List:MySQL++« Previous MessageNext Message »
From:Warren Young Date:July 4 2014 4:53am
Subject:Re: RFC to add option to use mysql_config
View as plain text  
On 7/2/2014 18:13, Adrian Cornish wrote:
> My autoconf stuff is not very good, this seems to work for me - but
> what should I do to improve it

I'd prefer that it worked like this:

1. If --with-mysql* is given, use those values without question.

2. If mysql_config runs, use its outputs without question.

3. If we don't get paths from either the configure script flags or 
mysql_config, use the current list of plausible paths to try and find 
the headers and C API library automatically.

Explicit error checks aren't needed for options 1 and 2.  They give us 
inputs for the purpose of overriding the default behavior, so we should 
believe whoever told us that these are good values.  Later configure 
script checks will implicitly test their correctness.

Option 1 is first to allow you to override the more automatic methods.

It should be extremely uncommon for there to be multiple mysql_config 
scripts in the PATH, so making the user specify the location is 
optimizing for the uncommon case.  If there are multiple such scripts in 
the system, we can assume that the local sysadmin put the "best" one 
first in the PATH.

To use a secondary mysql_config script, you can say things like this:

     CFLAGS="/path/to/mysql_config --cflags" LDFLAGS="..." ./configure

By the way, you don't need to put the patch inline.  Mailers often wreck 
inline patches, so attaching them is the best way.
RFC to add option to use mysql_configAdrian Cornish3 Jul
  • Re: RFC to add option to use mysql_configWarren Young4 Jul
    • Re: RFC to add option to use mysql_configWarren Young4 Jul
      • Re: RFC to add option to use mysql_configAdrian Cornish4 Jul
        • Re: RFC to add option to use mysql_configWarren Young9 Jul