docs/metadata.edge.example - Issue 29345751: Issue 4028 - Add support for Edge extensions to buildtools

Delta Between Two Patch Sets: docs/metadata.edge.example

Issue 29345751: Issue 4028 - Add support for Edge extensions to buildtools (Closed)

Left Patch Set: Address comments on patch set 6 Created July 13, 2016, 10:33 a.m.

Right Patch Set: Address Windows Store issues with blockmap and devbuild display name Created Oct. 14, 2016, 12:07 p.m.

Left:
Right:

Use n/p to move between diff chunks; N/P to move between comments.

Jump to:

Left: Side by side diff | Download
Right: Side by side diff | Download

LEFT	RIGHT
1 [default]	1 [default]

2 # Base metadata file whose sections and keys we inherit but can override.	2 # Base metadata file whose sections and keys we inherit but can override.

3 inherit = metadata.common	3 inherit = metadata.common

4

5 # App identity assigned by the Windows Dev Center.

6 [package_identity]

7 name = ACME.MyExt
Sebastian Noack 2016/07/13 18:10:37 I just noticed, this setting is redundant. The par I just noticed, this setting is redundant. The part before the dot is redundant with "author" in the "general" section. And the part after the dot can be retrieved from "name" message in the source translation. Just remove any spaces. Vasily Kuznetsov 2016/10/04 14:46:45 If I understand correctly this thing is more like Show quoted text On 2016/07/13 18:10:37, Sebastian Noack wrote: > I just noticed, this setting is redundant. The part before the dot is redundant > with "author" in the "general" section. And the part after the dot can be > retrieved from "name" message in the source translation. Just remove any spaces. If I understand correctly this thing is more like an id. I think it might be useful to have it here explicitly so it doesn't get messed up because of changes to UI text. Sebastian Noack 2016/10/04 19:58:03 Frankly, I'm not sure whether the Windows Store wo Show quoted text On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > I just noticed, this setting is redundant. The part before the dot is > redundant > > with "author" in the "general" section. And the part after the dot can be > > retrieved from "name" message in the source translation. Just remove any > spaces. > > If I understand correctly this thing is more like an id. I think it might be > useful to have it here explicitly so it doesn't get messed up because of changes > to UI text. Frankly, I'm not sure whether the Windows Store would except a build that uses a different name than has been reserved on the store and is the part of this generated ID. Same for the author. If it in fact doesn't have to be the same, I see your point. But even then one could argue, that it would be most likely an accident if it's inconsistent, and therefore good to require it to be consistent. Also I wonder whether it's worth to have a separate [package_identity] section. For reference, on Firefox we also have a unique ID and it is under [general]. And "publisher" is just the unique ID for the "author" which is already under [general]. What do you think? Vasily Kuznetsov 2016/10/11 15:58:26 Probably being consistent with Firefox is a good t Show quoted text On 2016/10/04 19:58:03, Sebastian Noack wrote: > On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > > I just noticed, this setting is redundant. The part before the dot is > > redundant > > > with "author" in the "general" section. And the part after the dot can be > > > retrieved from "name" message in the source translation. Just remove any > > spaces. > > > > If I understand correctly this thing is more like an id. I think it might be > > useful to have it here explicitly so it doesn't get messed up because of > changes > > to UI text. > > Frankly, I'm not sure whether the Windows Store would except a build that uses a > different name than has been reserved on the store and is the part of this > generated ID. Same for the author. > > If it in fact doesn't have to be the same, I see your point. But even then one > could argue, that it would be most likely an accident if it's inconsistent, and > therefore good to require it to be consistent. > > Also I wonder whether it's worth to have a separate [package_identity] section. > For reference, on Firefox we also have a unique ID and it is under [general]. > And "publisher" is just the unique ID for the "author" which is already under > [general]. > > What do you think? Probably being consistent with Firefox is a good thing. I moved `name` and `publisher` into [general] renaming them to `app_id` and `publisher_id` (that's what they really are). I also added a warning in case the `app_id` seems mismatched with `author` (from the metadata) and `display_name` (from the default translation). Sebastian Noack 2016/10/12 11:07:48 Fair enough, for keeping the app ID, rather then a Show quoted text On 2016/10/11 15:58:26, Vasily Kuznetsov wrote: > On 2016/10/04 19:58:03, Sebastian Noack wrote: > > On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > > > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > > > I just noticed, this setting is redundant. The part before the dot is > > > redundant > > > > with "author" in the "general" section. And the part after the dot can be > > > > retrieved from "name" message in the source translation. Just remove any > > > spaces. > > > > > > If I understand correctly this thing is more like an id. I think it might be > > > useful to have it here explicitly so it doesn't get messed up because of > > changes > > > to UI text. > > > > Frankly, I'm not sure whether the Windows Store would except a build that uses > a > > different name than has been reserved on the store and is the part of this > > generated ID. Same for the author. > > > > If it in fact doesn't have to be the same, I see your point. But even then one > > could argue, that it would be most likely an accident if it's inconsistent, > and > > therefore good to require it to be consistent. > > > > Also I wonder whether it's worth to have a separate [package_identity] > section. > > For reference, on Firefox we also have a unique ID and it is under [general]. > > And "publisher" is just the unique ID for the "author" which is already under > > [general]. > > > > What do you think? > > Probably being consistent with Firefox is a good thing. I moved `name` and > `publisher` into [general] renaming them to `app_id` and `publisher_id` (that's > what they really are). > > I also added a warning in case the `app_id` seems mismatched with `author` (from > the metadata) and `display_name` (from the default translation). Fair enough, for keeping the app ID, rather then auto generating it. Not sure whether the warning is worth the extra complexity though. Vasily Kuznetsov 2016/10/12 14:13:42 Yeah, me neither, actually. I coded it just so we Show quoted text On 2016/10/12 11:07:48, Sebastian Noack wrote: > On 2016/10/11 15:58:26, Vasily Kuznetsov wrote: > > On 2016/10/04 19:58:03, Sebastian Noack wrote: > > > On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > > > > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > > > > I