On 9/1/22 19:21, Simon Marchi wrote: > On 9/1/22 10:58, Tom de Vries via Gdb-patches wrote: >> On 9/1/22 16:43, Simon Marchi wrote: >>> On 9/1/22 05:51, Tom de Vries via Gdb-patches wrote: >>>> Hi, >>>> >>>> Currently, we cannot build gdb without makeinfo installed. >>>> >>>> It would be convenient to work around this by using the configure flag >>>> MAKEINFO=/usr/bin/true or some such, but that doesn't work because top-level >>>> configure requires a makeinfo of at least version 4.7, and that version check >>>> fails for /usr/bin/true, so we end up with MAKEINFO=missing instead. >>>> >>>> Work around this by adding a script gdb/contrib/makeinfo-dummy.sh that can be >>>> used instead, like so: >>>> ... >>>> $ ./src/configure MAKEINFO=$src/gdb/contrib/makeinfo-dummy.sh >>>> ... >>>> >>>> The script merely prints the version string that satisfies the version check >>>> in $src/configure. >>>> >>>> Tested on x86_64-linux, with makeinfo removed. >>>> >>>> Any comments? >>> >>> I wouldn't mind having this if it helps you. This file would probably >>> belong in $top_level/contrib more than $top_level/gdb/contrib. >>> >> >> Hi Simon, >> >> thanks for the review. >> >> Agreed, but does that also mean submitting to gcc-patches instead? All the other scripts in /contrib seems to be copies from /contrib. > > IMO, it would be fine to add a file like this just in binutils-gdb, and > if someone wants it in gcc they can copy it there. When making changes > to files existing in both repos, then yeah it's better to involve gcc to > so that they stay in sync. > >> >>> Just in case that would work for you, in this case I have been using: >>> >>> $ ./configure >>> $ make MAKEINFO=/bin/true >>> >>> and that worked fine so far. >> >> I see, I suppose that'll work as well, I didn't think of that. >> >> I guess it matters more for me to have a documented way of dealing with this, than some specific solution. I ran into this today, knew I worked around this before, but didn't remember how. I'm hoping that having this file in the repo will be something that is easy to find back and use. >> >> It would be possible to handle 'true' in the toplevel configure, which would make configure MAKEINFO=true work, but that would have to go past gcc-patches review, and I'm not sure if it's worth the effort to try this, given my suspicion that it will not be accepted. > > Also more efforts, but what we would want is a --disable-doc or > --without-makeinfo configure switch, I think. I gave the --without-makeinfo a try. WDYT? Thanks, - Tom