Add quakespasm, and OpenGL+SDL-based port of quake1.

Add quakespasm, an OpenGL+SDL-based port of quake1.

Add zip.fs to DEPS which is used to extract the quake
data zip file, if its not found alongside the executable.

BUG= http://code.google.com/p/naclports/issues/detail?id=96
R=binji@chromium.org

Committed: https://code.google.com/p/naclports/source/detail?r=1462

[binji, 2014-06-30 19:43:14]

lgtm

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/build.sh
File ports/quakespasm/build.sh (right):

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/build.sh#ne...
ports/quakespasm/build.sh:29: quakespasm${NACL_EXEEXT} -o quake.nmf
quakespasm.{nmf,html}

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/pkg_info
File ports/quakespasm/pkg_info (right):

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/pkg_info#ne...
ports/quakespasm/pkg_info:1: NAME=quakespasm
LICENSE?

[Sam Clegg, 2014-07-18 23:16:45]

Added zip.js and javascript code that handle uploading and
unziping of quake data files to html5 local storage before
launching the game.

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/build.sh
File ports/quakespasm/build.sh (right):

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/build.sh#ne...
ports/quakespasm/build.sh:29: quakespasm${NACL_EXEEXT} -o quake.nmf
On 2014/06/30 19:43:14, binji wrote:
> quakespasm.{nmf,html}

Done.

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/pkg_info
File ports/quakespasm/pkg_info (right):

https://codereview.chromium.org/337323007/diff/1/ports/quakespasm/pkg_info#ne...
ports/quakespasm/pkg_info:1: NAME=quakespasm
On 2014/06/30 19:43:14, binji wrote:
> LICENSE?

Done.

[binji, 2014-07-18 23:48:04]

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:18: var embed = document.createElement('object');
s/object/embed/

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:24: var param = document.createElement('param');
Use

    embed.setAttribute('PWD', pwd);

instead?

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:25: param.name = 'PWD';
seems a bit weird that PWD changes the directory instead of CHDIR or CWD or
something.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:32: function extractZipFile(file, filesystem) {
Yikes! I'd break this up into smaller functions personally. It's pretty hard to
follow now.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:58: runQuake('/tmp');
probably a little more javascript-y to have a callback that is called here. Then
the caller can decide what to do after all files are unzipped.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:78: function(bytes) {
what if the quota isn't granted? I guess the user knows they didn't allow it so
it doesn't really need to notify, hm?

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:79:
window.webkitRequestFileSystem(window.TEMPORARAY, bytes, function(fs) {
sp: TEMPORARY

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:110: window.TEMPORARAY, 3 * 1024 * 1024,
TEMPORARY

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:120: req.onreadystatechange = function() {
I prefer using onload/onerror. onreadystatechange is kinda ugly, IMO

https://codereview.chromium.org/337323007/diff/40001/ports/sdl/nacl.patch
File ports/sdl/nacl.patch (right):

https://codereview.chromium.org/337323007/diff/40001/ports/sdl/nacl.patch#new...
ports/sdl/nacl.patch:684: +  // This could fail if the sdl_mount_http attributes
Hm, this seems like a significant change. Mention it in the CL description?

[Sam Clegg, 2014-07-19 00:59:50]

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:18: var embed = document.createElement('object');
On 2014/07/18 23:48:04, binji wrote:
> s/object/embed/

We use object in naclterm.js .. and it seems that setting embed.data doesn't
work unless I use object here.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:24: var param = document.createElement('param');
On 2014/07/18 23:48:04, binji wrote:
> Use
> 
>     embed.setAttribute('PWD', pwd);
> 
> instead?

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:25: param.name = 'PWD';
On 2014/07/18 23:48:03, binji wrote:
> seems a bit weird that PWD changes the directory instead of CHDIR or CWD or
> something.

PWD is what bash uses in its environment and its what nacl_spawn uses to
communicate CWD.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:32: function extractZipFile(file, filesystem) {
On 2014/07/18 23:48:04, binji wrote:
> Yikes! I'd break this up into smaller functions personally. It's pretty hard
to
> follow now.

