lib/db: use integer port argument with mysql_real_connect in mysql driver

[fweimer-rh]

The port argument is not a string, but a plain integer. In the past, compilers ignore such type errors and produced a warning only, but this is changing, so this change also addresses a potential build failure.

Note that I can't really test this, but at least the compiler error is gone. Found as part of:

• https://fedoraproject.org/wiki/Changes/PortingToModernC
• https://fedoraproject.org/wiki/Toolchain/PortingToModernC

[wenzeslaus]

Thank you for the fix!

This is consistent with mysql_real_connect and db_get_login2 which creates the string.

[nilason]

Thanks @fweimer-rh !
mysql_real_connect() does indeed use unsigned int for the port (in comparison to PQsetdbLogin() which uses char *).

[wenzeslaus]

Anyone has a trick for quickly fixing clang-format issues without pulling the code locally? Unlike with Black, I'm not good at guessing what the right formatting is from the message.

[echoix]

Its ugly. I wasn't able to create a good suggestion.

[neteler]

(For the record for those not familiar with gh: it is fairly straighforward)

gh pr checkout 3325
utils/grass_clang_format.sh db/drivers/mysql/db.c
git diff
git commit -m"utils/grass_clang_format.sh db/drivers/mysql/db.c" db/drivers/mysql/db.c
git push 

Done in fc9803f

[nilason]

Anyone has a trick for quickly fixing clang-format issues without pulling the code locally? Unlike with Black, I'm not good at guessing what the right formatting is from the message.

#3255 ;-)

[nilason]

I believe we have no CI test with MySQL driver, I’ll look into this. I suppose this should have been caught.

[wenzeslaus]

[The formatting is] ugly. I wasn't able to create a good suggestion.

I agree. We didn't necessarily choose the format. We used whatever was already there and what clang-format supported at the time hoping to get back to creating a nicer style with a more powerful version of clang-format.

[nilason]

[The formatting is] ugly. I wasn't able to create a good suggestion.

I agree. We didn't necessarily choose the format. We used whatever was already there and what clang-format supported at the time hoping to get back to creating a nicer style with a more powerful version of clang-format.

Attractive or not, the point of using formatters like ClangFormat and Black is to avoid non-productive discussions on formatting with potentially no end ...... :)

[wenzeslaus]

...the point of using formatters like ClangFormat and Black is to avoid non-productive discussions on formatting with potentially no end ...... :)

Again, I deny ever taking a part in such discussion.