[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Jit cache, issue when gbs3 --nocache is specified
[Thread Prev] | [Thread Next]
- Subject: Re: Jit cache, issue when gbs3 --nocache is specified
- From: Benoît Minisini <benoit.minisini@xxxxxxxxxxxxxxxx>
- Date: Fri, 27 Jun 2025 18:55:00 +0200
- To: user@xxxxxxxxxxxxxxxxxxxxxx
Le 22/06/2025 à 20:29, Benoît Minisini a écrit :
Le 21/06/2025 à 23:10, Brian G a écrit :In gbs3 the --nocache flag causes the scripter to recreate the compiled script, but when the script is fast the jit cache can not be ignored and overrides the scripts nocache parameter.When specified as Fast a jit cache is created under the users /tmp/ gambas-<uid>/gb.jit. This is never cleared and always used. There is no way to tell gbx3 or gbr3 to not cache the jit .so files.Which in fact renders the --nocache option for scripter useless in that case. Should scripter ignore the --nocache flag in the case of Fast being used or should it print a warning?Not sure what should happen here. Add --nocache option to gbx3 or gbr3 same thingOr is nothing required? Or just drop this from the scripter itself as it really may not be useful except in testing.Why don'y you use the 'GB_JIT_FORCE' environment variable to force the recompilation of JIT code?
Did you check that?Sorry for having not yet looked at your updated scripter, I am very busy with personal stuff at the moment.
Regards, -- Benoît Minisini.
Jit cache, issue when gbs3 --nocache is specified | Brian G <brian@xxxxxxxxxxxxxxxx> |
Re: Jit cache, issue when gbs3 --nocache is specified | Benoît Minisini <benoit.minisini@xxxxxxxxxxxxxxxx> |