Sujet : Re: Python recompile
De : antispam (at) *nospam* fricas.org (Waldek Hebisch)
Groupes : comp.lang.cDate : 10. Mar 2025, 14:00:57
Autres entêtes
Organisation : To protect and to server
Message-ID : <vqmnq7$3acv2$1@paganini.bofh.team>
References : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
User-Agent : tin/2.6.2-20221225 ("Pittyvaich") (Linux/6.1.0-9-amd64 (x86_64))
Michael S <
already5chosen@yahoo.com> wrote:
Pay attention that all this slow, complicated 'configure' business
didn't originate on Windows. It was invented in order to cover variety
of Unixen. Which (variety) no longer exists, but religious 'free
software' people like to pretend that it is still relevant and continue
to use configure.
There is a lot of variety on "single" system, that is Linux.
Supporting this variety is important, it allows experimentation
and fast evolution.
And there are "legacy systems", significant fraction of which runs
proprietary Unices. There are people supporting them, money
is spent, etc. So there is practial use, and as long as
there are people spending effort to make programs run on
them it makes sense to keep support in configure and
similar tools.
I understand commercial attitude, where loosing %1 of the maket
is normally deemed not worth of developement cost of supporting
rare system. And related attitude that unknow bugs do not
matter (=> only bugs that customers report are worth fixing).
For free software situation is different: specific support for
rare systems is done by people using those system. The "cost"
for mainstream systems is needing more generality and possibly
using compatiblity constructs. Generality has its costs, but
usually leads to better architecture in long time. Some
compatiblity constructs do complicate code. OTOH rare systems
help in finging bugs which may be present on all systems, but
features of less usual systems can make them more visible.
-- Waldek Hebisch