In D23109#506462, @yuripv wrote:

I wonder if we could instead check if we have the table already loaded and return without doing everything over and over again. Something like the following right at the start:

if (table->__collate_load_error == 0 && table->header != NULL && strcmp(table->header.locale, encoding) == 0)
    return (_LDP_LOADED);

Not completely sure about the table->header check here as we already checked that the table was loaded, but it's just an idea.
(P.S. I think encoding is really a wrong argument name here, as it's full locale name.)

I wonder if we could instead check if we have the table already loaded and return without doing everything over and over again. Something like the following right at the start:

Warner, does this make sense to you/won't interfere with other changes you might be working on?

actually split the synopsis into separate options that can't be used together

As noted by @imp, there are better ways to do this.

Correct the .Fo field (remove *). Ingo did agree that .Fo (f) shouldn't trigger the warning and it will be supressed in the future. For the moment, we'll just have a non-fatal warning.

• use sbuf functions
• also handle am_pm
• mark all keywords not specified by POSIX (e.g. abmon_1 through abmon_12, am_str, pm_str, ...) as local additions as other systems (at least Linux and Solaris) do not handle them
• add tests

In D21569#469935, @0mp wrote:

Seems fine to me.

@ian I'm not sure if I should put you in 'submitted by' or 'reviewed by' line, you did both :)

In D21452#466886, @ian wrote:

IMO, the much easier fix is to change the call to bus_generic_attach() at the end of the ichsmb_attach() routine to:

config_intrhook_oneshot((ich_func_t)bus_generic_attach, dev);
return (0);

We really need to do it to all i2c/smb controller drivers. I've done most of the arm ones already.

In D21452#466863, @hselasky wrote:

If this driver doesn't need to started early have a look at this:
We could modify SI_SUB_DRIVERS in the macro below to load the driver at a later point in time!

#define EARLY_DRIVER_MODULE_ORDERED(name, busname, driver, devclass,    \
    evh, arg, order, pass)                                              \
                                                                        \
static struct driver_module_data name##_##busname##_driver_mod = {      \
        evh, arg,                                                       \
        #busname,                                                       \
        (kobj_class_t) &driver,                                         \
        &devclass,                                                      \
        pass                                                            \
};                                                                      \
                                                                        \
static moduledata_t name##_##busname##_mod = {                          \
        #busname "/" #name,                                             \
        driver_module_handler,                                          \
        &name##_##busname##_driver_mod                                  \
};                                                                      \
DECLARE_MODULE(name##_##busname, name##_##busname##_mod,                \
               SI_SUB_DRIVERS, order)
/sys/sys/bus.h:#define	DRIVER_MODULE_ORDERED(name, busname, driver, devclass, evh, arg,\

In D21452#466857, @hselasky wrote:

Can you upload full context?

In D18705#401939, @imp wrote:

This is fine, but there's a chance we'll rework the arg parsing in this program in the near future... Better to get this in now though.

@eadler any comments on this as you requested pre-commit review for top?

@kevans any thoughts here? I'm still hesitant to add the symbol, so are you ok with the change only being in libc and libregex using the old behavior (strict character comparison instead of equivalence class)?

In D18408#400188, @evfbsd_yandex.ru wrote:

Hi @yuripv, are there plans to commit this?

@imp anything else I need to change here?

Are we *guaranteed* to have the floppy drives coming only at the start of the list?

Address comments from @imp.

In D18705#399280, @bcran wrote:

-N (delete BootNext) doesn’t require -b since it doesn’t operate on a bootnum.

-N doesn't need the -b bootnum.

Drop the xlocale_private.h change, it's not really needed.

In D18611#396895, @pfg wrote:

For the record: while it is good practice to request reviews before committing, IMHO they dont make much sense, and are not required at all, for reverts. After all they stand for things all the initial reviewers missed in the first place ;).

Looks good to me, I have even applied the patch and tested some corner cases I could think of :-)

@kevans are you OK with the test case added to libc/regex instead and using sed? :)

Oops, I didn't mean to add Ed back to this review :-)

Reading through the discussion, I do agree with @ed -- we should have just one version in UTF-8, and convert to current encoding on the fly, though I do agree with @AMDmi3 as well -- making sure we use only the characters that can be converted to the related single-byte locale shouldn't be that hard, we don't need to use the fancy unicode chars in messages.

Sorry for the delay. I have spent quite a bit more time on this, including testing. Summary and the change updated accordingly.

In D18334#393097, @asomers wrote:

LGTM. Do you you need someone to commit this for you?