NewsForumsCodeLogsGallerySpecsSearch

current github issues

BlitzMax Forums/BlitzMax Programming/current github issues

skidracer(Posted 2016) [#1]
As a contributor to the official blitz-research blitzmax repository on github I have been looking at a few minor tweaks such as improved linux audio support.

Before this however I would like to consider a few standards / normalisations.

1. I already have suffered multiple skull fractures due to end of line madness possible in git. Best coverage is < CR >< LF > pairs.

2. UTF16 encoding. What were we on? I am willing to modify MaxIDE to save UTF8,

2a Inclusion of BOM will touch everything....

2b How does non-ansii source affect legacy C tool chains?

For starters, I plan to count number of files of all file types that are non-ansi.


dw817(Posted 2016) [#2]
What I read was modify MaxIDE to allow UTF8. This will certainly make the foreign language programmers quite happy !

As for myself, does this mean that characters 128-255 will be read back as characters 128-255 and not hashes of UTF8 as it is now ?


skidracer(Posted 2016) [#3]
MaxIDE already support unicode using UTF-16 so no change in happiness of non-ansii coders is expected.


As for myself, does this mean that characters 128-255 will be read back as characters 128-255 and not hashes of UTF8 as it is now ?



I have no idea what you are asking. MaxIDE switches to UTF16 to save .bmx as soon as you add an emoji or what have you to source as it uses standard BlitzMax SaveText function.

With other file types MaxIDE should respect the encoding but probably doesn't.


dw817(Posted 2016) [#4]
Sorry, I should clarify. Where you can have characters (single-byte) in your source-code in the IDE, say for instance in a DEFDATA statement where within quotes you have characters that by ASCII code are larger than 127.

So it would be possible to store characters 32-255 within a DEFDATA statement, with of course the exception of ~ and quote "

Here, run this code and scroll up to see the 4-digit returns to see what I am talking about. One of the real advantages of source code to show a situation or solution. :)
Strict
Local fp:TStream=WriteStream("ins.txt"),t$="defdata ~q",i,n=32
For i=32 To 255
  If i<>34 And i<>126 Then t$:+Chr$(i)
Next
WriteLine fp,t$+"~q"
CloseStream fp

ReadData t$
For i=1 To Len(t$)
  If i=3 Or i=94 Then n:+1
  Print n+" "+Asc(Mid$(t$,i))
  n:+1
Next

DefData " !#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}"



Derron(Posted 2016) [#5]
For improvements have a look at github.com/maxmods

@utf16 and maxide

Maxide does not play nicely with non-bom utf8. it corrupts content.

Next to the utf trouble the broken undo functionality is of a big concern too (it is unuseable for me except starting a debug build so see where something crashes). Better use scintilla as textarea replacement..it handles undo...and surely plays nicely with utf8/16.


Nice to see someone "official" wanting to make his hands dirty.


Btw maxmods-repo contains adjustments for linux audio.

Bye
Ron


Derron(Posted 2016) [#6]
Regarding MaxIDE and utf8 check this post of mine.


@coverage of LF/CR
You either do convert them all one time, or you ignore line ending adjustments when committing your changes (similar things are possible for whitespace / tab conversions).


bye
Ron