This page reflects errors that were found in the 4th edition of the book. If you found an error that is not listed here, please send me a note at welch@acm.org with the word "book" in the subject and mention that you did not see it here.
Special thanks to the readers noted in italics that reported these errors.
Last update Mon Aug 18 20:21:30 PDT 2014
| pli | 3rd bullet point in the Why Tcl? section should say
"available on the Internet" instead of "available on the network".
Sean Mitchell |
| plx | 4th Edition Thanks section, 2nd paragraph. Third word
from end of paragraph should be "and" not "an".
Gary Roth |
| p37 | Example 3-3 botches the first line of output. It should be puts "Content-Type: text/html"Marty Leisner |
| p42 | The first paragraph should say
"loop variable name"
instead of
"loop variable x"
Pierre-Marie Le Biot |
| p50 | The string first description should say
that -1 is returned if subString is not found (not string) |
| p52 | The text next to the Hot Tip may be confusing.
Kevin suggests this, which reads pretty well:
"Despite the quotes, the expression operators that work on numbers and strings first try converting items to numbers if possible, and then the operators convert them back if they detect a case of string comparison."I would add that it helps to know that Tcl represents values internally as both a string and some other native value, like a double-precision floating point, or 32-bit integer, or Tcl list, or byte code, or binary blob. Tcl Math expressions compute values that are native double-precision values, and are only converted to strings for printing. The Hot Tip is reminding us that there are some automatic conversions between native representations and strings, and in some subtle cases those conversions might suprise you. Tcl_Obj values are described in Chapter 47 on C programming. Kevin Jennings |
| p55 | The caption for Example 4-4 says "Microsoft World"
instead of "Microsoft Word".
Marcin Siebert |
| p57 | Near the bottom of the page, the format %#08x should result in:
0x00000abecause the field width of 8 includes the leading 0x. Hemang Lavana |
| p65 | Example 5-1. The last line of the example has an
extra }. It should be:
=> {{1 2} {a b} {$foo}} {1 2}
Simon Klempert |
| p78 | There is a reference to regular expressions and
Chapter 11 with a stray "page" in it. The closest page reference would
be to page 145, Regular Expression Syntax.
Steve Munnings |
| p84 | Example 6-15. The paragraph after this example
says that the catch command is not grouped with
curly braces, but it is. (It wasn't in earlier editions.)
It turns out you can get away without the
curly braces, as in:
if [catch { blah } result ] { # handle error }
But it is better style and slightly more efficient
to always enclose the if expression in braces.
Simon Klempert |
| p89 | The cross reference for namespaces should
be Chapter 14, not Chapter 7.
Simon Klempert |
| p99 | The second-to-last line says "Example 5-8 uses on
array", which should be "Example 5-8 uses one array."
Yang Wang |
| p100 | In the last paragraph, the word implement should be implemented
Damien Lynch |
| p102 | The first paragraph says "Example 8-10 defines RecordInsert"
but it should say it defines RecordAppend.
Yang Wang |
| p109 | Table 9-2. The file volumes command
does not take a name argument.
Marty Leisner |
| p123 | Example 9-11. The second foreach
erroneously has curly braces around its [glob -nocomplain *] list value argument.
Alex Brown |
| p133 | Example 10-2 has a typo in the rename
command, which should be rename $procName $procName-orig |
| p136 | In the middle of the page, after
PackedButton .save "Save" { Save $file } {-side left}
The next two in-line examples should reference .save
instead of .new. E.g., they should be:
button .save -text Save -command { Save $file }
pack .save -side left
Yang Wang |
| p150 | In the last paragraph, the in-line reference
to the pattern [[:<:]] is missing a right bracket.
Mike Schincariol |
| p162 | In Table 11-7 the pattern for hexadecimal numbers is incorrect.
It includes the character ranges a-hA-H, which of course should
be a-fA-F. This same bug also appears on p164.
Andrew Ching |
| p174 | In the first paragraph of the Packages Implemented in C Code
section, there is an incorrect reference to Chapter 37. It should refer to Chapter 47.
Damien Lynch |
| p191 | In the paragraph before Example 13-6, the last sentence
is missing an "and":
Chapter 10 describes callbacks and why the curly braces are used with eval as they are in this example: Steve Munnings |
| p211-212 |
The in-line examples for namespace code return a new command
that begins with ::namespace inscope. The
book omits the leading ::, which is present to ensure
the command is not ambiguous. If you evaluated that command in
a namespace that
(foolishly) had its own namespace command, you
would not want to use that redefined namespace command.
Marty Leisner |
| p221 |
In Conversions Between Encodings, the gb12345 encoding
is incorrectly referred to as the Big5 encoding.
Those are in fact two different encodings. See
www.ascc.net
Andrew Tannenbaum |
| p240 |
In Example 17-1 it is safer to cancel the after event when the socket becomes writeable,
otherwise it could be possible for the after event to run after the fileevent and
incorrectly change the value of the connected variable.
set ::afterid [after $timeout {set connected timeout}]
set sock [socket -async $host $port]
fileevent $sock w {after cancel $::afterid ; set connected ok}
vwait connected
Zoran Slavulj
|
| p265 |
In Example 18-5 the call to Mtype_Add has its arguments reversed. It should read:
Mtype_Add .junk application/myjunkRicardo Jorge |
| p275 |
In Example 18-16 the two uses of html::varEmpty should really be ncgi::empty
Hank Liu |
| p276 |
In Example 18-17 there is an extra double-quote (") after addr2
in the arguments to Form_Simple
Ricardo Jorge |
| p278 |
In Table 18-2 the procedure name Url_PrefixInstall is misspelled.
