summaryrefslogtreecommitdiff
path: root/info/impatient/errors.tex
blob: f2e59773363c7b82bffeae6dacb763aec8f5f969 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
% This is part of the book TeX for the Impatient.
% Copyright (C) 2003 Paul W. Abrahams, Kathryn A. Hargreaves, Karl Berry.
% See file fdl.tex for copying conditions.

\input macros
\chapter{Making sense \linebreak of error messages}

\chapterdef{errors}

\codefuzz = 4pc % for this chapter only (in scope of \chapter)
\bix^^{error messages}
Interpreting \TeX's error messages can sometimes 
be like going to your physician
with a complaint that you're feeling fatigued and being handed,
in response,
a breakdown of your blood chemistry.  The explanation of your distress
is probably there, but it's not easy to figure out what it is.
A few simple rules will go a long way in helping you to understand \TeX's
error messages and get the most benefit from them.

Your first goal should be
to understand what you did that caused \TeX\ to complain.
Your second goal (if you're working interactively)
should be to catch as many errors as you can in a single run.

Let's look at an example.  Suppose that your input
contains the line:
\csdisplay
We skip \quid a little bit.
|
You meant to type `|\quad|', but you typed `|\quid|' instead.
Here's what you'll get from \TeX\ in response:
\csdisplay
!! Undefined control sequence.
l.291 We skip \quid
                   a little bit.
?
|
This message will appear both at your terminal and in your log file.
^^{log file//error messages}
The first line, which always starts with an exclamation point (|!!|),
tells you what the problem is.  The last two lines before the `|?|' prompt
(which in this case are also
the next two lines) tell you how far \TeX\ has gotten
when it found the error.  
It found the error on line $291$ of the current input file,
and the break between the two message lines indicates
\TeX's precise position within line $291$, namely, just after |\quid|.
The current input file is the one just after the most recent unclosed
left parenthesis in the terminal output of your run (see \xref{infiles}).

This particular error, an undefined control
sequence, is one of the most common ones you can get.
If you respond to the prompt with another~`|?|',
\TeX\ will display the following message:
{\hfuzz = 2in
\csdisplay
Type <return> to proceed, S to scroll future error messages,
R to run without stopping, Q to run quietly,
I to insert something, E to edit your file,
1 or ... or 9 to ignore the next 1 to 9 tokens of input,
H for help, X to quit.
|
}%
Here's what these alternatives mean:
\ulist
\li If you type \asciichar{return},
\TeX\ will continue processing your document.
In this case it will just ignore the |\quid|.
\li If you type `|S|' (or `|s|'---uppercase and lowercase are equivalent
here), \TeX\ will process your document without stopping \emph{except}
if it encounters a missing file.  Error messages will still appear at your
terminal and in the log file.
\li If you type `|R|' or `|r|',
you'll get the same effect as `|S|' except that
\TeX\ won't even stop for missing files.
\li If you type `|Q|' or `|q|',
\TeX\ will continue processing your document but
will neither stop for errors nor display them at your terminal.  The
errors will still show up in the log file.
\li If you type `|X|' or `|x|',
\TeX\ will clean up as best it can, discard the page
it's working on, and quit.  You can still print or view the
pages that \TeX\ has already processed. 
\li If you type `|E|' or `|e|', \TeX\ will clean up and terminate
as it would for `|X|' or `|x|' and then enter your text editor,
positioning you at the erroneous line.
(Not all systems support this option.)
\li If you type `|H|' or `|h|',
you'll get a further explanation of the error displayed
at your terminal and
possibly some advice about what to do about it.  This explanation will also
appear in your log file.  For the undefined control sequence above,
you'll get:
{\hfuzz = 2in
\csdisplay
The control sequence at the end of the top line
of your error message was never \def'ed. If you have
misspelled it (e.g., `\hobx'), type `I' and the correct
spelling (e.g., `I\hbox'). Otherwise just continue,
and I'll forget about whatever was undefined.
|
}
\li If you type `|?|', you'll get this same message again.
\endulist
\noindent
The other two alternatives, typing `|I|' or a small integer, provide ways of
getting \TeX\ back on the track so that your error won't cause further errors
later in your document:
\ulist
\li If you type `|I|' or `|i|'
followed by some text, then \TeX\ will insert
that text as though it had occurred just after the point of the error,
at the innermost level where \TeX\ is working.
In the case of the example above, that means at \TeX's
position in your original input, namely, just after `|\quid|'.
Later you'll see an example that shows
the difference between inserting something at the innermost level and
inserting it into your original input.  In the example above of the undefined
control sequence, if you type:
\csdisplay
I\quad
|
\TeX\ will carry out the |\quad| command and produce a quad space 
where you intended to have one.
\li If you type a positive integer less than $100$ (not less than $10$ as the
message misleadingly suggests),
\TeX\ will delete that number of tokens from the innermost
level where it is working.
(If you type an integer greater than or equal to $100$, \TeX\ will
delete $10$ tokens!)
\endulist