just noticed, this setting is redundant. The part before the dot is > > > > redundant > > > > > with "author" in the "general" section. And the part after the dot can > be > > > > > retrieved from "name" message in the source translation. Just remove any > > > > spaces. > > > > > > > > If I understand correctly this thing is more like an id. I think it might > be > > > > useful to have it here explicitly so it doesn't get messed up because of > > > changes > > > > to UI text. > > > > > > Frankly, I'm not sure whether the Windows Store would except a build that > uses > > a > > > different name than has been reserved on the store and is the part of this > > > generated ID. Same for the author. > > > > > > If it in fact doesn't have to be the same, I see your point. But even then > one > > > could argue, that it would be most likely an accident if it's inconsistent, > > and > > > therefore good to require it to be consistent. > > > > > > Also I wonder whether it's worth to have a separate [package_identity] > > section. > > > For reference, on Firefox we also have a unique ID and it is under > [general]. > > > And "publisher" is just the unique ID for the "author" which is already > under > > > [general]. > > > > > > What do you think? > > > > Probably being consistent with Firefox is a good thing. I moved `name` and > > `publisher` into [general] renaming them to `app_id` and `publisher_id` > (that's > > what they really are). > > > > I also added a warning in case the `app_id` seems mismatched with `author` > (from > > the metadata) and `display_name` (from the default translation). > > Fair enough, for keeping the app ID, rather then auto generating it. Not sure > whether the warning is worth the extra complexity though. Yeah, me neither, actually. I coded it just so we could see how it would look. Happy to remove it if you think it's not worth it. Sebastian Noack 2016/10/12 14:43:25 I leave it up to you, but FWIW, no I don't think i Show quoted text On 2016/10/12 14:13:42, Vasily Kuznetsov wrote: > On 2016/10/12 11:07:48, Sebastian Noack wrote: > > On 2016/10/11 15:58:26, Vasily Kuznetsov wrote: > > > On 2016/10/04 19:58:03, Sebastian Noack wrote: > > > > On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > > > > > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > > > > > I just noticed, this setting is redundant. The part before the dot is > > > > > redundant > > > > > > with "author" in the "general" section. And the part after the dot can > > be > > > > > > retrieved from "name" message in the source translation. Just remove > any > > > > > spaces. > > > > > > > > > > If I understand correctly this thing is more like an id. I think it > might > > be > > > > > useful to have it here explicitly so it doesn't get messed up because of > > > > changes > > > > > to UI text. > > > > > > > > Frankly, I'm not sure whether the Windows Store would except a build that > > uses > > > a > > > > different name than has been reserved on the store and is the part of this > > > > generated ID. Same for the author. > > > > > > > > If it in fact doesn't have to be the same, I see your point. But even then > > one > > > > could argue, that it would be most likely an accident if it's > inconsistent, > > > and > > > > therefore good to require it to be consistent. > > > > > > > > Also I wonder whether it's worth to have a separate [package_identity] > > > section. > > > > For reference, on Firefox we also have a unique ID and it is under > > [general]. > > > > And "publisher" is just the unique ID for the "author" which is already > > under > > > > [general]. > > > > > > > > What do you think? > > > > > > Probably being consistent with Firefox is a good thing. I moved `name` and > > > `publisher` into [general] renaming them to `app_id` and `publisher_id` > > (that's > > > what they really are). > > > > > > I also added a warning in case the `app_id` seems mismatched with `author` > > (from > > > the metadata) and `display_name` (from the default translation). > > > > Fair enough, for keeping the app ID, rather then auto generating it. Not sure > > whether the warning is worth the extra complexity though. > > Yeah, me neither, actually. I coded it just so we could see how it would look. > Happy to remove it if you think it's not worth it. I leave it up to you, but FWIW, no I don't think it's worth it. Vasily Kuznetsov 2016/10/13 11:56:35 Me neither. Done Show quoted text On 2016/10/12 14:43:25, Sebastian Noack wrote: > On 2016/10/12 14:13:42, Vasily Kuznetsov wrote: > > On 2016/10/12 11:07:48, Sebastian Noack wrote: > > > On 2016/10/11 15:58:26, Vasily Kuznetsov wrote: > > > > On 2016/10/04 19:58:03, Sebastian Noack wrote: > > > > > On 2016/10/04 14:46:45, Vasily Kuznetsov wrote: > > > > > > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > > > > > > I just noticed, this setting is redundant. The part before the dot > is > > > > > > redundant > > > > > > > with "author" in the "general" section. And the part after the dot > can > > > be > > > > > > > retrieved from "name" message in the source translation. Just remove > > any > > > > > > spaces. > > > > > > > > > > > > If I understand correctly this thing is more like an id. I think it > > might > > > be > > > > > > useful to have it here explicitly so it doesn't get messed up because > of > > > > > changes > > > > > > to UI text. > > > > > > > > > > Frankly, I'm not sure whether the Windows Store would except a build > that > > > uses > > > > a > > > > > different name than has been reserved on the store and is the part of > this > > > > > generated ID. Same for the author. > > > > > > > > > > If it in fact doesn't have to be the same, I see your point. But even > then > > > one > > > > > could argue, that it would be most likely an accident if it's > > inconsistent, > > > > and > > > > > therefore good to require it to be consistent. > > > > > > > > > > Also I wonder whether it's worth to have a separate [package_identity] > > > > section. > > > > > For reference, on Firefox we also have a unique ID and it is under > > > [general]. > > > > > And "publisher" is just the unique ID for the "author" which is already > > > under > > > > > [general]. > > > > > > > > > > What do you think? > > > > > > > > Probably being consistent with Firefox is a good thing. I moved `name` and > > > > `publisher` into [general] renaming them to `app_id` and `publisher_id` > > > (that's > > > > what they really are). > > > > > > > > I also added a warning in case the `app_id` seems mismatched with `author` > > > (from > > > > the metadata) and `display_name` (from the default translation). > > > > > > Fair enough, for keeping the app ID, rather then auto generating it. Not > sure > > > whether the warning is worth the extra complexity though. > > > > Yeah, me neither, actually. I coded it just so we could see how it would look. > > Happy to remove it if you think it's not worth it. > > I leave it up to you, but FWIW, no I don't think it's worth it. Me neither. Done
8 publisher = CN=4F066043-8AFE-41C9-B762-6C15E77E3F88

