[Scilab-users] Regex problem

[Samuel Gougeon]

Le 19/08/2016 13:26, Gerhard Kreuzer a écrit :
>
> Hi,
>
> I have a data file containing metadata and binary data. I successfully 
> scanned the file with .net Regex class, now I want to scan it with 
> SciLab, but ….
>
> The relevant part of  the file looks like:  #data# .. here comes 
> binary data … #EOC#  .. here comes binary data …  #EOC# .. and so on.
>
> My regex in .net notation looks like: #data#((?<data>(?s:.*?))#EOC#)+
> Ok, in SciLab the notation is little different: 
> '/#data#((?P<data>(?s:.*?))#EOC#)+/'
> This regex expression didn’t match at all. I start experimenting and 
> it looks like that the regex engine stops at a new line character 
> (0x0A) which i spart of  the binary data block. As far as I know (and 
> that isn’t that far) the clause (?s:.*?)means, take any character 
> until you find #EOC# but as least as possible.
> No interpretation on 0x0A ….
> Any idea how I can parse my file and get the binary data blocks into 
> variables, or at least get pointers to  the starting points, so I am 
> able to read the binary data with some file read function?

Here is a working example supporting ascii(10). It looks tha "ms" 
modifiers must be used together:
--> s = "abcd" + ascii(10) + "efghijkClmnop" + ascii(10) + "fg hiJkl"
  s  =
  abcd
efghijkClmnop
fg hiJkl

--> [trash,trash,captures] = regexp(s, "/c.*?i/*ms*i"); captures
  captures  =
!cd
efghi      !
!              !
!Clmnop
fg hi  !

So, in your case, you may try with:

regexp(s, "/#data#(.*?)#EOC#)*/ms*");

May be the hardest thing will be to get your binary content as a string.
I don't think that Scilab's regexp() will accept anything else than a 
string.
May be there are some Scilab regexp features in reading binary files. To 
be investigated.

HTH
Samuel Gougeon

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20160819/a901f971/attachment.htm>