Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:121206
MIME-Version: 1.0
References: <CAEPPVa3d6okq-wmeFSjjcuix_TyiU7pK45TrquvgMjKY1eSX3A@mail.gmail.com>
 <B5AAA3E4-525B-4D01-8873-051EFE599BBB@sakiot.com> <CAEPPVa12Rv6bwG8rzSA8R9kdmXP+CatvMiQNr279-jVzez9AOg@mail.gmail.com>
 <CAJmy8YEe3EOyEgH1hnCnctqYOPGhzy0HTpQ=LX77vHC9Cb40+Q@mail.gmail.com>
 <CAJmy8YHAvjMWLc1Tbb4qNj1650orvAExdO_Ef8t2CdNriPutdw@mail.gmail.com> <CAEPPVa3EMrJFnCnir9jSVA_UiLWM4EoCea7MriTU3WyCOdxR9A@mail.gmail.com>
In-Reply-To: <CAEPPVa3EMrJFnCnir9jSVA_UiLWM4EoCea7MriTU3WyCOdxR9A@mail.gmail.com>
Date: Mon, 2 Oct 2023 16:24:18 +0200
Message-ID: <CAJmy8YFG1uXUDsey4bQBP+WmNUKOcuwoyOikgr8H8Oe-ANJHBQ@mail.gmail.com>
To: youkidearitai <youkidearitai@gmail.com>
Cc: internals@lists.php.net
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Subject: Re: [PHP-DEV] trim support for multibyte spaces
From: divinity76@gmail.com (Hans Henrik Bergan)

add a
void str_dump(const size_t strlen, const char *str)
{
    printf("string(%zu) \"", strlen);
    fwrite(str, strlen, 1, stdout);
    printf("\"\n");
}

then replace the string_view stuff with

    size_t teststrlen =3D strlen(teststr);
    str_dump(teststrlen, teststr);
    mb_trim(&teststrlen, &teststr, trim_lengths_num, trim_lengths, trim_cha=
rs);
    str_dump(teststrlen, teststr);

