IRC channel logs
2014-03-30.log
back to list of logs
<phant0mas>patching files for double declarations,bypassing missing header files by puting them in the right place <phant0mas>I might need someting to keep me awake after all :P *Digit798 gets curious as to how easy or extensive it is to simply "drop-in" guix into an existing installed distro, whatever it might be, barebones busybox, or some more conventional os laden with already installed apps viaits own packagemanager <rigelk>hi civodul ; i've been in exams last week, and I'm still coding on the GNUnet service. <rigelk>civodul: going to vote for now ;) <rigelk>i'll be back for tonight ; i have a roadmap to clarify if I recall <phant0mas>had some issues with double declarations of variables in glibc, which I pathed, some files were looking for the files in the wrong place <phant0mas>I am actually reading the errors I get during building and I apply the coresponding patches:P <civodul>esp. in terms of experimentation with Guix and GNUnet <civodul>phant0mas: if there are genuine glibc build issues, make sure to report them to bug-hurd <civodul>by genuine, i mean that are likely not due to the build setup <civodul>anyway, it seems you're making good progress :-) <phant0mas>about the misplasements of the header files it may be something wrong in my side <phant0mas>but about the double declaration I will report it <phant0mas>I asked yestarday in #hurd but took no answer <phant0mas>for now I commented the one out and it seems to work <civodul>make sure to document build issues you have to paper around <guest`>why there's 2 '("python" ,python-2)' in icecat's native-inputs field? <civodul>presumably Python 2 is needed at build time <guest`>civodul: redundancy, or two times means it is REALLY needed? <phant0mas>civodul: the amounts of patches used in debian and not present in upstream is unbelievable <civodul>phant0mas: yes, that's a problem :-/ <civodul>guest`: the string "python" allows it to be refer from build-side code, whereas ,python-2 is a reference to the actual package definition for Python 2 <guest`>civodul: yes, but there's two lines like that <guest`>civodul: I can do that someday, I need to learn git :P <phant0mas>civodul: I am starting to believe that the upstream cannot be built without the debian patches <phant0mas>I am hand picking them now to see which ones we need <civodul>phant0mas: that's very well possible, unfortunately <civodul>when you have concrete data, we should email bug-hurd with the details <phant0mas>ok, I am writing the issues on a paper so I can concetrate them *civodul makes good progress on signed substitutes <zacts>what reliability features does dmd offer? <zacts>civodul: does dmd offer crash recovery of services, or even of the dmd program itself? <civodul>zacts: well, like all init systems it tries to respawn services that fail, up to a limit <civodul>if it crashes itself, you're screwed <phant0mas>came that close...and I think it failed at linking <phant0mas>civodul: I am too tired to continue today :P <phant0mas>tomorrow I am gonna send an email to bug hurd to report the things I had to change and ask about the part that fails <civodul>make sure it's solid enough, so as to not bother them <civodul>mark_weaver: comment on the signed substitutes thing welcome! :-) <mark_weaver>civodul: I've been distracted for the last couple of days, but I'm glad to hear there's progress with that :)