Closed Bug 458917 Opened 11 years ago Closed 11 years ago

Windows CE Cleanup


(Core :: General, defect, critical)

Windows CE
Not set





(Reporter: dougt, Assigned: dougt)



(2 files)

Attached patch patch v.1Splinter Review
This patch remove some #ifndef WINCE code across the tree.  Basically, with the latest SDK's there is more stuff available to us as developers allowing us to more fully implement the mozilla platform.

1) disabling a hack in xpconnect that forced wince to build as debug (backing out bug 304714)

2) time is supported, so we can remove the working around in nsSimplePageSequenceFrame.

3) removing MAX_REFLOW_DEPTH workaround in nsIHTMLContentSink

4) Permissions are basically a noop now on windows mobile using the nsLocalFile.  So, we can remove a SetPermissions() call from the profile directory service profile.

5) in the RDF file system data source, we can treat file urls similar to unix (windows mobile doesn't have drive letters)

6) We have fixed up the registry code in xpcom a while, but forget to enable it. (WindowsRegKey and friends)

7) GetTempPathW is supported on windows mobile now -- no need for a special wince case

8) MOZ_USER_DIR is available during a windows mobile build, so no need to have a special DEFAULT_PRODUCT_DIR case for windows mobile

9)  No need to special case the profile directory under windows mobile

10) fixed up the drive enumerator to return TRUE the first time HasMoreElements is called.  This allows people do to:

while(has more elements)
  get element
  do something with element

Dispite all of these changes, the patch is easy to follow.
Attachment #342101 - Attachment is patch: true
Attachment #342101 - Attachment mime type: application/octet-stream → text/plain
Attachment #342101 - Flags: superreview?(pavlov)
Attachment #342101 - Flags: review?(blassey)
Comment on attachment 342101 [details] [diff] [review]
patch v.1

bsmedberg: can you look at the nsLocalFile changes here? the rest looks ok
Attachment #342101 - Flags: superreview?(pavlov)
Attachment #342101 - Flags: superreview?(benjamin)
Attachment #342101 - Flags: review?(blassey)
Attachment #342101 - Flags: review+
Comment on attachment 342101 [details] [diff] [review]
patch v.1

The intent of the drive-enumerator changes is to return an empty enumerator, right? If so, please ifdef all the drive-enumerator code away and just use NS_NewEmptyEnumerator on WINCE... or re-request review if I'm misreading.

sr=me for all the rest of it, if you want to land that pending the drive issue
Attachment #342101 - Flags: superreview?(benjamin) → superreview-
bsmedberg -- na, i need to drive enumerator to return / for the first time it is asked.
Comment on attachment 342101 [details] [diff] [review]
patch v.1

Can you use a NS_NewSingletonEnumerator? It seems really odd to share the nsDriveEnumerator class between windows/WINCE since they do entirely different things.
Attached patch patch v.2Splinter Review
same patch as above, but i have ifdef out the drive enumerator on wince -- we don't have drives at all.

as a side note -- this is a good example of overuse of COM.  No where outside of this file is the drive enumerator class accessed, yet we use xpcom.
Attachment #342149 - Flags: superreview?(benjamin)
File a followup bug on deCOMtaminating that then?
Attachment #342149 - Flags: superreview?(benjamin) → superreview+
66d756c2c44d	Doug Turner — Bug 458917 - Windows CE Cleanup. r/sr=stuart+bsmedberg
Closed: 11 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.