[PATCH] eliminate false positive in compile.el

[PATCH] eliminate false positive in compile.el

[Daniel Colascione]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

=== modified file 'lisp/progmodes/compile.el'
- --- lisp/progmodes/compile.el	2010-03-24 13:41:07 +0000
+++ lisp/progmodes/compile.el	2010-03-25 03:22:22 +0000
@@ -262,7 +262,7 @@
 \\(?:-\\([0-9]+\\)?\\(?:\\.\\([0-9]+\\)\\)?\\)?:\
 \\(?: *\\(\\(?:Future\\|Runtime\\)?[Ww]arning\\|W:\\)\\|\
  *\\([Ii]nfo\\(?:\\>\\|rmationa?l?\\)\\|I:\\|instantiated
from\\|[Nn]ote\\)\\|\
- -\[0-9]?\\(?:[^0-9\n]\\|$\\)\\|[0-9][0-9][0-9]\\)"
+\[0-9]?\\(?:[^0-9\n]\\)\\|[0-9][0-9][0-9]\\)"
      1 (2 . 5) (4 . 6) (7 . 8))

     ;; The `gnu' style above can incorrectly match gcc's "In file

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (Darwin)

iEYEARECAAYFAkuq4ZsACgkQ17c2LVA10VvzmwCfZ/SDqMc4EkLjwdiDuxARiWLV
LuAAoNYkH3yw0VtDaPNpczxQ1xMAUaa0
=27+E
-----END PGP SIGNATURE-----

Re: [PATCH] eliminate false positive in compile.el

[Stefan Monnier]

> === modified file 'lisp/progmodes/compile.el'
> - --- lisp/progmodes/compile.el	2010-03-24 13:41:07 +0000
> +++ lisp/progmodes/compile.el	2010-03-25 03:22:22 +0000
> @@ -262,7 +262,7 @@
>  \\(?:-\\([0-9]+\\)?\\(?:\\.\\([0-9]+\\)\\)?\\)?:\
>  \\(?: *\\(\\(?:Future\\|Runtime\\)?[Ww]arning\\|W:\\)\\|\
>   *\\([Ii]nfo\\(?:\\>\\|rmationa?l?\\)\\|I:\\|instantiated
> from\\|[Nn]ote\\)\\|\
> -\[0-9]?\\(?:[^0-9\n]\\|$\\)\\|[0-9][0-9][0-9]\\)"
> +\[0-9]?\\(?:[^0-9\n]\\)\\|[0-9][0-9][0-9]\\)"

which simplifies to

\[0-9]?[^0-9\n]\\|[0-9][0-9][0-9]\\)"

Could you give some rationale for this change other than just "eliminate
false positive"?  I mean, someone went to the trouble to add the
\\(?:...\\|$\\), so there should be a good reason to remove it,


        Stefan

Re: [PATCH] eliminate false positive in compile.el

[Daniel Colascione]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 3/25/10 3:24 PM, Stefan Monnier wrote:
>> -\[0-9]?\\(?:[^0-9\n]\\|$\\)\\|[0-9][0-9][0-9]\\)"
>> +\[0-9]?\\(?:[^0-9\n]\\)\\|[0-9][0-9][0-9]\\)"
> 
> which simplifies to
> 
> \[0-9]?[^0-9\n]\\|[0-9][0-9][0-9]\\)"

Right; I just figured I'd make the change minimal.

> 
> Could you give some rationale for this change other than just "eliminate
> false positive"?  I mean, someone went to the trouble to add the
> \\(?:...\\|$\\), so there should be a good reason to remove it,

Well, I prefer to not match "included from" lines from gcc, but with the
current gnu compile matcher, it's impossible not to do that. What these
lines all have in common is that they end with ":" or ":,". gnu doesn't
match the latter, but it does match the former, which is always the last
line before the real error message.

By removing |$, we prevent the gnu matcher from matching compile error
lines that appear to have an empty message, which fixes the above issue.
The gcc-include matcher does catch the "included from" lines if people
want to use it.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (Darwin)

iEYEARECAAYFAkurvj8ACgkQ17c2LVA10VtzqwCdH70b7c4C8PI16DNNBCh1fIm/
j9UAn365appc78wzgCaPB324IltUUk57
=U7eK
-----END PGP SIGNATURE-----

Re: [PATCH] eliminate false positive in compile.el

[Daniel Colascione]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 3/25/10 5:03 PM, Dan Nicolaescu wrote:
> So what should happen with those lines?
> I hop this does not mean that they are not highlighted at all, I use
> them all the time to look at all inclusion points looking for
> problems.

'gnu' never matched all those lines reliably anyway. 'gcc-include' does
what you want, and it's still enabled by default.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (Darwin)

iEYEARECAAYFAkur0AgACgkQ17c2LVA10VuJnwCdGWwS9yyflLkM2ctvcZX/arm2
1EgAoOZKSNxUyPL0aJqyyP1YDyRlHlup
=qtNN
-----END PGP SIGNATURE-----