LGTM.
On Sat, Sep 20, 2008 at 2:17 AM, Freeland Abbott <[EMAIL PROTECTED]
> wrote:
> r3669.... which promptly broke the build. :-/ Turns out the obsolete
> svnversion on that machine demands an argument, instead of assuming it's run
> in the workspace.
> TBR r3670 for Scott to review:
>
> Index: build-tools/ant-gwt/src/com/google/gwt/ant/taskdefs/SvnInfo.java
>> ===================================================================
>> --- build-tools/ant-gwt/src/com/google/gwt/ant/taskdefs/SvnInfo.java
>> (revision 3669)
>> +++ build-tools/ant-gwt/src/com/google/gwt/ant/taskdefs/SvnInfo.java
>> (working copy)
>> @@ -123,7 +123,7 @@
>> private String getSvnVersion(File workdir) {
>> String line = null;
>>
>> - LineNumberReader svnout = runCommand(workdir, "svnversion");
>> + LineNumberReader svnout = runCommand(workdir, "svnversion", ".");
>> try {
>> line = svnout.readLine();
>> } catch (IOException e) {
>>
>
>
>
>>
>>
>> On Thu, Sep 18, 2008 at 2:27 PM, Scott Blum <[EMAIL PROTECTED]> wrote:
>>
>>> LGTM, although I only skimmed SvnInfo.java. I would have thought,
>>> however, that Ant would have already had really good process launching
>>> infrastructure. I'm surprised you had to reinvent the wheel with
>>> runCommand().
>>> On Thu, Sep 18, 2008 at 12:01 PM, Freeland Abbott <
>>> [EMAIL PROTECTED]> wrote:
>>>
>>>> It seems that getting ant 1.6.5 (which, although old, is still the basis
>>>> of what's in our tools stuff) is known not to like my <propertyregex>
>>>> forms,
>>>> so here's a rework of that as a custom AntTask. Scott will be glad to know
>>>> that the build file stuff we thrashed through isn't changed, it's just a
>>>> reimplementation of <gwt.svninfo>.
>>>> Users of obsolescent tools can now be happy! (More importantly, we're
>>>> self-consistent again, and I think that common.ant.xml is much clearer.)
>>>>
>>>
>>>
>>
>
--~--~---------~--~----~------------~-------~--~----~
http://groups.google.com/group/Google-Web-Toolkit-Contributors
-~----------~----~----~----~------~----~------~--~---
