Sounds good to me.
On 03 Mar 2008 02:25:49 +0300, Egor Pasko <[EMAIL PROTECTED]> wrote: > On the 0x3FA day of Apache Harmony Nathan Beyer wrote: > > > It's still too easy to make a mistake. The best bet is get the RI's > > fixed and get the build to force the correct thing. > > > > Yes, I do use svn diff, but not on generated code, which is what the > > HTML files are. In the future, all I'll do is just NOT change these > > pages. > > As a probably convenient alternative, we might require the doc build > to run under Harmony, so that it breaks otherwise. That way it is not > too easy to break anything. The error message could say something like: > > "Building website is not supported on runtimes other than > Harmony. Please, set your JAVA_HOME accordingly. You can download the > latest Harmony release here: <...>" > > How does it sound? > > > > > -Nathan > > > > On 01 Mar 2008 23:04:21 +0300, Egor Pasko <[EMAIL PROTECTED]> wrote: > > > On the 0x3FA day of Apache Harmony Nathan Beyer wrote: > > > > > > > Here's the change I made - > > > > http://svn.apache.org/viewvc?view=rev&revision=628236. > > > > > > yes, you can see from the patch that Russian text has changed in many > > > places where was not intended to. So, "svn diff" helped to see > > > that. This way I kind of proposed the way how one can make sure if > > > their changes is OK for Russian. Using this technique it is not "too > > > > > > easy for people to innocently do the wrong thing here". > > > > > > So I do not see the reason of making it completely automatic (since > > > looking at svn diff is generally a good practice). Automation would be > > > like put a non-visible test string to the header of each Russian page > > > and verify that the diff does not contain it. Is it easy? > > > > > > But the situation is awkward, I agree. > > > > > > > > > > > > > On 01 Mar 2008 12:14:11 +0300, Egor Pasko <[EMAIL PROTECTED]> wrote: > > > > > On the 0x3FA day of Apache Harmony Nathan Beyer wrote: > > > > > > This happened to me. I don't know Russian, but when I updated > the site > > > > > > for the changes to the ant targets, I figured it was fine to > change > > > > > > the bit that said "ant XXXX". > > > > > > > > > > Natan, I did not understand this phrase quite well. Would you, plz, > > > > > put it in other words so I could understand what happened? I am so > > > > > sorry :) > > > > > > > > > > > > > > > > -Nathan > > > > > > > > > > > > On 29 Feb 2008 23:09:31 +0300, Egor Pasko <[EMAIL PROTECTED]> > wrote: > > > > > > > > > > > > > > On the 0x3F7 day of Apache Harmony Tim Ellison wrote: > > > > > > > > Gregory Shimansky wrote: > > > > > > > > > Alexei Zakharov said the following on 22.02.2008 13:48: > > > > > > > > >> Hm, I can build Russian pages without any corrupted > symbols in it > > > > > > > > >> using RI on Win32. What this bug is about? > > > > > > > > > This is a strange bug and it manifests differently on > different > > > > > > > > > windows installations. I think it also depends on your > system > > > > > > > > > language settings on windows. The bug is in wrong > generation of > > > > > > > > > UTF-8 characters for some of Cyrillic letters. See [1] and > > > > > > > > > predecessors for more details. > > > > > > > > > [1] http://harmony.markmail.org/message/lcjnldxiisawfzxq > > > > > > > > > > > > > > > > Is there any way we can put a check in the build script to > ensure the > > > > > > > > Russian text has not been broken? Even if it is just > checking that > > > > > > > > you are running on a Harmony JRE. > > > > > > > > > > > > > > > > It sounds too easy for people to innocently do the wrong > thing here. > > > > > > > > > > > > > > why? > > > > > > > > > > > > > > if you do not know Russian, you probably would not edit it. > So, "svn > > > > > > > diff" would tell you if characters in Russian text changed. > And if > > > > > > > they did, hey, the wrong thing was about to get in :) > > > > > > > > > > > > > > -- > > > > > > > Egor Pasko > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > Egor Pasko > > > > > > > > > > > > > > > > > > > > -- > > > Egor Pasko > > > > > > > > > > -- > Egor Pasko > >
