byte-compiler does *not* complain about missing case or loop

byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

1. emacs -Q
2. C-x C-f hx.el  (hx.el relies on cl but doesn't require it)
3. See the error log below.

Why is the byte-compiler silent about the "missing" `case' and `loop'?

----------------------------------------------------------------


If this should be filed as a report, I will oblige. 

Compiling file /home/kjambunathan/.emacs.d/lisp/hx.el at Wed Mar  5 10:32:47 2014

In hx-export:
hx.el:85:36:Warning: `t' called as a function
hx.el:197:9:Warning: reference to free variable `for'
hx.el:197:13:Warning: reference to free variable `markers'
hx.el:197:21:Warning: reference to free variable `in'
hx.el:198:13:Warning: reference to free variable `org-element-type'
hx.el:199:13:Warning: reference to free variable `hx-element-type'
hx.el:200:9:Warning: reference to free variable `collect'

Re: byte-compiler does *not* complain about missing case or loop

[Stefan Monnier]

> Why is the byte-compiler silent about the "missing" `case' and `loop'?

AFAIK it does report them as missing, at the end of the compilation
(it can't report them right away because `loop' and `case' might be
defined as functions further down in the file).


        Stefan

Re: byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 562 bytes --]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Why is the byte-compiler silent about the "missing" `case' and `loop'?
>
> AFAIK it does report them as missing, at the end of the compilation
> (it can't report them right away because `loop' and `case' might be
> defined as functions further down in the file).

Try the attached test.el (with emacs -Q).

1. The compiler doesn't get to "the end of data" ritual.

2. How would I interpret the error?

   1:40 ends points to param line but the actual error report - the
   `marker' variable - is elsewhere.


[-- Attachment #2: test.el --]
[-- Type: application/emacs-lisp, Size: 460 bytes --]

Re: byte-compiler does *not* complain about missing case or loop

[Andreas Schwab]

Jambunathan K <kjambunathan@gmail.com> writes:

> 2. How would I interpret the error?
>
>    1:40 ends points to param line but the actual error report - the
>    `marker' variable - is elsewhere.

(types . marker) is a malformed function call.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 704 bytes --]

Andreas Schwab <schwab@linux-m68k.org> writes:

> Jambunathan K <kjambunathan@gmail.com> writes:
>
>> 2. How would I interpret the error?
>>
>>    1:40 ends points to param line but the actual error report - the
>>    `marker' variable - is elsewhere.
>
> (types . marker) is a malformed function call.

The LINE:COL position of the error report is wrong.

The presence of that defun prevents the "end of data" report from being
triggered.

FWIW, see the attached report

1. With the offending defun compiled in, the "end of data" report is
*absent*.

2. With offending defun ";;"-ed out, the "end of data" report is
*present*.

I wonder what would explain the presence/absence of "end of data"
report.


[-- Attachment #2: with-destructuring.txt --]
[-- Type: text/plain, Size: 801 bytes --]

\f
Compiling file /home/kjambunathan/.emacs.d/lisp/hx.el at Sat Mar  8 17:50:52 2014

In hx-export:
hx.el:85:36:Warning: `t' called as a function
hx.el:210:9:Warning: reference to free variable `for'
hx.el:210:13:Warning: reference to free variable `markers'
hx.el:210:21:Warning: reference to free variable `in'
hx.el:211:13:Warning: reference to free variable `org-element-type'
hx.el:212:13:Warning: reference to free variable `hx-element-type'
hx.el:213:9:Warning: reference to free variable `collect'

In hx-org-li:
hx.el:300:42:Warning: reference to free variable `for'
hx.el:300:46:Warning: reference to free variable `html'
hx.el:300:51:Warning: reference to free variable `in'
hx.el:300:70:Warning: reference to free variable `count'
hx.el:418:40:Error: Wrong type argument: sequencep, marker

[-- Attachment #3: without-destructuring.txt --]
[-- Type: text/plain, Size: 1963 bytes --]

\f
Compiling file /home/kjambunathan/.emacs.d/lisp/hx.el at Sat Mar  8 17:52:15 2014

