a > chance to run at least some of them if I can (as expected), 'borrow' a > workstation there, as they are fibre-to-the-premises it's cheap. I can > try to get up to date with what your working on/at then. > > > > ----------------------------------------------------------- If you wish > to unsubscribe from this mailing, send mail to > [email protected] with a subject of: unsubscribe t2 > ----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [email protected] with a subject of: unsubscribe t2
t2.diffs
Description: Binary data
----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [email protected] with a subject of: unsubscribe t2
