Re: [bug #40639] GNU Make with profiling information

[Tim Murphy]

On 15 December 2013 16:07, Paul Smith <address@hidden> wrote:

> In other words, I prefer to take a page from Git, GDB, and other
> projects where the default output is human readable but probably not
> easily parsed by tools, and then provide a different output format
> option that provides machine-parse-able formats.  I'm not interested in
> trying to create one output format that solves both of those problems.

The way I have seen this addressed before is to have a filters that
render the comprehensive output into human-readable forms of varying
kinds. e.g. if you like the Linux kernel-style where your output is
only "CC file.o" unless there's an error, then a filter can be made
that produces this. If you prefer gnu makes' standard output then that
would be the default. If you want coloration or to produce output as
HTML then those are also possible.

For various reasons the filters we used had to parse the "structured
output" into readable form but if you were designing such a thing into
make then you could do it without a print-then-parse. It's fairly key
for recipes to be able to inject information into this output e.g.
attribute=value pairs so that logical information about the build can
be passed into the log or on towards reporting tools.  e.g. the name
of the team to which a particular target belongs or whether it is an
intermediate, internal or releasable file.

Even more interesting would be an indexing filter which would generate
a table that allowed one to jump directly to the point in a log where
some file was mentioned or where some error occurred. When builds get
really big this would be nice as even grep can be overwhelmed by
something large enough.

I have gone off-topic. I apologise.

Regards,

Tim

-- 
You could help some brave and decent people to have access to
uncensored news by making a donation at:

http://www.thezimbabwean.co.uk/friends/