In response to your latest report, I"ate my own dogfood" by downloading/extracting/building from the current github zipfile.
Still (again) 64-bit antiX 16 here (dunno how/why 32 vs 64bit might behave differently though) and it it successfully compiled.
Code: Select all
desktop-menu --write-out-global
make[1]: desktop-menu: Command not found
dpkg-query -S /usr/local/bin/desktop-menu
^---v
desktop-session-antix: /usr/local/bin/desktop-menu
Your assessment seems correct. According to this:
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"https://github.com/BitJam/Build-iso/search?utf8=%E2%9C%93&q=desktop-session-antix&type="
linktext was:"https://github.com/BitJam/Build-iso/sea ... ntix&type="
====================================
the"desktop-session-antix" package is preinstalled to only"base" and"full" editions.
If you installed just that one missing package to your"core" edition, breakage (idunno about build failure) might still occur.
As mentioned in the INSTALL file & my early posts in this topic, handling this"test build" is awkward, because antiX fluxbox
implementation involves bits-n-pieces provided by multiple, separate .deb packages (e.g."desktop-defaults-fluxbox-antix").
Hmm, I had anticipated that even if eventually this ships pre-installed, those packages should remain separate.
(Avoids needing to rebuild the package containing executables when preinstalled themes are swapped in/out)
I can't make a decision"how to accommodate core edition?" on my own.
As is, this fluxbox package would rightfully need to declare"desktop-defaults-fluxbox-antix" as a dependency.
(and could"squeak by", listing"fluxbox-themes-antix" as a
Recommends, but I'm not keen on that prospect)
That package (d-d-f-a) is only 300Kb or so; hopefully its pre-installation in"core" edition wouldn't be objectionable
or a mechanism is already in place
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"https://github.com/BitJam/Build-iso/blob/53d86d50b050a2d30c3ae83e97b347b036bf02fd/Template/COMMON/delete-files.list"
linktext was:"https://github.com/BitJam/Build-iso/blo ... files.list"
====================================
to nix its content during the automated build process, when generating the"core edition" iso.
Again, thanks for testing and helping to flesh out these prospective"gotchas" and"whatifs".
ps:
Today, in response to the earlier build failure you described
(dangit, presence of second arg should be moot. The gcc v4.9+ compiler should just ignore, or raise notice, but not fail)
I tested:
-- edit"configure.ac" and remove the second argument
-- edit its counterpart (at bottom of"ax_cxx_compile_stdcxx_11.m4")
and confirmed the build is successful for me, with or without presence of the second argument.