In hx-export:
hx.el:85:36:Warning: `t' called as a function
hx.el:210:9:Warning: reference to free variable `for'
hx.el:210:13:Warning: reference to free variable `markers'
hx.el:210:21:Warning: reference to free variable `in'
hx.el:211:13:Warning: reference to free variable `org-element-type'
hx.el:212:13:Warning: reference to free variable `hx-element-type'
hx.el:213:9:Warning: reference to free variable `collect'

In hx-org-li:
hx.el:300:42:Warning: reference to free variable `for'
hx.el:300:46:Warning: reference to free variable `html'
hx.el:300:51:Warning: reference to free variable `in'
hx.el:300:70:Warning: reference to free variable `count'

In hx-export-to-file:
hx.el:477:39:Warning: reference to free variable `hx-backends-alist'
hx.el:523:13:Warning: `:constructor' called as a function
hx.el:523:38:Warning: reference to free variable `hx-create-backend'
hx.el:523:38:Warning: `:copier' called as a function
hx.el:525:3:Warning: reference to free variable `name'
hx.el:525:8:Warning: reference to free variable `parent'
hx.el:525:15:Warning: reference to free variable `transcoders'
hx.el:525:27:Warning: reference to free variable `content-selector'
hx.el:525:44:Warning: reference to free variable `post-processor'

In hx-tex-post-processor:
hx.el:683:25:Warning: reference to free variable `for'
hx.el:683:29:Warning: reference to free variable `field'
hx.el:683:35:Warning: reference to free variable `in'
hx.el:684:25:Warning: reference to free variable `always'

In end of data:
hx.el:773:1:Warning: the following functions are not known to be defined: case, t,
    hx-backend-transcoders, cdddr, caddr, remove-if,
    pp-display-expression, loop, hx-org-text-markup, h1, h2, h3, h4,
    h5, h6, ol, ul, otherwise, hx-backend-content-selector,
    hx-backend-post-processor, defstruct, hx-backend, :constructor,
    :copier, hx-create-backend, assert

Re: byte-compiler does *not* complain about missing case or loop

[Stefan Monnier]

> 1. The compiler doesn't get to "the end of data" ritual.

No, because it hits an actual error and just stops there, without even
generating a .elc file.


        Stefan

Re: byte-compiler does *not* complain about missing case or loop

[Andreas Schwab]

Jambunathan K <kjambunathan@gmail.com> writes:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>
>> Jambunathan K <kjambunathan@gmail.com> writes:
>>
>>> 2. How would I interpret the error?
>>>
>>>    1:40 ends points to param line but the actual error report - the
>>>    `marker' variable - is elsewhere.
>>
>> (types . marker) is a malformed function call.
>
> The LINE:COL position of the error report is wrong.

I wouldn't have noticed if you didn't call my attention.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> 1. The compiler doesn't get to "the end of data" ritual.
>
> No, because it hits an actual error and just stops there, without even
> generating a .elc file.

Should it stop just there?

It can pretend as though the faulty defun was not there in first place
and go through the motions (without generating the .elc file).  It would
help if the compiler dumps an "end of data" report.

Re: byte-compiler does *not* complain about missing case or loop

[Stefan Monnier]

>>> 1. The compiler doesn't get to "the end of data" ritual.
>> No, because it hits an actual error and just stops there, without even
>> generating a .elc file.
> Should it stop just there?

You mean, should it have better "error recovery"?  I'd agree
that it would be better if it could recover in such a case.

But I think I'd only accept a patch for it if it can be implemented in
a simple/clean way, because the benefit is fairly small.


        Stefan

Re: byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> "error recovery"?

Did you notice the *wrong* LINE:COL reporting?  Does that fall under
"error recovery" as well.

http://lists.gnu.org/archive/html/emacs-devel/2014-03/msg00269.html

Re: byte-compiler does *not* complain about missing case or loop

[Stefan Monnier]

>> "error recovery"?
> Did you notice the *wrong* LINE:COL reporting?

Didn't you notice that *all* LINE:COL data is "approximate"
in the byte-compiler's output?


        Stefan

Re: byte-compiler does *not* complain about missing case or loop

[Jambunathan K]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>>> "error recovery"?
>> Did you notice the *wrong* LINE:COL reporting?
>
> Didn't you notice that *all* LINE:COL data is "approximate"
> in the byte-compiler's output?

This is an excuse I can allow :-).