in reply to Re^2: PerlMonks Editor
in thread PerlMonks Editor

The line with whitespace on it (an empty line doesn't do it for me) is fairly trivial - but silly not to have found. Thanks for finding it for me!

The other issue I should have fixed before uploading, but I sort of forgot about it. On investigation it could be more important than I'd thought and has now been fixed in my current version. It probably doesn't affect current functionality, but my test suite runs against a module version of the code so I've not run the combined version against the test suite to check.

I'll upload an updated version to CPAN soon.

Thanks for the help!

Update: an updated version has been uploaded as http://cpan.perl.org/authors/id/G/GR/GRANDPA/PMEdit/PMEdit-001.000202-1.pl and should appear soonish at a mirror near you. ;)

DWIM is Perl's answer to Gödel

Replies are listed 'Best First'.
Re^4: PerlMonks Editor
by parv (Vicar) on Dec 01, 2006 at 07:09 UTC

    Hey, happy to help. But my helping you is not done yet. ;}

    Program stops after printing "bad event type" message .. . 

    bad event type or keysym "_" at PMEdit-001.000202-1.pl line 269, <DATA
+> line 81.

    [download]
    ... but that goes away (and a Tk window emerges) if I change _ to one of [a-zA-Z0-9] but not [-+].

    Other minor problems ... (1) chomp fails to clear ^M in __DATA__ ... 

    1057 sub LoadConfig {
1058     my $ok = 1;
1059
1060     while (<DATA>) {
1061         # Load the default configuration stuff
1062         chomp;
1063         next if ! length;
     ...
1067         my ($tag, $htmlTag, $name, $flagsField, @options) = split
+ /\s*,\s*/;
1068
1069         if (! defined $flagsField) {
1070             print "Missing entries in tag line ($.): $_";
     ...

    [download]
    ... which causes above print() to act; line numbers reported are 32 & 58.

    (2) Later in LoadConfig(), chomp() is rendered useless by use of s/\s+$//; ... 

    1136     while (<DATA>) {
1137         # Load key binding information
1138         next if /^#/;
1139
1140         chomp;
1141         s/^\s+//;
1142         s/\s+$//;

    [download]
    In both cases, I suggest just use s/\s+$// & chop the chomp().

    (3) There seems to spurious tab characters on lines 228, 769, 771-772, 774-775.

    diff -u output follows ...

    Read more... (2 kB)

[reply]
[d/l]
[select]

      Again thank you! The first issue is related to different key binding support due either to OS differences or Tk version differences. I've now wrapped the troublesome code in an eval to trap the error and handle it in a rather friendlier manner.

      The ^M issue is a Windows/*nix line end issue I presume and I have used the fix you suggest - the two chomps are gone.

      I'm not sure where the tab characters came from, but they have now been removed.

      I'll sit on these changes for a day or two in case anything else comes to light. ;)

      DWIM is Perl's answer to Gödel
[reply]
        The first issue is related to different key binding support due either to OS differences or Tk version differences.

        Yeah, that is really a downer, currently, that I could not use '_' in key binding. I will look into that some more, no doubt. Quite possibly that would have to do Tk installed here. I found a patch while searching for reason|solution for Tk dieing with sig 11 when compiled with Xft, which may or may not fix the this key binding issue.

        Anyway, please do ignore the change in __DATA__ related to underline key binding line. I should not have included that in the patch posted earlier.

[reply]

      Program stops after printing "bad event type" message .. .

      bad event type or keysym "_" at PMEdit-001.000202-1.pl line 269, <DATA> line 81.

      ... but that goes away (and a Tk window emerges) if I change _ to one of [a-zA-Z0-9] but not [-+].

      So after source diving, I found that the key symbol is being retrieved, in Tk-804.027/pTk/mTk/generic/tkBind.c:TkKeysymToString(), via XStringToKeysym(3X11) which uses X11/keysymdef.h which does not have any X_[-+_] entry but does have X_(underscore|plus|minus). Needless to say, after changing '_' in the underline key binding to 'underscore', Ctrl-_ worked as expected.

[reply]
[d/l]
[select]

In Section Cool Uses for Perl