[Date Prev][Date Next][Thread Prev][][Date Index][Thread Index]

Re: [emacs-w3m/emacs-w3m] Bb history display (#4)



@yamaoka : Diff'ing against 7fa00f9, I like much of what your changes accomplish. Here are the points that I have questions about (all line numbers are in file w3m.el):

  1. [line 10821] (width (- (w3m-display-width) 21))
    I had 19 instead of 21, to squeeze in more space; was it giving you trouble with wide characters or something?

  2. [lines 11005-11021] I think my version is better

2.1) There exists a variable w3m-db-history-display-size which is by default set nil, and is passed to the function as SIZE when called interactively. When a person sets that default variable as non-nil, your version does not allow the default to be changed (ie. enter a number instead of pressing return at the prompt), and more importantly, does not prompt for START, so the user loses that functionality.

2.2) It seems that you're checking for an X11 version of emacs, and in that case you are performing an extra check, and possibly prompting the user. If you really do all that, it should also be done for non-X11 emacs also?

2.3) The check seems to have a bug because it is being performed based upon START, but that argument will always be zero at that point. It should be performed after the user is prompted for START.

2.4) The check is based upon a size comparison, and seems to be in order to avoid something being to long or too large, but it isn't taking into account SIZE, or the ability of the user to begin at any START point and click on the 'prev' link to go back in the list. I don't see how this check will ever be useful. If the purpose is to avoid a long wait while a single page is being processed, then the check should be after prompting for SIZE, only if the user accepted the default of 0, and be based upon some MAX-PAGE-DISPLAY-SIZE (not max database size), and the message should be a warning to the effect "That would be a very long page (n lines). Proceed anyway?", and upon a no, exit.

2.5) What happens when the user answers no to the prompt? In that case, the use has told us not to continue, but then the function does continue! It proceeds to prompt for START and SIZE.

  1. function w3m-db-history-fix-indentation I don't know how to test, so I'll take your word for it.


You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.