9	4

10 # Icons to be placed into /Assets and referenced in AppxManifest.xml.	5 # Icons to be placed into /Assets and referenced in AppxManifest.xml.

11 [appx_assets]	6 [appx_assets]

12 # 50x50 logo for the store.	7 # 50x50 logo for the store.

13 logo_50.png = path/to/logo_50.png	8 logo_50.png = path/to/logo_50.png

14	9

15 # 44x44 and 150x150 logos for Windows UI.	10 # 44x44 and 150x150 logos for Windows UI.

16 logo_44.png = path/to/logo_44.png	11 logo_44.png = path/to/logo_44.png

17 logo_150.png = path/to/big_logo.png	12 logo_150.png = path/to/big_logo.png

18	13

19 [compat]	14 [compat]

20 # Target Edge version (for manifest.json)	15 # Target Edge version (for manifest.json)

21 edge = 37.14332.1000.0	16 edge = 37.14332.1000.0

22 # Min/max windows version (for AppxManifest.xml)	17 # Min/max windows version (for AppxManifest.xml)

23 windows = 10.0.14332.0/12.0.0.0	18 windows = 10.0.14332.0/12.0.0.0

24	19

25 [general]	20 [general]

	21 # App identity and publisher ID assigned by the Windows Dev Center.

	22 app_id = ACME.MyExt

	23 publisher_id = CN=4F066043-8AFE-41C9-B762-6C15E77E3F88

	24

26 # Base name of the plugin file (version and extension will be added to this).	25 # Base name of the plugin file (version and extension will be added to this).

27 basename = myext-edge	26 basename = myext-edge

28	27

29 # Display name and description for the manifest files.

