Le 24/07/2023 à 15:45, Pierre Gruet a écrit :
Hi again,
Le 24/07/2023 à 14:15, Sylvestre Ledru a écrit :
Le 24/07/2023 à 14:07, Pierre Gruet a écrit :
Hi Sylvestre,
[...]
As it is changing the behavior of the language, I feel that it is a
bit risky to do such changes this way.
It might break scilab script between the upstream implementation and
the Debian/Ubuntu packages.
For example:
-[Mss, comments] = csvRead(filename, ";" , [] , "string" , [] ,"/foo
bar/");
+[Mss, comments] = csvRead(filename, ";" , [] , "string" , [] ,"foo
bar");
Maybe you could automatically remove the trailing and ending "/"
before passing them to PCRE, no?
Thanks again for looking at the patch.
What you suggest is to write some handle_pattern function taking
patterns with delimiters and returning them without the delimiters and
possibly without the backslashes escaping the delimiter character
inside the pattern, right?
Also, logically following would be looking at modifiers after the
final delimiter and translating them into the right options for
compiling the pattern?
yeah, in the gateway, removing / / on the fly when processing the input
arguments
Do you think one needs to account for modifiers othen than "imx"?
Dunno, I don't know how often they are used and how :(
Cheers
S
