On Sat, Mar 20, 2021 at 12:37:47AM +0100, Ralf Hemmecke wrote:
> > Concerning "low profile developers",
> > I expect most of FriCAS developement to be done independently
> > of "make".
> :-(
>
> Why do you expect that? Because that is your development style or
> because the build system is not smart enough to only compile the stuff
> quickly that a developer changed?
>
> Thanks to Qian we have CI working at github, so how local development
> happens is mostly to the taste of the developer. Still, it would be
> better if the build system was clever enough to just do the right thing
> when one (re-)types "make".
1) IME even in projects heavily geared toward "make" much of
developement was independent of "make". Namely, it was
running program under debugger, trying various tests and
observing state. Only after collecting enough data it
was possible to write code fixing a bug or implementing
new functionality.
2) Our algebra build have to solve bootstrap problem. Due
to this it is hard to avoid rebuilding whole algebra
after seemingly trivial change. Developer should have
reasonably good idea if change requires rebuilding
other files. "make" sees only file date/time and with
this must act conservatily and rebuild a lot.
3) Beside rebuild time, there is also problem of test data
and command line history. "gdb" keeps some testing
state (breakpoints, etc) and command line history and
picks new new executable after change. Lisp works by
loading changed code, which is nice but does not fit
well with "make". By recompiling file that I work on
I can keep my testing state in Lisp image and get
reasonably fast developement cycle: most time is my
thinking time, biggest delay is Spad compilation
which for single file is usually short enough to
not disturb thinking. From that point of view
"make" (or ruther time spent running it) is a cost,
which most projects must pay, but fortunately we do
not have to.
BTW: I suspect that Python folks also are less eager to
run "make" (OTOH Python bytcode compilation means that
make can be quite fast).
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/fricas-devel/20210320002145.GA35450%40math.uni.wroc.pl.
