mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "James Peach" <jpe...@apache.org>
Subject Re: Review Request 33752: MESOS-2537: Fix AC_ARG_ENABLED option processing
Date Wed, 26 Aug 2015 21:14:57 GMT


> On Aug. 26, 2015, 9:01 p.m., Timothy St. Clair wrote:
> >
> 
> Timothy St. Clair wrote:
>     Have you tested all combinations here?

No, only some of the possible combinations. The addition of buld flags output makes is easier
to tell when it it doing the right thing. This change doesn't really change any logic except
for ```--enable-debug``` and ```--enable-optimize```, since the absence of defaults for those
is confusing.


> On Aug. 26, 2015, 9:01 p.m., Timothy St. Clair wrote:
> > configure.ac, line 145
> > <https://reviews.apache.org/r/33752/diff/5/?file=1055235#file1055235line145>
> >
> >     doesn't this change default toggle to an evaluation of the output variable?
 
> >     
> >     Shouldn't this just be a toggle on existance?  
> >     
> >     I'm ok on uses of "$withval" but do you have an arguement for the enableval?

The ```--enable-foo``` syntax also supports ```--disable-foo```. So if the build option is
not specified, the default should apply, otherwise the operator can specify the build option
must be present or absent. For example, our internal RPM build was specifying ```--disable-perftools```,
but this had the unintended effect of enabling perftools.


- James


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/33752/#review96588
-----------------------------------------------------------


On Aug. 26, 2015, 8:11 p.m., James Peach wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/33752/
> -----------------------------------------------------------
> 
> (Updated Aug. 26, 2015, 8:11 p.m.)
> 
> 
> Review request for mesos, Cody Maloney and Timothy St. Clair.
> 
> 
> Bugs: MESOS-2537
>     https://issues.apache.org/jira/browse/MESOS-2537
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> In a number of places, the Mesos configure script passes "$foo=yes"
> to the 2nd argument of AC_ARG_ENABLED. However, the 2nd argument
> is invoked when the option is provided in any form, not just when
> the --enable-foo form is used. One result of this is that
> --disable-optimize doesn't work.
> 
> The correct handling of the 2nd argument is to save the value of
> "$enableval". This change sets the value of all the enable variables
> using $enableval, and sets the default value based on the option
> name.
> 
> There are a number of enable options that were internally named
> "$with_foo" and "$without_foo". Rename these to "$enable_foo" for
> clarity and to remove the need for both a with_ and a without_
> version.
> 
> Finally, emit the compilation flags at the end of the configure
> phase so it is easier to see the results of your configuration
> options.
> 
> 
> Diffs
> -----
> 
>   configure.ac 87461d73ed04c4cf176c3475ded9f98dadcda608 
> 
> Diff: https://reviews.apache.org/r/33752/diff/
> 
> 
> Testing
> -------
> 
> Configure and build on CentOS 7 and Mac OS X 10.10.3. Verify that the status summary
reflects the expected compiler flags. Verify that --enable-foo and --disable-foo do different
things.
> 
> 
> Thanks,
> 
> James Peach
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message