E474: Failed to parse file

[pseregiet]

I'm working with u-boot, it's a bootloader for linux (and other) and it happens to have some non-ascii stuff in one of the headers (a font definition). AnyJumps fails to search for anything in this codebase because of this

Error detected while processing function 15_Jump[30]..search#SearchDefinitions[33]..46_RunRgDefinitionSearch[8]..46_Parse
RgResults:
line 7:
E474: Unidentified byte: include/video_font_data.h: PCRE2: error matching: UTF-8 error: isolated byte with 0x80 bit set
E474: Failed to parse include/video_font_data.h: PCRE2: error matching: UTF-8 error: isolated byte with 0x80 bit set
Press ENTER or type command to continue

Would it be possible to just ignore such errors and not include that file in the result ?

[pseregiet]

Bumping this issue because it happens in more repositories and it's really annoying. This plugin really rules and I don't want to change to another :( Please help.

Error detected while processing function 15_Jump[30]..search#SearchDefinitions[33]..46_RunRgDefinitionSearch[8]..46_ParseRgResult
s:
line 7:
E474: Unidentified byte: lib/fonts/font_8x8.c: PCRE2: error matching: UTF-8 error: isolated byte with 0x80 bit set
E474: Failed to parse lib/fonts/font_8x8.c: PCRE2: error matching: UTF-8 error: isolated byte with 0x80 bit set

[pechorin]

i cannot reproduce anything with https://gitlab.archlinux.org/dvzrv/linux-rt/-/blob/fe82702ad8dd2918a86949902fc2d1c0bc375274/lib/fonts/font_8x8.c#L474

can you provide example and search keyword?

[pseregiet]

i cannot reproduce anything with https://gitlab.archlinux.org/dvzrv/linux-rt/-/blob/fe82702ad8dd2918a86949902fc2d1c0bc375274/lib/fonts/font_8x8.c#L474

can you provide example and search keyword?
--

Sorry for the very late reply... I have fixed this problem myself by adding --no-pcre2-unicode argument to rg in the autoload/search.vim file.

diff --git a/autoload/search.vim b/autoload/search.vim
index bd95509..fad4a85 100644
--- a/autoload/search.vim
+++ b/autoload/search.vim
@@ -3,7 +3,7 @@
 let s:regexp_keyword_word = 'KEYWORD'
 let s:engines             = ['rg', 'ag']

-let s:rg_base_cmd = "rg -n --auto-hybrid-regex --json"
+let s:rg_base_cmd = "rg -n --auto-hybrid-regex --json --no-pcre2-unicode"
 let s:ag_base_cmd = "ag --nogroup --noheading"

 let s:rg_filetype_convertion_map = {
@@ -39,6 +39,7 @@ let s:non_standard_ft_extensions_map = {
       \}

 let s:filetypes_comments_map = {
+      \"c":             ["//"],
       \"cpp":           ["//"],
       \"elisp":         [";"],
       \"commonlisp":    [";"],

I've pulled the origin/master and this error is still present. This is the offending file: https://elixir.bootlin.com/u-boot/latest/source/include/video_font_data.h
This error shows up when searching (coma J) for any word, as it (I assume) calls rg for the whole repo and rg fails without that argument on that font file. Perhaps you have PCRE2 disabled in config of rg itself ?

[dkalowsk]

@pseregiet thanks for the initial leg work here. I was seeing a similar issue for C files that I was using as well, although the error message wasn't providing a file or line... just error. From what I could see, it looked like the C regexs for any-jump were all PCRE2 only.

I ended up changing the line to support rg v12 and 13 and got it working locally with:

let s:rg_base_cmd = "rg -n --pcre2 --engine auto --json --unicode"