Here's an example of another common error:
\csdisplay
Skip across \hskip 3cn by 3 centimeters.
|
The error message for this is:
\csdisplay
!! Illegal unit of measure (pt inserted).
<to be read again> 
                   c
<to be read again> 
                   n
l.340 Skip across \hskip 3cn
                             by 3 centimeters.
|
^^|<to be read again>|
In this case \TeX\ has observed that `|3|' is followed by something that
isn't a proper unit of measure, and so it's assumed the unit of measure to be
points.  \TeX\ will read the tokens of `|cn|' again and insert them into
your input, which is not what you want.  In this case you can get a better
result by first typing `|2|' to bypass the `|cn|'.
You'll get the message:
\csdisplay
<recently read> n
                 
l.340 Skip across \hskip 3cn
                           by 3 centimeters.
|

Now you can type
`|I\hskip 3cm|' to get the skip you wanted (in addition to the |3pt| skip
that you've already gotten).\footnote
{By typing `|I\unskip\hskip 3cm|' you can get rid of the |3pt| skip.}

If you type something that's only valid in math mode, \TeX\ will switch over
to math mode for you whether or not that's what you really wanted.
For example:
\csdisplay
So \spadesuit s are trumps.
|
Here's \TeX's error message:
\csdisplay
!! Missing $ inserted.
<inserted text> 
                $
<to be read again> 
                   \spadesuit 
l.330 So \spadesuit
                  s are trumps.
|
Since the |\spadesuit| symbol is only allowed in math mode,
\TeX\ has inserted a `|$|' in front of it.
After \TeX\ inserts a token, it's positioned in \emph{front} of
that token, in this case the `|$|', ready to read it.
Typing `|2|' will cause \TeX\ to skip both the `|$|'
and the `|\spadesuit|' tokens, leaving it ready to process the `|s|'
in `|s are trumps.|'.
(If you just let \TeX\ continue, it will typeset `|s are trumps|'
in math mode.)

Here's an example where \TeX's error diagnostic is downright wrong:
\csdisplay
\hbox{One \vskip 1in two.}
|
The error message is:
\csdisplay
!! Missing } inserted.
<inserted text> 
                }
<to be read again> 
                   \vskip 
l.29 \hbox{One \vskip
                     1in two.}
|
^^|<inserted text>|
The problem is that you can't use |\vskip| when \TeX\ is in 
restricted horizontal mode, i.e, constructing an hbox.
But instead of rejecting the |\vskip|, \TeX\ has inserted a right brace
in front of it in an attempt to close out the hbox.
If you accept \TeX's correction, \TeX\ will complain again
when it gets to the correct right brace later on. It will also complain
about anything before that right brace that isn't allowed in vertical mode.
These additional complaints will be particularly confusing
because the errors they indicate are bogus, a
result of the propagated effects of the inappropriate
insertion of the right brace.
Your best bet is to type `|5|', skipping past all the tokens
in `|}\vskip 1in|'.

