linux: don't always print dlopen errors from LoadNativeLibrary

linux: don't always print dlopen errors from LoadNativeLibrary

Instead, return them to the caller and let the caller decide whether
the error is worth notifying the user about.

BUG=79068
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=82008

[awong, 2011-04-15 18:57:01]

adding more yaks...

http://codereview.chromium.org/6864020/diff/1/base/native_library_mac.mm
File base/native_library_mac.mm (right):

http://codereview.chromium.org/6864020/diff/1/base/native_library_mac.mm#newc...
base/native_library_mac.mm:25: if (!dylib)
Shouldn't this also grab dlerror()?

The mac library loading API is posix afaik.

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc
File base/native_library_win.cc (right):

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc#newc...
base/native_library_win.cc:44: return LoadNativeLibraryHelper(library_path,
LoadLibraryW);
If the return is NULL, populate with Populate with result of
FormatString(GetLastError()), or at least GetLastError()?

Maybe do this in the Helper function?

BTW, why in the world do we botther with a Helper function here?

[Evan Martin, 2011-04-15 19:00:04]

http://codereview.chromium.org/6864020/diff/1/base/native_library_mac.mm
File base/native_library_mac.mm (right):

http://codereview.chromium.org/6864020/diff/1/base/native_library_mac.mm#newc...
base/native_library_mac.mm:25: if (!dylib)
On 2011/04/15 18:57:01, awong wrote:
> Shouldn't this also grab dlerror()?
> 
> The mac library loading API is posix afaik.

I figured I wouldn't try to change behavior.  And I wasn't sure how to get the
error for the below code branches.

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc
File base/native_library_win.cc (right):

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc#newc...
base/native_library_win.cc:44: return LoadNativeLibraryHelper(library_path,
LoadLibraryW);
On 2011/04/15 18:57:01, awong wrote:
> If the return is NULL, populate with Populate with result of
> FormatString(GetLastError()), or at least GetLastError()?

(Same argument here as for Mac.)

> 
> Maybe do this in the Helper function?
> 
> BTW, why in the world do we botther with a Helper function here?

Probably to share code with LoadNativeLibraryDynamically.

[awong, 2011-04-15 19:03:56]

LGTM

since we're preserving behavior.

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc
File base/native_library_win.cc (right):

http://codereview.chromium.org/6864020/diff/1/base/native_library_win.cc#newc...
base/native_library_win.cc:44: return LoadNativeLibraryHelper(library_path,
LoadLibraryW);
On 2011/04/15 19:00:04, Evan Martin wrote:
> On 2011/04/15 18:57:01, awong wrote:
> > If the return is NULL, populate with Populate with result of
> > FormatString(GetLastError()), or at least GetLastError()?
> 
> (Same argument here as for Mac.)
> 
> > 
> > Maybe do this in the Helper function?
> > 
> > BTW, why in the world do we botther with a Helper function here?
> 
> Probably to share code with LoadNativeLibraryDynamically.

Oh duh.

[Evan Martin, 2011-04-18 16:35:23]

I need owners approval from *both* jam and agl

[agl, 2011-04-18 16:46:34]

LGTM

[Evan Martin, 2011-04-18 19:05:27]

timing out jam, trying brett for owners approval

[brettw, 2011-04-18 19:52:00]

LGTM

[jam, 2011-04-18 21:27:23]

On 2011/04/18 19:52:00, brettw wrote:
> LGTM

sorry just saw this, feel free to IM me next time