30 display_name = MyExt

31 description = MyExt is the greatest extension
Sebastian Noack 2016/07/13 18:10:38 This is redundant as well. The description can be This is redundant as well. The description can be retrieved from the source translation. Vasily Kuznetsov 2016/10/04 14:46:44 Done Show quoted text On 2016/07/13 18:10:38, Sebastian Noack wrote: > This is redundant as well. The description can be retrieved from the source > translation. Done
32

33 # Background color for the Appx manifest.	28 # Background color for the Appx manifest.

34 background_color = pink	29 background_color = pink

35

36 # Extension ID for Appx manifest.

37 extension_id = 1.0
Sebastian Noack 2016/07/13 18:10:37 This can be hard-coded to 1.0 in the template. Sin This can be hard-coded to 1.0 in the template. Since there is only one entry point it doesn't have any effect anyway. Sebastian Noack 2016/09/13 10:41:23 This attribute just changed, for Edge extensions i Show quoted text On 2016/07/13 18:10:37, Sebastian Noack wrote: > This can be hard-coded to 1.0 in the template. Since there is only one entry > point it doesn't have any effect anyway. This attribute just changed, for Edge extensions it has to be set to "EdgeExtension", rather than being floating point number, now. Even less useful it is to have it configurable. Vasily Kuznetsov 2016/10/04 14:46:44 Done. Show quoted text On 2016/09/13 10:41:23, Sebastian Noack wrote: > On 2016/07/13 18:10:37, Sebastian Noack wrote: > > This can be hard-coded to 1.0 in the template. Since there is only one entry > > point it doesn't have any effect anyway. > > This attribute just changed, for Edge extensions it has to be set to > "EdgeExtension", rather than being floating point number, now. Even less useful > it is to have it configurable. Done.
38	30

39 # Permissions of the extension to place into manifest.json	31 # Permissions of the extension to place into manifest.json

40 # (see https://developer.chrome.com/extensions/declare_permissions).	32 # (see https://developer.chrome.com/extensions/declare_permissions).

41 permissions =	33 permissions =

42 tabs	34 tabs

43 bookmarks	35 bookmarks

44 http:///	36 http:///

45 https:///	37 https:///

46	38

47 # UI definitions.	39 # UI definitions.

48 options = options.html	40 options = options.html

49 devtools = devtools.html

50	41

51 # Icon for the browser toolbar. Used to produce browser_action key in	42 # Icon for the browser toolbar. Used to produce browser_action key in

52 # manifest.json (see https://developer.chrome.com/extensions/browserAction).	43 # manifest.json (see https://developer.chrome.com/extensions/browserAction).

53 browserAction = icons/icon-19.png icons/icon-38.png popup.html	44 browserAction = icons/icon-19.png icons/icon-38.png popup.html

54	45

55 # Icons for manifest.json.	46 # Icons for manifest.json.

56 icons = icon-16.png icon-32.png icon-48.png icon-64.png icon-128.png	47 icons = icon-16.png icon-32.png icon-48.png icon-64.png icon-128.png

57

58 # Schema for managed storage (storage.managed_schema in manifest.json).

59 managedStorageSchema = schema.json
Sebastian Noack 2016/07/13 18:10:37 Since Edge doesn't support managed storage (the ke Since Edge doesn't support managed storage (the key in the generated manifest.json is ignored) it's rather misleading to document it here. Same for the devtools option above. Or perhaps, just change the name of this file to be platform agnostic. Vasily Kuznetsov 2016/10/04 14:46:45 Making this example platform agnostic is more comp Show quoted text On 2016/07/13 18:10:37, Sebastian Noack wrote: > Since Edge doesn't support managed storage (the key in the generated > manifest.json is ignored) it's rather misleading to document it here. Same for > the devtools option above. Or perhaps, just change the name of this file to be > platform agnostic. Making this example platform agnostic is more complicated since different metadata files have different keys. I'll keep it simple and remove the redundant stuff. Done.
60	48

61 # Files that are renamed/moved when placed into the package.	49 # Files that are renamed/moved when placed into the package.

62 [mapping]	50 [mapping]

63 target/path/file.png = source/path/icon.png	51 target/path/file.png = source/path/icon.png

64	52

65 # Javascript source files to be compiled/bundled by JSHydra.	53 # Javascript source files to be compiled/bundled by JSHydra.

66 [convert_js]	54 [convert_js]

67 # Two files bundled into one.	55 # Two files bundled into one.

68 path/bundle.js = source/path/src1.js src2.js	56 path/bundle.js = source/path/src1.js src2.js

69	57

70 # Remove a file from a bundle defined in common.metadata.	58 # Remove a file from a bundle defined in common.metadata.

71 path/common-bundle.js -= removed/src.js	59 path/common-bundle.js -= removed/src.js

LEFT	RIGHT