Here's a similar example in which the error message is longer
than any we've seen so far:
\csdisplay
\leftline{Skip \smallskip a little further.} But no more.
|
The mistake here is that |\smallskip| only works in a vertical mode.
The error message is something like:
\csdisplay
!! Missing } inserted.
<inserted text> 
                }
<to be read again> 
                   \vskip 
\smallskip ->\vskip 
                    \smallskipamount 
<argument> Skip \smallskip 
                           a little further.
\leftline #1->\line {#1
                       \hss }
l.93 ...Skip \smallskip a little further.}
                                           But no more.
|
The error messages here give you a tour through the macros that are used in
\plainTeX's implementation of |\leftline|---macros
that you probably don't care about.
The first line tells you that \TeX\ intends to cure the problem by
inserting a right brace.
\TeX\ hasn't actually read the
right brace yet, so you can delete it if you choose to.
Each component of the message after the first line
(the one with the `!') occupies a pair of lines.
Here's what the successive pairs of lines mean:
\olist
\li The first pair indicates that \TeX\ has inserted, but not yet read,
a right brace.
\li The next pair indicates that after reading the right brace, \TeX\ will
again read a `|\vskip|' command (gotten from the macro definition
of |\smallskip|).
\li The third pair indicates that \TeX\ was expanding the |\smallskip|
macro when it found the error.  The pair also displays the definition of
|\smallskip| and indicates how far \TeX\ has gotten in expanding and
executing that definition.  Specifically, it's just attempted
unsuccessfully to execute the |\vskip| command.  In general, a
diagnostic line that starts with a control sequence followed by `|->|'
indicates that \TeX\ has been expanding and executing a macro by that
name.
\li The fourth pair indicates that \TeX\ was processing a macro argument
when it found the |\smallskip| and also indicates \TeX's position in that
argument, i.e., it's just processed the |\smallskip| (unsuccessfully).
By looking ahead to the next pair of lines
we can see that the argument was passed to
|\leftline|.
\li The fifth pair indicates that \TeX\ was expanding
the |\leftline|
macro when it found the error.
(In this example the error occurred while \TeX\ was 
in the middle of interpreting several
macro definitions at different levels of expansion.)
Its position after |#1| indicates that the last thing it saw was the
first (and in this case the only) argument to |\leftline|.
\li The last pair indicates where \TeX\ is positioned in your input file.
Note that this position is well beyond the position where it's
inserting the right brace and reading `|\vskip|' again.
That's because \TeX\ has already read the entire argument to |\leftline|
from your input file, even though it's only processed part of that argument.
The dots at the beginning of the pair indicate a preceding part of the
input line that isn't shown.  This preceding part, in fact, includes
the |\leftline| control sequence that made the |\vskip| illegal.
\endolist
\noindent
In a long message like this, you'll generally find only the first line and the
last pair of lines to be useful; but it sometimes helps to know what the other
lines are about.  Any text that you insert or delete will be 
inserted or deleted at the
innermost level.  In this example the insertion or deletion would occur
just before the
inserted right brace.
Note in particular that in this case \TeX\ puts any text you might insert 
\emph{not} into your input text but into a macro definition
several levels down. (The original macro definition is of course not modified.)

You can use the ^|\errorcontextlines| command \ctsref{\errorcontextlines}
to limit the
number of pairs of error context lines that \TeX\ produces.
If you're not interested in all the information that \TeX\ is giving you,
you can set |\error!-contextlines| to $0$. That will give you just the first
and last pairs of lines.

Finally, 
we'll mention two other indicators that can appear at the start of a pair
of error message lines:
\ulist
\li ^|<output>| indicates that \TeX\ was in the middle of its output routine
when this error occurred.
\li ^|<write>| indicates that \TeX\ was in the middle of executing a
|\write| command when this error occurred.
\TeX\ will detect such an error when it is actually doing the |\write|
(during a |\shipout|), rather than when it first encounters the |\write|.
\endulist


\eix^^{error messages}
\endchapter
\byebye