Andy Goth |
| p355 |
Example 22-5 is missing the $ in the second reference to $starkit::topdir
Rainer Schwarzinger |
| p355 |
Example 22-6 doesn't create the "data" file. One way to do that would be
echo "Hello, World!" > write.vfs/dataRainer Schwarzinger |
| p378 | Example 24-1 does not work as written
on Windows because of limitations in pipes on that platform.
Here is a version that works on Windows
by using the cmd command processor. Note, however,
that output is still batched up and so the fileevent is
not really necessary.
Ben Ginsberg |
| p385 | The CD-ROM version of Example 24-3
has an out of date cross reference to the Scrolled_Text
procedure in a comment. The cross reference in
the printed example is correct.
Marty Leisner |
| p393 | The screen shots in examples
24-5, 24-6, 24-7 are not updated, so they incorrectly
refer to Example 31-3, which is now really Example 37-3.
Marty Leisner |
| p401 | The second sentence in the first paragraph
should read:
"Due to other constraints the request might not be honored, ..." Mike Beaufort |
| p412-414 | The screen shots in this chapter were
taken on Windows where wish has a white default background.
If you are on a Unix platform you will need to preface the
examples with . config -bg white Marty Leisner |
| p415 | Example 26-7 could usefully
cross reference the numeric color specifications shown
on page 622.
Marty Leisner |
| p416 | Example 26-9 is supposed to be
tried interactively so you can see the effect of each of
the three grid commands. If you run the whole example you'll
end up with the appearance in the lower right figure.
In addition, the text of the labels should be "left side"
and "right side".
Marty Leisner |
| p418 | Example 26-11 is missing from the
CD-ROM archive.
Marty Leisner |
| p500 | The CD-ROM version of Example 33-1
loads the "/etc/passwd" file, while the printed example
loads the sample Tk file "colors.tcl".
Marty Leisner |
| p502 | The CD-ROM version of Example 33-3
It is missing the three toplevel commands at the end of the example.
Marty Leisner |
| p509 | Table 34-1 does not list %S,
which is "The text string being inserted/deleted, if any, {} otherwise."
Marty Leisner |
| p510 | The paragraph before Example 34-3 should say
"Example 34-3 demonstrates..."
Marty Leisner |
| p558 | At the top of the page, the tk scale command should be tk scaling
Andrew Tannenbaum |
| p604 | The bindings for Tab and Shift-Tab are
not correct because tk_focusNext and tk_focusPrev don't actually
set the focus (they did in an early Tk 4.0 beta...). The current
bindings are now:
bind all <Tab> {tkTabToWindow [tk_focusNext %W]}
bind all <Shift-Tab> {tkTabToWindow [tk_focusPrev %W]}
Andrew Tannenbaum |
| p606 | Using both Example 39-1 and 39-2 together reveals an
inconsistency in the value of the variable used to wait. In 39-1 it is set
to "cancel", but in 39-2 it expect it to be either "0" or "1". The simplest
fix is to change Dialog_Wait in Example 39-1 to contain:bind $topAndreas Brodmann |
| p660 | In the first paragraph
under the Icons section, the state
operation returns iconic, not "iconified"
Ron Hart |
| p632 | On Windows,
the no cursor should result
in no cursor being displayed. However, due to a bug in Tk 8.3 and 8.4,
the cursor is a circle with a slash through it.
Roy Terry |
| p689 | Example 46-7: The Bind_Save procedure takes args
but is never passed anything. Instead, it should use the global
bind(class) value.
proc Bind_Save { dotfile } {
global bind
set out [open $dotfile.new w]
foreach w $bind(class) {
# etc
}
}
Greg Edwards
|
| p701 | In Example 47-2, the sprintf call specifies buf as its first argument. That should be buffer.
Art Morel |
| p765 | In Example 49-15 the ClockEventProc should cancel the timer event associated with the
TICKING flag in the case where it schedules the ClockDisplay. This avoids queuing up lots and lots of extra
calls to the Display procedure.
redraw:
if ((clockPtr->tkwin != NULL) &&
!(clockPtr->flags & REDRAW_PENDING)) {
if (clockPtr->flags & TICKING) {
Tk_DeleteTimerHandler(clockPtr->token);
clockPtr->flags &= ~TICKING;
}
Tk_DoWhenIdle(ClockDisplay, (ClientData) clockPtr);
clockPtr->flags |= REDRAW_PENDING;
}
Dave Wolff
|
| p837 | There are no entries for < or > under Symbols. These are used
as event delimiters on page 439.
Andrew Tannenbaum |
| p843 | The clock command is missing from the index,
nor is it listed as a sub-entry under the Tcl Command index entry.
However, clock is listed in the command index on
page 22, and is described starting on page 183. |
| p838 | Entry for "arrow on canvas" should reference page 572.
RonHart |
| p849 | Entry for "file: is plain" is really about "file: isfile. This confusion shows up for the "plain file" cross reference, which should really be a "see also" reference.
RonHart |
| p856 | Entry for "lsearch, Tcl command" should reference page 69.
RonHart |