Can we get a just slightly more verbose commit message? Something along the lines of:
Shave 39% from snapshot size by representing 0 bits by |. On Thu, Oct 20, 2011 at 1:01 PM, <[email protected]> wrote: > > http://codereview.chromium.**org/8344079/diff/1/src/heap.h<http://codereview.chromium.org/8344079/diff/1/src/heap.h> > File src/heap.h (right): > > http://codereview.chromium.**org/8344079/diff/1/src/heap.h#**newcode67<http://codereview.chromium.org/8344079/diff/1/src/heap.h#newcode67> > src/heap.h:67: V(Map, global_property_cell_map, GlobalPropertyCellMap) > \ > On 2011/10/20 10:06:54, Lasse Reichstein wrote: > >> Is this ordering change just due to "popularity", or is there another >> > reason? > >> (In 64-bit code, we generate shorter code for accessing the first 32 >> > elements, > >> in 32-bit code it's the first 64, so this changes which elements that >> > is). > > It's due to popularity in the snapshot. The first 32 are encoded in > just one byte. > > > http://codereview.chromium.**org/8344079/diff/1/src/**serialize.cc<http://codereview.chromium.org/8344079/diff/1/src/serialize.cc> > File src/serialize.cc (right): > > http://codereview.chromium.**org/8344079/diff/1/src/** > serialize.cc#newcode923<http://codereview.chromium.org/8344079/diff/1/src/serialize.cc#newcode923> > src/serialize.cc:923: *current++ = > isolate->heap()->roots_**address()[root_id]; > On 2011/10/20 10:06:54, Lasse Reichstein wrote: > >> The name roots_address suggests that it returns an Address. Could it >> > be called > >> roots_array_start or _origin instead? >> > > Done. > > > http://codereview.chromium.**org/8344079/diff/1/src/** > serialize.cc#newcode1325<http://codereview.chromium.org/8344079/diff/1/src/serialize.cc#newcode1325> > src/serialize.cc:1325: root_index < 0x20 && > On 2011/10/20 10:06:54, Lasse Reichstein wrote: > >> Name the constants. >> > > Done. > > > http://codereview.chromium.**org/8344079/diff/1/src/** > serialize.cc#newcode1435<http://codereview.chromium.org/8344079/diff/1/src/serialize.cc#newcode1435> > src/serialize.cc:1435: if (repeat_count > kMaxRepeats) { > On 2011/10/20 10:06:54, Lasse Reichstein wrote: > >> A generic repeat takes five bytes. Up to 4*kMaxRepeats, it's smaller >> > to do > >> repeated constant repeats. >> > > No, a generic repeat takes 2 bytes if the repeat count is less than 128. > > > http://codereview.chromium.**org/8344079/diff/1/src/**serialize.h<http://codereview.chromium.org/8344079/diff/1/src/serialize.h> > File src/serialize.h (right): > > http://codereview.chromium.**org/8344079/diff/1/src/** > serialize.h#newcode488<http://codereview.chromium.org/8344079/diff/1/src/serialize.h#newcode488> > src/serialize.h:488: void set_root_index_wave_front(int value) { > root_index_wave_front_ = value; } > On 2011/10/20 10:06:54, Lasse Reichstein wrote: > >> Assert that the wave-front moves forward? >> > > Done. > > > http://codereview.chromium.**org/8344079/<http://codereview.chromium.org/8344079/> > > -- > v8-dev mailing list > [email protected] > http://groups.google.com/**group/v8-dev<http://groups.google.com/group/v8-dev> > -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