and it should be pure C, i think

On Mon, 2 Oct 2023 at 15:36, youkidearitai <youkidearitai@gmail.com> wrote:
>
> Hi, Hans.
>
> Thank you very much for your code. And sorry for late.
> I confirmed works fine.
> https://gist.github.com/youkidearitai/0018dee27353c00aebaff3bf57c5b8c6
>
> However, this code is C++17, php-src is C code.
> If you would like contribute, I would like written to C code.
>
> Regards
> Yuya
>
> 2023=E5=B9=B410=E6=9C=881=E6=97=A5(=E6=97=A5) 19:46 Hans Henrik Bergan <d=
ivinity76@gmail.com>:
> >
> > .. probably a bunch of stuff that *could* be optimized or done better,
> > but one i saw just now is that instead of 2x nested loops and goto,
> > the outer loop and labels could be removed and the `goto
> > remove_from_start_continue_2;` could be replaced with `i=3D-1;` eg
> >
> >     size_t local_strlen =3D *strlen;
> >     char *local_str =3D *str;
> >     for (size_t i =3D 0; i < trim_lengths_num; ++i)
> >     {
> >         if (local_strlen >=3D trim_lengths[i] && memcmp(local_str,
> > trim_chars[i], trim_lengths[i]) =3D=3D 0)
> >         {
> >             local_strlen -=3D trim_lengths[i];
> >             local_str +=3D trim_lengths[i];
> >             i =3D -1;
> >         }
> >     }
> >
> > 2x nested loops reduced to 1 loop, and goto removed~
> >
> > On Sun, 1 Oct 2023 at 10:43, Hans Henrik Bergan <divinity76@gmail.com> =
wrote:
> > >
> > > > If have any idea, feel free to comment to me.
> > >
> > > i think the C code would look something like
> > >
> > >
> > > void mb_trim(size_t *strlen, char **str, const size_t
> > > trim_lengths_num, const size_t *trim_lengths, const char **trim_chars=
)
> > > {
> > >     size_t local_strlen =3D *strlen;
> > >     char *local_str =3D *str;
> > >     for (;;)
> > >     {
> > >         for (size_t i =3D 0; i < trim_lengths_num; ++i)
> > >         {
> > >             if (local_strlen >=3D trim_lengths[i] && memcmp(local_str=
,
> > > trim_chars[i], trim_lengths[i]) =3D=3D 0)
> > >             {
> > >                 local_strlen -=3D trim_lengths[i];
> > >                 local_str +=3D trim_lengths[i];
> > >                 goto remove_from_start_continue_2;
> > >             }
> > >         }
> > >         break;
> > >     remove_from_start_continue_2:;
> > >     }
> > >     for (;;)
> > >     {
> > >         for (size_t i =3D 0; i < trim_lengths_num; ++i)
> > >         {
> > >             if (local_strlen >=3D trim_lengths[i] && memcmp(((local_s=
tr
> > > + local_strlen) - trim_lengths[i]), trim_chars[i], trim_lengths[i]) =
=3D=3D
> > > 0)
> > >             {
> > >                 local_strlen -=3D trim_lengths[i];
> > >                 goto remove_from_end_continue_2;
> > >             }
> > >         }
> > >         break;
> > >     remove_from_end_continue_2:;
> > >     }
> > >     memmove(*str, local_str, local_strlen);
> > >     char *newstr =3D (char *)realloc(*str, local_strlen);
> > >     if (newstr !=3D nullptr)
> > >     {
> > >         *strlen =3D local_strlen;
> > >         *str =3D newstr;
> > >     }
> > >     else
> > >     {
> > >         // some error handling
> > >     }
> > > }
> > >
> > > with my simple testcode looking like
> > >
> > > int main()
> > > {
> > >     const char *trim_chars[] =3D {
> > >         " ",
> > >         "!",
> > >         // utf8 whitespace:
> > >         "\xE2\x80\x80", // EN QUAD
> > >         "\xE2\x80\x81", // EM QUAD
> > >         "\xE2\x80\x82", // EN SPACE
> > >         "\xE2\x80\x83", // EM SPACE
> > >         "\xE2\x80\x84", // THREE-PER-EM SPACE
> > >         "\xE2\x80\x85", // FOUR-PER-EM SPACE
> > >         "\xE2\x80\x86", // SIX-PER-EM SPACE
> > >     };
> > >     size_t trim_lengths[] =3D {
> > >         strlen(trim_chars[0]),
> > >         strlen(trim_chars[1]),
> > >         strlen(trim_chars[2]),
> > >         strlen(trim_chars[3]),
> > >         strlen(trim_chars[4]),
> > >         strlen(trim_chars[5]),
> > >         strlen(trim_chars[6]),
> > >         strlen(trim_chars[7]),
> > >         strlen(trim_chars[8]),
> > >     };
> > >     size_t trim_lengths_num =3D sizeof(trim_lengths) / sizeof(trim_le=
ngths[0]);
> > >     char *teststr =3D strdup("  !  \xE2\x80\x80\xE2\x80\x81\xE2\x80\x=
82
> > >  Hello World !  \xE2\x80\x83\xE2\x80\x84\xE2\x80\x85\xE2\x80\x86  !
> > > ");
> > >     // char *teststr =3D strdup("      !    Hello World !        !  "=
);
> > >     size_t teststrlen =3D strlen(teststr);
> > >     std::cout << teststrlen << ": \"" << std::string_view(teststr,
> > > teststrlen) << "\"" << std::endl;
> > >     mb_trim(&teststrlen, &teststr, trim_lengths_num, trim_lengths, tr=
im_chars);
> > >     std::cout << teststrlen << ": \"" << std::string_view(teststr,
> > > teststrlen) << "\"" << std::endl;
> > >     return 0;
> > > }
> > >
> > >
> > > On Sat, 30 Sept 2023 at 13:16, youkidearitai <youkidearitai@gmail.com=
> wrote:
> > > >
> > > > 2023=E5=B9=B49=E6=9C=8830=E6=97=A5(=E5=9C=9F) 17:42 Saki Takamachi =
<saki@sakiot.com>:
> > > > >
> > > > > > I also want to trim function of multibyte trim functions.
> > > > >
> > > > > > I think that in addition to mb_trim,
> > > > > mb_ltrim and mb_rtrim are also necessary.
> > > > >
> > > > > Hi.
> > > > >
> > > > > Having a new option besides regex sounds like a good idea for me,=
 as a user of a language that benefits from `mb_trim()`.
> > > > >
> > > > > Perhaps users are more intuitive about the usefulness of those fu=
nctions who have spent more time cleaning "multibyte spaces".
> > > >
> > > > Hi
> > > >
> > > > Thanks for reply.
> > > > I'm trying implements to trim for multibyte function.
> > > > Please give me some time.
> > > >
> > > > If have any idea, feel free to comment to me.
> > > >
> > > > Regards
> > > > Yuya
> > > >
> > > > --
> > > > ---------------------------
> > > > Yuya Hamada (tekimen)
> > > > - https://tekitoh-memdhoi.info
> > > > - https://github.com/youkidearitai
> > > > -----------------------------
> > > >
> > > > --
> > > > PHP Internals - PHP Runtime Development Mailing List
> > > > To unsubscribe, visit: https://www.php.net/unsub.php
> > > >
>
>
>
> --
> ---------------------------
> Yuya Hamada (tekimen)
> - https://tekitoh-memdhoi.info
> - https://github.com/youkidearitai
> -----------------------------
>
> --
> PHP Internals - PHP Runtime Development Mailing List
> To unsubscribe, visit: https://www.php.net/unsub.php
>