Problem transfering files with brackets in filename

[Mark-Five]

Platform : Win10 x64
Java : 8/181 x64 (latest)
Using ADTPro release 2.0.3 on apple //e with SSC
Filename : SHERWOOD FOREST (1982)(PHOENIX SOFTWARE).dsk

Content of trace window :
Exception in thread "Thread-2" java.util.regex.PatternSyntaxException: Unclosed group near index 40
SHERWOOD FOREST (1982)(PHOENIX SOFTWAR..
at java.util.regex.Pattern.error(Unknown Source)
at java.util.regex.Pattern.accept(Unknown Source)
at java.util.regex.Pattern.group0(Unknown Source)
at java.util.regex.Pattern.sequence(Unknown Source)
at java.util.regex.Pattern.expr(Unknown Source)
at java.util.regex.Pattern.compile(Unknown Source)
at java.util.regex.Pattern.(Unknown Source)
at java.util.regex.Pattern.compile(Unknown Source)
at java.util.regex.Pattern.matches(Unknown Source)
at java.lang.String.matches(Unknown Source)
at org.adtpro.gui.Gui$1.accept(Unknown Source)
at java.io.File.list(Unknown Source)
at org.adtpro.gui.Gui.getFiles(Unknown Source)
at org.adtpro.CommsThread.sendDirectoryWide(Unknown Source)
at org.adtpro.CommsThread.dispatchCommand(Unknown Source)
at org.adtpro.CommsThread.commandLoop(Unknown Source)
at org.adtpro.CommsThread.run(Unknown Source)

Renaming the file (removal of brackets) solves the problem.

[david-schmidt]

Good catch. Is this because you had selected the file via wildcard on the client end, or did you type it all out?

[Mark-Five]

Pressed R for receive, then RETURN.
Then selected the file using arrow keys and RETURN again.
Funny thing is that it's starting something on the serial port.
After that, I have to disconnect/reconnect (on the PC software) to regain access to ADTPro.

[david-schmidt]

Got it, this is via the arrow/return interface. Another thing is also true - if the file (name) is actually longer than the screen, it will simply fail to find the file (since the file name is passed as-is-viewed to the server).

[Mark-Five]

Yep. Also noticed this one but thought the filename was just too long.
I can help on the 6502 part but I suck at java.

[david-schmidt]

The Java side is trivial. But more interesting might be to send a page+offset index to the file rather than screen-scraping as it does today. That would solve both problems.

[Mark-Five]

That would sure be a great enhancement.
At least for me since I'm far too lazy to type a filename :-)

[david-schmidt]

Notes:
In hostfns.asm: NIBPCNT holds the current page number, NDULASTROW is current vertical position on screen; it will be 3-indexed (first row is actually 4th on screen due to header).

In conio.asm: SCRAPE: can be retrained to encode the page and row; a payload like 0x01<page_byte><vertical_byte>0x00 or something like that.

Protocol extension to Query File Size and Image Get will understand the leading 0x01 in the payload to mean this coordinate system rather than a text name. Reply from the host can remain unchanged.

[david-schmidt]

Merged - will be part of next release!