@RN_
@T3s|4_
@orignal
@postman
%acetone
%mareki2p
+NiceBoat
+Onn4l7h
+Over
+altec
+nyaa2pguy
+poriori
+profetikla
+qend-irc2p
+r00tobo
+uop23ip
+waffles
Arch
FreefallHeavens
Gully
Irc2PGuest16019
Irc2PGuest20377
Irc2PGuest33925
Irc2PGuest77921
Liorar_
MatrixBot
Onn4|7h
StormyCloud
T3s|4
Teeed
WaAmH
ahiru
anontor2
cims
dr|z3d
floatyfloatilla
mahlay
makoto
n2_
nZDoYBkF
nilbog
not_bob_afk2
o3d3
r00tobo[2]
rednode
solidx66
stormycloud[m]
vivid_reader56
zzz
cumlord
idk if true or not but is this your theory on irc d/c?
cumlord
I was running a test against ilita/postman with different routers to count the d/cs but accidentally cut it short
not_bob
Well then.
orignal
mareki2p I have replaced translations to string_view
orignal
thank you for the hint
mareki2p
Nice. In theory, there should be much less/fewer dynamic memory allocations on program startup copying data from read-only memory to read-write memory. Now, the data stays in read-only memory. Meaning improving both memory consumption and CPU time consumptions. Do you notice any such improvements in practice?
orignal
I don't think so
orignal
it might be noticable improvement
orignal
however it's always better to use string_view instead string
mareki2p
I could try to sort the strings in translations at compile-time. Replacing the dynamic std::map by static std::array + binary search.
zzz
or just switch to po's and libgettext?
orignal
why is it better?
zzz
don't know that it is. but maybe.
orignal
that's not my busines anyway