[fpc-pascal] FPC_PCHAR_LENGTH dog slow

[Ryan Joseph]

I was just parsing something and found that a call to length on a pchar was taking 90+% of the time in FPC_PCHAR_LENGTH.

Why is that so crazy slow? Does it need to iterate the entire length of the string to know it’s length?

Regards,
Ryan Joseph

_______________________________________________
fpc-pascal maillist - fpc-***@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman

[Tomas Hajny]

Post by Ryan Joseph
I was just parsing something and found that a call to
length on a pchar was taking 90+% of the time in
FPC_PCHAR_LENGTH.
Why is that so crazy slow? Does it need to iterate the
entire length of the string to know it’s length?

Yes, that's indeed the difference between Pascal strings
(Short/Ansi/Wide/UnicodeString) and C-like PChars.

Tomas


_______________________________________________
fpc-pascal maillist - fpc-***@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pasca

[w***@windstream.net]

Post by Tomas Hajny

Post by Ryan Joseph
Why is that so crazy slow? Does it need to iterate the entire length of the
string to know it’s length? >

Yes, that's indeed the difference between Pascal strings
(Short/Ansi/Wide/UnicodeString) and C-like PChars.

absolutely! i truly hated having to add loops to my string handling code when
working with ""C strings"" as we used to call them back in the day... the huge
advantage of "Pascal strings" is their length being specified at the beginning
of the character array they are stored in...

i think i'm still waiting for the string "length byte" becoming a "length word"
or possibly a "length long" so the speed of pascal strings can be reacquired -=B-)

[leledumbo via fpc-pascal]

Post by w***@windstream.net
i think i'm still waiting for the string "length byte" becoming a "length word"

or possibly a "length long" so the speed of pascal strings can be reacquired
-=B-)

Ansi and any other dynamic strings already have length longint that allows
2GB string length with O(1) length retrieval.

[w***@windstream.net]

Post by leledumbo via fpc-pascal

Post by w***@windstream.net
i think i'm still waiting for the string "length byte" becoming a "length
word" or possibly a "length long" so the speed of pascal strings can be
reacquired -=B-) >

Ansi and any other dynamic strings already have length longint that allows
2GB string length with O(1) length retrieval.

excellent... i've been confused by all the various string changes that have been
made in the recent years since i joined the lists... the UTF stuff really spun
my head around with all the various discussions and arguments... i'm still not
sure i understand it... i don't need UTF but i hope that my newer programs are
written properly to handle UTF if it ever comes up... they're generally console
applications reading textual data from elsewhere and processing for local output...

[Mattias Gaertner]

On Tue, 3 Jul 2018 09:15:16 -0600

Post by Ryan Joseph
I was just parsing something and found that a call to length on a pchar was taking 90+% of the time in FPC_PCHAR_LENGTH.
Why is that so crazy slow? Does it need to iterate the entire length of the string to know it’s length?

Yes.

Mattias
_______________________________________________
fpc-pascal maillist - fpc-***@lists.freepascal.org
http://lists.freepas