15576 2024-08-11 20:40:51 +0000 e2fsprogs calls not fully capable makeinfo 2025-06-05 15:30:50 +0000 1 1 7 Build System, Metadata & Runtime OE-Core core unspecified All Multiple NEW Low minor Future 1 adrian.freihofer unassigned meta.mr.watcher meta.watcher randy.macleod ross.burton --- --- No (bug/feature does not impact docs) oldest_to_newest 99540 0 adrian.freihofer 2024-08-11 20:40:51 +0000 When building e2fsprogs or e2fsprogs-native the following message is found in the log.do_install file: Traceback (most recent call last): File "...poky/build/tmp/work/x86_64-linux/e2fsprogs-native/1.47.0/recipe-sysroot-native/bin/makeinfo", line 98, in <module> assert this_binary in simple_binaries, \ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: Don't know how to get default output file name from input file! This stand-in version of makeinfo is not yet fully capable of emulating the real version from the GNU texinfo suite. If you see this message, file a bug report with details on the recipe that failed. The assertion is thrown by: https://git.yoctoproject.org/poky/tree/meta/recipes-extended/texinfo-dummy-native/texinfo-dummy/template.py#n98 99542 1 randy.macleod 2024-08-15 14:38:36 +0000 This looks like an error but can you confirm that e2fsprogs actually builds properly. If so, we might add a comment in the log generation to explain that the log can be safely ignored or somehow eliminate the log in this case. 99732 2 randy.macleod 2024-09-26 15:02:16 +0000 There's a bug here but it's a low priority. 99733 3 randy.macleod 2024-09-26 15:05:18 +0000 ... 99735 4 ross.burton 2024-09-26 15:51:23 +0000 So the problem is that our texinfo stub is very dumb. e2fsprogs invokes "makeinfo ../../git/doc/libext2fs.texinfo" without telling it where a file will be written: proper makeinfo can parse the file and find the @setfilename command but our stub does not, this is what the error is trying to say. We could just assume ${basename}.info and hope for the best?