06:44 <metacollin> wpwrak: undelete is a reserved function in POSIX systems and is in one of the standard headers that gets included, so it has to be changed to a different name for it to compile on a POSIX compatible system

13:33 <wpwrak> metacollin: okay, that means that no #ifdef is really needed. now, the 2nd problem is that i can't find any indication that undelete is defined in posix, i.e., here: http://pubs.opengroup.org/onlinepubs/9699919799/

13:38 <wpwrak> so it's either NOT in posix, meaning that the mac headers don't conform, but and there may be a way to make them conform (e.g., through a feature test macro), or it's anticipating some unpublished version of POSIX. alas, http://www.manpagez.com/man/2/undelete/ (if that's the right reference) doesn't sat how it relates to POSIX

14:40 <metacollin> https://en.wikipedia.org/wiki/Unistd.h

14:41 <metacollin> https://www.freebsd.org/cgi/man.cgi?query=undelete&sektion=2

14:41 <metacollin> Also, my patch isn't meant to be a suggested code change per se, just to illustrate the issue

14:42 <metacollin> I don't know how you use or planned to use undelete within fped, so its really just a matter up to you. If you're ok with simply renaming it, then there is no need for an #ifdef at all

14:42 <metacollin> I just don't like changing people's function names willy-nilly :)

14:43 <Joerg-Neo900> wpwrak: ^^^ :-)

14:44 <Joerg-Neo900> wpwrak: and please mind the "release" tag

14:44 <metacollin> Oh, and one other change: using flex on macos is done via -ll rather than -lm, but I believe that is specific to macOS. And I am probably the only person who wants/cares about macOS support heh

14:45 <metacollin> er not -lm, -lfl

14:45 <metacollin> https://gist.github.com/metacollin/abf049d340e074b30593c68e01a16ab1 is the change, but after that, fped works great on macOS :)

14:45 <Joerg-Neo900> on a sidenote: did you manage to build eeshow?

14:46 <Joerg-Neo900> we had some weird issue with libgit changed API or somesuch

14:47 <Joerg-Neo900> to be seen in last few commits

14:47 <metacollin> yeah I have eeshow working too... I think. its giving a parse error for the ee-full repo's neo900.pro

14:48 <Joerg-Neo900> I think that's exactly this issue

14:48 <metacollin> neo900.sch:61: cannot parse

14:48 <metacollin> "F0 "Charger/OTG" 80"

14:49 <Joerg-Neo900> iirc it was something similar to "cd doesn't like a filename as parameter any more"

14:49 <Joerg-Neo900> ooh, that's sth different

14:49 <metacollin> oh i think I know what this is

14:51 <Joerg-Neo900> nested quotes?

14:51 <Joerg-Neo900> meh, the outer ones are prolly from eeshow itself

14:55 <Joerg-Neo900> metacollin: when you share diagnostic output to wpwrak, he could have a look and prolly spot the issue in no time

14:59 <metacollin> is Charger/OTG supposed to be referencing a file?

15:00 <Joerg-Neo900> wpwrak: for fped, the patch is prolly >>sed -i 's/ undelete(/ fped_undelete(/g' *.{ch} #avoid collision with https://www.freebsd.org/cgi/man.cgi?query=undelete&sektion=2<<

15:00 <metacollin> yeah, that would be totally fine if he doesn't mind just renaming his function

15:02 <Joerg-Neo900> comment "fix undelete() collision for MacOS/freebsd"; tag "stable release"

15:03 <Joerg-Neo900> nevermind OSIX, not worth investigating what's going on there

15:03 <Joerg-Neo900> Possix*

15:04 <Joerg-Neo900> ugh, charger/OTG shouldn't reference a file for obvious reasons

15:04 <Joerg-Neo900> the file would be "charger.sch"

15:05 <Joerg-Neo900> the sheet name however is possibly "charger/OTG"

15:07 <metacollin> hmmmm

15:07 <Joerg-Neo900> http://wstaw.org/m/2017/08/12/plasma-desktopcl2346.png http://wstaw.org/m/2017/08/12/plasma-desktopLO2346.png

15:11 <metacollin> yeah

15:12 <Joerg-Neo900> ;-)

15:13 <metacollin> I should be using the version of eeshow from http://neo900.org/git/eeshow correct?

15:14 <Joerg-Neo900> yes, that's canonical master (wording?) repo

15:16 <Joerg-Neo900> http://neo900.org/git/eeshow/tree/README?id=71c92484bd32e4006c485702212ec94a7f845916

15:18 <wpwrak> the parse error may be another POSIX-noncompliance. i use the "m" modifier, which is defined in POSIX.1-2008

15:18 <Joerg-Neo900> hi wpwrak! :-)

15:18 <wpwrak> metacollin: can't you just install glibc ? ;-)

15:21 <wpwrak> lack of support for %m is a problem. there's no easy way to replace it (short of rewriting *scanf)

15:23 <Joerg-Neo900> mind to link to a definition ofd what %m does?

15:23 <Joerg-Neo900> man scanf ?

15:24 <wpwrak> here is a discussion: https://stackoverflow.com/questions/38685724/difference-between-ms-and-s-scanf

15:25 <Joerg-Neo900> ta :)

15:26 <wpwrak> anyway, gotta run. the undelete problem may be fixable with feature test macros.

15:27 <Joerg-Neo900> hmmm >>assignment-allocation character m, which shall cause a memory buffer to be allocated to hold the string<< indeed that's nontrivial