Slow decoding results for binary files

[UmamiAppearance]

          Hi,

First of all, thank you! I tested and for all graphics it appears to work
now. However, it is not working in other case of uuencoded binaries. I
double checked everything (including used regexp).

In some cases it is just return 3 bytes and those bytes always representing
this uint8array [ 80, 75, 3 ]. For example the first lines when just
returning those 3 bytes are:
M4$L#!!0 ( ..@.*** +$ 0 9&]C4')O<',O87!P+GAM
M;$V./0L",1!$_\IQO;=!P4)B0-!2L+(/>QLOD&1#LD)^OCG!CVX>;QA&WPIG
MNI#BV&5(C(I(/ !47BK9.7:=N')=HI6-Y #OGD<Z,STA)8O4'J@)I9GF
@.**>K7A.YNJQ<&4GPZ4A!0WJ=0U[R;UEA_6#MI7E!+ P04
M " #CAFI6/HK7?.\ K @ $0 &1O8U!R;W!S+V-O&ULS9+/
M3L,P#(=?!>7>.FT9AZC+96BG(2$Q"<0M2KPMHOFCQ*C=V].6K1."!^ 8^Y?/
[...]

In other cases, the CPU is running hot and decoding is not stopping or it
takes a very long time but the decoding result is not correct.
Not sure how I can help to find the reason.

Am So., 21. Mai 2023 um 12:33 Uhr schrieb UmamiAppearance <
@.***>:

Give it a try now.

—
Reply to this email directly, view it on GitHub
#52 (comment),
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABI2CFKRFQIIEQBEL6647RLXHHVQVANCNFSM6AAAAAAYIAFCKI
.
You are receiving this because you authored the thread.Message ID:
@.***>

Originally posted by @ecofi in #52 (comment)

[UmamiAppearance]

The problem are big input files for decoding. The decoder has a general purpose for all the different base systems. In that process the input gets manipulated multiple times, which means the whole input gets copied multiple times before it gets decoded - this makes the process slow and CPU intensive.
When the decoder was initially designed, the idea was to use it for many multiple base systems but with short input sequences in mind. To solve the issue with bigger files it has to be redesigned. Which isn't a bad idea anyway, but is a little bit more work.

I need further investigation on this. Initial string manipulation is not the bottleneck.

[UmamiAppearance]

Actually, there was a bad array concatenation in the generic decoding function, which turned out to be the reason for slow decoding.

This was resolved in #55

@ecofi: If there are still issues for the particular case, please open a new issue and provide the problematic files/encoded strings via an external service.