GN: Add the ability to trim whitespace from script results.

GN: Add the ability to trim whitespace from script results.

Adds an optional "trim" mode that can trim whitespace from the beginnings and endings of script results when converting them to GN values.

This also adds default arguments for exec_script to simplify calling in cases where the arguments are not needed.

R=bbudge@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=244586

[bbudge, 2014-01-08 23:22:31]

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
File tools/gn/function_exec_script.cc (right):

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
tools/gn/function_exec_script.cc:239: "  exec_script(filename [, arguments [,
input_conversion [,\n"
I can't grok these brackets.

https://codereview.chromium.org/129043005/diff/30001/tools/gn/input_conversio...
File tools/gn/input_conversion.cc (right):

https://codereview.chromium.org/129043005/diff/30001/tools/gn/input_conversio...
tools/gn/input_conversion.cc:150: *err = Err(input_conversion_value, "Not a
valid input conversion.",
Shouldn't this be 'input_conversion'?
The comment above this function wouldn't be correct then, since we would trim
off the "trim " before hitting this.

[brettw, 2014-01-09 23:48:53]

New snap up.

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
File tools/gn/function_exec_script.cc (right):

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
tools/gn/function_exec_script.cc:239: "  exec_script(filename [, arguments [,
input_conversion [,\n"
I changed this to use a more C++ style of annotating default arguments.

https://codereview.chromium.org/129043005/diff/240001/tools/gn/input_conversi...
File tools/gn/input_conversion.cc (right):

https://codereview.chromium.org/129043005/diff/240001/tools/gn/input_conversi...
tools/gn/input_conversion.cc:150: *err = Err(input_conversion_value, "Not a
valid input_conversion.",
I think this is correct. The first argument is what gets blamed in the error
message. We want to point to the string that the user specified in the call,
which this will do, and is why I pass this Value around.

We could identify just the substring after removing the "trim " prefix but
that's pretty tricky. I do this for string interpolation (using $ in strings to
replace with variables). But I don't think it's worth it in this case. Here, if
you say "trim foo" it will point to the entire string and say it was invalid.

[bbudge, 2014-01-10 00:37:54]

lgtm

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
File tools/gn/function_exec_script.cc (right):

https://codereview.chromium.org/129043005/diff/30001/tools/gn/function_exec_s...
tools/gn/function_exec_script.cc:239: "  exec_script(filename [, arguments [,
input_conversion [,\n"
On 2014/01/09 23:48:53, brettw wrote:
> I changed this to use a more C++ style of annotating default arguments.

Sorry, I somehow didn't think the left/right brackets matched up. I see now what
this means. Feel free to use this style if you like.

https://codereview.chromium.org/129043005/diff/240001/tools/gn/input_conversi...
File tools/gn/input_conversion.cc (right):

https://codereview.chromium.org/129043005/diff/240001/tools/gn/input_conversi...
tools/gn/input_conversion.cc:150: *err = Err(input_conversion_value, "Not a
valid input_conversion.",
On 2014/01/09 23:48:53, brettw wrote:
> I think this is correct. The first argument is what gets blamed in the error
> message. We want to point to the string that the user specified in the call,
> which this will do, and is why I pass this Value around.
> 
> We could identify just the substring after removing the "trim " prefix but
> that's pretty tricky. I do this for string interpolation (using $ in strings
to
> replace with variables). But I don't think it's worth it in this case. Here,
if
> you say "trim foo" it will point to the entire string and say it was invalid.

OK, I see now. Maybe a better name would be original_input_conversion?

[brettw, 2014-01-13 21:33:18]

Committed patchset #4 manually as r244586 (presubmit successful).