- Remove unneeded library tags when concatenating libraries.

- Remove unneeded library tags when concatenating libraries.
Committed: https://code.google.com/p/dart/source/detail?r=15368

[siva, 2012-11-27 02:18:07]

https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
File runtime/tools/concat_library.py (right):

https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
runtime/tools/concat_library.py:37: if line.startswith('#source') or
line.startswith('part '):
There is the danger of legal lines which start with part being removed here:
e.g:
func1(
param1 ,
part ,
sum);

Also, is it a requirement that 'part' has to be the first word without any
whitespace?
e.g:
   part of deepshit;  // Is this legal.

why not do:
line = re.sub("part of .*;", "", line);

[Ivan Posva, 2012-11-27 05:38:19]

https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
File runtime/tools/concat_library.py (right):

https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
runtime/tools/concat_library.py:37: if line.startswith('#source') or
line.startswith('part '):
On 2012/11/27 02:18:07, siva wrote:
> There is the danger of legal lines which start with part being removed here:
> e.g:
> func1(
> param1 ,
> part ,
> sum);
> 
> Also, is it a requirement that 'part' has to be the first word without any
> whitespace?
> e.g:
>    part of deepshit;  // Is this legal.
> 
> why not do:
> line = re.sub("part of .*;", "", line);

I understand all of these limitations (they went away with the new syntax which
is easier on the eyes and harder for tools). In any case we also need to replace
the equivalent of #source, which is "part 'uri';" so this hits two birds with
one stone. Eventually this should go away when we load the sources directly from
disk for dart_no_snapshot and gen_snapshot. We still will need to find a way to
load the non-snapshotted sources at that point though.

Somebody with better regular expression understanding than me can certainly
refine the above test. Currently it seems to be safe.

[Ivan Posva, 2012-11-27 05:40:53]

On 2012/11/27 05:38:19, Ivan Posva wrote:
>
https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
> File runtime/tools/concat_library.py (right):
> 
>
https://chromiumcodereview.appspot.com/11411176/diff/1/runtime/tools/concat_l...
> runtime/tools/concat_library.py:37: if line.startswith('#source') or
> line.startswith('part '):
> On 2012/11/27 02:18:07, siva wrote:
> > There is the danger of legal lines which start with part being removed here:
> > e.g:
> > func1(
> > param1 ,
> > part ,
> > sum);
> > 
> > Also, is it a requirement that 'part' has to be the first word without any
> > whitespace?
> > e.g:
> >    part of deepshit;  // Is this legal.
> > 
> > why not do:
> > line = re.sub("part of .*;", "", line);
> 
> I understand all of these limitations (they went away with the new syntax
which
> is easier on the eyes and harder for tools). In any case we also need to
replace
> the equivalent of #source, which is "part 'uri';" so this hits two birds with
> one stone. Eventually this should go away when we load the sources directly
from
> disk for dart_no_snapshot and gen_snapshot. We still will need to find a way
to
> load the non-snapshotted sources at that point though.
> 
> Somebody with better regular expression understanding than me can certainly
> refine the above test. Currently it seems to be safe.

TBR so Lasse can make forward progress.