[regex] Unable to replace just at the beginning/at the end of a string.

[axbender]

Describe the bug

Matching at the start/end of a string doesn't work as expected; it seems that ^ and $ are not correctly honored (see
sample below under "Reproduction Steps").

The following Python code does what I intend to do:

import re
g_name = r"_#abc.#_ab#"

g_re = r"(^[#.]+)|([#.]+$)" # Replace either at the beginning or at the end of the string
print(re.sub(g_re, "*", g_name))

Correctly prints: abc.#_ab

g_re = r"^[_#.]+" # Replace only at the beginning of the string
print(re.sub(g_re, "*", g_name))

Correctly prints: *abc.#_ab#

Reproduction Steps

import regex

g_name := r"_#abc.#_ab#"

mut g_re := regex.regex_opt(r"(^[#.]+)|([#.]+$)")!
println(g_re.replace(g_name, "*"))
// Prints: _#abc.#_ab#, i.e. nothing is replaced.

g_re = regex.regex_opt(r"^[_#.]+")!
println(g_re.replace(g_name, ""))
// Prints: abcab, i.e. EVERY occurrence is replaced. What about the start-of-line anchor?

Expected Behavior

Replace correctly (cmp. the Python results).

Current Behavior

Are documented in "Reproduction Steps".

Possible Solution

Additional Information/Context

No response

V version

0.5.0 6be2fb

Environment details (OS name and version, etc.)

Windows 10.

Note

You can use the 👍 reaction to increase the issue's priority for developers.

Please note that only the 👍 reaction to the issue itself counts as a vote.
Other reactions and those to comments will not be taken into account.

[axbender]

Similar (but not entirely) to #22119

[penguindark]

I'll check it this weekend.

[penguindark]

OK I'm working on it and I confirm PCRE is a total mess! 👎
python test:

g_re = r"(^[#.]+)|([#.]+$)"
g_name = r"_#abc.#_ab#"
print(re.sub(g_re, "*", g_name))
# Output: _#abc.#_ab*

vregex can't do this because in its architecture ^ and $ are mean to be used only for whole expression not in | alternatives,
this choice let vregex to achieve a top speed on big data.
On the other hand I understand the users' requests for a PCRE compliant regex, thus I'm working on a regex PCRE compliant that achieve the same result of Python at a cost of 20 times slower performance on big data.
Anyway thanks to this issue I've found a bug linked to ^ and $ in vregex not related to this issue and I'm making a PR to fix it.
In conclusion I can say: vregex can not use start/end anchors with |, the new vpcre will do, i've yet tested your case in it and it work as you expect.
I'm sorry that I can not solve your issue in vregex.

[axbender]

Sad to hear that.
But what about the second example, w/o the "|"? "^" seems to be ignored altogether...

How can the new engine be used (is it available already)?

[JalonSolov]

import regex.pcre

[penguindark]

you can have a look at new regex.pcre docs here: https://github.com/vlang/v/blob/master/vlib/regex/pcre/README.md