bug#71844: [PATCH] 29.1; "cucumber" in compilation-error-regexp-alist-alist is too aggressive

bug#71844: [PATCH] 29.1; "cucumber" in compilation-error-regexp-alist-alist is too aggressive

[Matthew Bauer]

[-- Attachment #1: Type: text/plain, Size: 710 bytes --]

I believe this regex for cucumber is not precise enough. Specifically,
it was picking up this indentation in an error message from GHC:

Main.hs:195:1: error: [GHC-44432]
    The type signature for ‘apathway’ lacks an accompanying binding
    Suggested fix:
      Perhaps use one of these:
        ‘foo’ (Defined at Main.hs:196:1),
        ‘foo’ (Defined at Main.hs:216:1)

It picked up this as being error messages which I believe it shouldn’t have:

   ‘foo’ (Defined at Main.hs:216
   ‘foo’ (Defined at Main.hs:216

This change just ignores it if there is remaining space. Probably it
still needs to be refined, but this at least avoids this false positive
for me.


[-- Attachment #2: Fix "cucumber" regexp --]
[-- Type: text/x-patch, Size: 1666 bytes --]

From 3a5e45ec2d836a38348e802c441d567e070b138c Mon Sep 17 00:00:00 2001
From: Matthew Bauer <mjbauer95@gmail.com>
Date: Sat, 29 Jun 2024 14:45:35 -0500
Subject: [PATCH] Try to limit false positives for cucumber compilation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

I believe this regex for cucumber is not precise enough. Specifically,
it was picking up this indentation in an error message from GHC:

Main.hs:195:1: error: [GHC-44432]
    The type signature for ‘apathway’ lacks an accompanying binding
    Suggested fix:
      Perhaps use one of these:
        ‘foo’ (Defined at Main.hs:196:1),
        ‘foo’ (Defined at Main.hs:216:1)

It picked up this as being error messages which I believe it shouldn’t have:

   ‘foo’ (Defined at Main.hs:216
   ‘foo’ (Defined at Main.hs:216

This change just ignores it if there is remaining space. Probably it
still needs to be refined, but this at least avoids this false positive
for me.
---
 lisp/progmodes/compile.el | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lisp/progmodes/compile.el b/lisp/progmodes/compile.el
index 2e4eb11811a..a03f1935971 100644
--- a/lisp/progmodes/compile.el
+++ b/lisp/progmodes/compile.el
@@ -473,7 +473,7 @@ of[ \t]+\"?\\([a-zA-Z]?:?[^\":\n]+\\)\"?:" 3 2 nil (1))
                    "     "))
              "#")
           " "
-          (group (not "(") (* nonl))          ; file
+          (group (not (in "( ")) (* nonl))          ; file
           ":"
           (group (in "1-9") (* (in "0-9"))))  ; line
      1 2)
-- 
2.39.3 (Apple Git-146)

bug#71844: [PATCH] 29.1; "cucumber" in compilation-error-regexp-alist-alist is too aggressive

[Eli Zaretskii]

> From: Matthew Bauer <mjbauer95@gmail.com>
> Date: Sat, 29 Jun 2024 14:55:05 -0500
> 
> I believe this regex for cucumber is not precise enough. Specifically,
> it was picking up this indentation in an error message from GHC:
> 
> Main.hs:195:1: error: [GHC-44432]
>     The type signature for ‘apathway’ lacks an accompanying binding
>     Suggested fix:
>       Perhaps use one of these:
>         ‘foo’ (Defined at Main.hs:196:1),
>         ‘foo’ (Defined at Main.hs:216:1)
> 
> It picked up this as being error messages which I believe it shouldn’t have:
> 
>    ‘foo’ (Defined at Main.hs:216
>    ‘foo’ (Defined at Main.hs:216

Why not? don't you want to be able to visit the definitions to examine
whether they are the right fix?

More generally, compilation-mode doesn't (yet) support the "fix
suggestion" features offered by some of the modern compilers.  IMO, we
should address this kind of situations in a general framework of
supporting fix suggestions, not just by ad-hoc changes in the regexps
that recognize error messages.  Patches welcome.