Re: [icu-support] [icu4c] Loading .dat file for data

['Shane C. v. icu-s. <icu...@un...>]

I see 2 potential issues:

1. You should use udata_setCommonData not udata_setAppData
2. You should pass in the whole file, including the header, not just the
part returned by udata_getMemory. Try passing the return value of
udata_open to udata_setCommonData.

On Mon, Dec 23, 2024 at 9:41 AM Bastien Durel <bas...@qu...>
wrote:

> Hello,
>
> I'm building an application that should be deployed on various Linux
> via tgz, so I cannot rely on system's ICU.
> We have a policy a reducing the linked-in shared objects, so I link
> against a static libicu. To minimize the binary size (because I will
> have a lots of them), I try to get the data itself out of the binary,
> so I compiled icu4c with `--with-data-packaging=archive`
>
> I then link the apps against libicuuc.a & libicudata.a, and I added
> this code inside the initialization routine :
>
>
> std::filesystem::path icu_data_path = GetICUDataPath();
> auto package_name = icu_data_path.filename();
> auto* data = udata_open(
>     icu_data_path.parent_path().c_str(),
>     "dat",
>     package_name.stem().c_str(),
>     &status);
> if (U_FAILURE(status) != 0) {
>     std::cout << "icu_data_path: " << icu_data_path << std::endl;
>     std::cout << "data: " << data << std::endl;
>     std::cout << "*data: " << udata_getMemory(data) << std::endl;
>     throw std::runtime_error(VE_OPEN_ICU_DATA(
>         icu_data_path.string(), u_errorName(status)));
> }
> udata_setAppData("icudt", udata_getMemory(data), &status);
> if (U_FAILURE(status) != 0) {
>     throw std::runtime_error(VE_LOAD_ICU_DATA(
>         icu_data_path.string(), u_errorName(status)));
> }
>
>
> The udata_open() returns a pointer as expected, but the
> udata_setAppData() fails, so I get this exception :
>
> Failed to load ICU data from '/usr/local/share/icu/76.1/icudt76l.dat':
> U_INVALID_FORMAT_ERROR
>
> Do you know if the archive is supposed to be loaded by some other way ?
> Or if it's not intended to be opened this way ?
>
> Thanks,
>
> NB: sorry for re-posting, but the google group didn't looked like if
> was fully initialized, so I'm not sure the mail was really forwarded.
>
> Regards,
>
> --
> Bastien Durel
>
> --
> You received this message because you are subscribed to the Google Groups
> "icu-support" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to icu...@un....
> To view this discussion visit
> https://groups.google.com/a/unicode.org/d/msgid/icu-support/04d010fba1dc5fea95c6d62c6d0c90e3f3a60f7a.camel%40quetastream.com
> .
>
> --
> You received this message because you are subscribed to the Google Groups
> "ICU - Team" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to icu...@un....
> To view this discussion visit
> https://groups.google.com/a/unicode.org/d/msgid/icu-team/04d010fba1dc5fea95c6d62c6d0c90e3f3a60f7a.camel%40quetastream.com
> .
>

-- 
You received this message because you are subscribed to the Google Groups "icu-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to icu...@un....
To view this discussion visit https://groups.google.com/a/unicode.org/d/msgid/icu-support/CABxsp%3DmgxGkBnj94R5Ay_7r4fTfiKAjbkkTyjCj8USNd5a6A7g%40mail.gmail.com.