Re: x5250 version 0.4.6 released -- LINUX5250

On Sun, 5 Dec 2004, Scott Klement wrote:

a) "typeahead" (typing keys ahead of time while the system is busy.)
doesn't work properly, the keys appear to get into the keyboard buffer,
but do not get printed on the screen when they finally do get displayed.

Oh yes, thanks for reminding me. The keystrokes do get into the buffer (pressing enter after WRKACTJOB in your example should do the right thing) but the charaters do not get drawn. I've added this to the list.

b) It's not possible to use the left-control key as the error reset key.
  This should be the default, as it's what pretty much everyone is
  used to.

I'm thinking of leaving key map related issues until after 0.5. I recognize that tn5250 has some keymapping abilities, but I'd rather address these issues later.

c) When you display something in reverse-image, the characters get
  reversed, but the underlines below them do not.  They should also
  be reversed.  When you have several lines of reverse-image on the
  screen, the reverse imaged spaces should look like a continuous area.
  Right now there are a few lines of black space in between them.

Are you saying that underlined fields do not get reversed when blank? Or that underlines should not be drawn in reverse fields? I think the "few lines of black space in between" reversed fields are related to your next item below.

d) There are extra/unnecessary blank pixels between each line of text on
  the screen.  Compare an xterm screen with the same font to an x5250
  screen, note that the x5250 screen spaces the lines further apart.
  This is a waste of screen real-estate.  Makes it more difficult to
  multitask because windows don't all fit ont he screen at once.  (And
  really bugs the heck out of me.)

Hmm... not sure how to respond here. I put them in on purpose. The squished look of xt5250 really bugs the heck out of me! :) Plus, it is designed so that no part of any character ever touches the underline below, hopefully improving readability. This is actually one of the parts of x5250 I like quite a bit. Maybe a resource is called for here.

e) The TN5250 colsep_style config keyword should be implemented so that
  matches what the Win32 terminal has.

  This keyword controls the way that column separators are displayed,
  and has the following possible values:
     colsep_style=none
     colsep_style=dots
     colsep_style=full

There is already an X resource for this. But I guess I need to get this from .tn5250rc as well. I didn't even realize such a keyword existed for tn5250. I guess only win32 uses this keyword?

f) It does not appear that the +local_print_key feature of TN5250 is
  implemented in x5250.  In tn5250 (or xt5250) you can specify this
  keyword, and several other tunables, to have the emulator generate
  a PostScript file with the screen image in it, and then direct that
  file to lpr or whatever.

  In Windows, this same feature exists, but instead of using PostScript
  it uses Windows GDI calls to send the print screen to your printer.
  This is why this print screen support was put into the terminal instead
  of into lib5250.

  Though, now that we've got tn5250, gtk-5250 and x5250, maybe it maeks
  sense to put the postscript routines in lib5250 (specifically,
  utility.c?) so that one copy of the code can be used for all of these
  terminal types.

  In any case, this needs to be implemented in x5250 if you want it to
  be feature-compatible.

I think that lib5250 is the place for this. I don't know how win32 works, but I suppose a win32 specific method could be called in place of the postscript method? That way, unless the terminal registers a local print method, lib5250 would simply run the default postscript local print function.

Most of these things should be relatively simple to fix. (I hope?)


I hope so too :)

James Rich

It's not the software that's free; it's you.
        - billyskank on Groklaw