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:58: runQuake('/tmp');
On 2014/07/18 23:48:04, binji wrote:
> probably a little more javascript-y to have a callback that is called here.
Then
> the caller can decide what to do after all files are unzipped.

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:78: function(bytes) {
On 2014/07/18 23:48:04, binji wrote:
> what if the quota isn't granted? I guess the user knows they didn't allow it
so
> it doesn't really need to notify, hm?

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:79:
window.webkitRequestFileSystem(window.TEMPORARAY, bytes, function(fs) {
On 2014/07/18 23:48:04, binji wrote:
> sp: TEMPORARY

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:110: window.TEMPORARAY, 3 * 1024 * 1024,
On 2014/07/18 23:48:04, binji wrote:
> TEMPORARY

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:120: req.onreadystatechange = function() {
On 2014/07/18 23:48:04, binji wrote:
> I prefer using onload/onerror. onreadystatechange is kinda ugly, IMO

Done.

https://codereview.chromium.org/337323007/diff/40001/ports/sdl/nacl.patch
File ports/sdl/nacl.patch (right):

https://codereview.chromium.org/337323007/diff/40001/ports/sdl/nacl.patch#new...
ports/sdl/nacl.patch:684: +  // This could fail if the sdl_mount_http attributes
On 2014/07/18 23:48:04, binji wrote:
> Hm, this seems like a significant change. Mention it in the CL description?

Split into separate CL

[binji, 2014-07-21 17:45:49]

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:18: var embed = document.createElement('object');
On 2014/07/19 00:59:49, Sam Clegg wrote:
> On 2014/07/18 23:48:04, binji wrote:
> > s/object/embed/
> 
> We use object in naclterm.js .. and it seems that setting embed.data doesn't
> work unless I use object here.

yeah object w/ a data element is the old way and embed w/ attribute is the new
way. We should probably move toward that, but I guess we can keep it as-is for
now, if object is still being used.

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
ports/quakespasm/quakespasm.js:29: if (i === entries.length) {
this is serialized now, right? (it seems to have been parallelized before) Is it
slower?

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
ports/quakespasm/quakespasm.js:127: } else {
does this work? I thought onload wouldn't fire if the request failed.

[Sam Clegg, 2014-07-21 18:05:28]

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/40001/ports/quakespasm/quakesp...
ports/quakespasm/quakespasm.js:18: var embed = document.createElement('object');
On 2014/07/21 17:45:48, binji wrote:
> On 2014/07/19 00:59:49, Sam Clegg wrote:
> > On 2014/07/18 23:48:04, binji wrote:
> > > s/object/embed/
> > 
> > We use object in naclterm.js .. and it seems that setting embed.data doesn't
> > work unless I use object here.
> 
> yeah object w/ a data element is the old way and embed w/ attribute is the new
> way. We should probably move toward that, but I guess we can keep it as-is for
> now, if object is still being used.

Done.  I figured out how to use the 'embed' element.

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
ports/quakespasm/quakespasm.js:29: if (i === entries.length) {
On 2014/07/21 17:45:49, binji wrote:
> this is serialized now, right? (it seems to have been parallelized before) Is
it
> slower?

There is only one file in this case.  Also it really needs to be serial since
there directories need to be created before the files that live in them.

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
ports/quakespasm/quakespasm.js:127: } else {
On 2014/07/21 17:45:49, binji wrote:
> does this work? I thought onload wouldn't fire if the request failed.

Yes.. noload fires with status == 404.  onerror does not seem to fire at all.

[binji, 2014-07-21 18:07:36]

lgtm

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
File ports/quakespasm/quakespasm.js (right):

https://codereview.chromium.org/337323007/diff/100001/ports/quakespasm/quakes...
ports/quakespasm/quakespasm.js:127: } else {
On 2014/07/21 18:05:27, Sam Clegg wrote:
> On 2014/07/21 17:45:49, binji wrote:
> > does this work? I thought onload wouldn't fire if the request failed.
> 
> Yes.. noload fires with status == 404.  onerror does not seem to fire at all.

Weird.

[Sam Clegg, 2014-07-21 18:11:14]

Committed patchset #7 manually as r1462 (presubmit successful).