use IS_SLASH in various places instead of string comparing

[LamentXU123]

we have a defined IS_SLASH function:

#define IS_SLASH(c)	((c) == '/' || (c) == '\\')

So to make things more legible we can use IS_SLASH(XXX) instead of XXX == '\\' || XXX == '/'

I search all occasions by this command:

rg -n "'/'.*'\\'|'\\'.*'/'|== '/' \|\| .*== '\\'|!= '/' && .*!= '\\'|== '\\' \|\| .*== '/'|!= '\\' && .*!= '/'"

and manually remove some improper occasions :)

[TimWolla]

You missed some, for example

php-src/ext/phar/phar_object.c

Line 3568 in 95a17af

if (next_char == '/' || next_char == '\\' || next_char == '\0') {

Please do the search with a robust tool, not with regular expressions. The following Coccinelle patch can be a good start:

@@
char x;
@@

* (x == '/' || x == '\\')

@@
char x;
@@

* (x == '\\' || x == '/')

[LamentXU123]

You missed some, for example

php-src/ext/phar/phar_object.c

Line 3568 in 95a17af

if (next_char == '/' || next_char == '\\' || next_char == '\0') {

Please do the search with a robust tool, not with regular expressions. The following Coccinelle patch can be a good start:

No this is intented... If we change this the test on ext/phar/tests/bug64931/bug64931.phpt will fail. See https://github.com/php/php-src/actions/runs/24648588020/job/72066292649 as I've said I manually deleted some changes

The following Coccinelle patch can be a good start:

Yeah I'll try now, thank you :)

[TimWolla]

If we change this the test on ext/phar/tests/bug64931/bug64931.phpt will fail

That's not how C works.

[iluuu1994]

This changes behavior, no? IS_SLASH() is defined as #define IS_SLASH(c) ((c) == '/') on non-Windows systems.

[TimWolla]

IS_SLASH() is defined as #define IS_SLASH(c) ((c) == '/') on non-Windows systems.

Ah, that explains the observed test failure then. And yes, this means that this PR is not a simple cleanup.

[iluuu1994]

Right, so I think this PR is inherently problematic. Closing, but thanks anyway.

[LamentXU123]

IS_SLASH() is defined as #define IS_SLASH(c) ((c) == '/') on non-Windows systems.

Ah, that explains the observed test failure then. And yes, this means that this PR is not a simple cleanup.

Uh that explains test failure. Then good to close this, and thank for reviewing anyways :)