directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Emmanuel Lecharny <elecha...@apache.org>
Subject Re: Empty trunk directories
Date Thu, 24 May 2007 20:53:58 GMT
Enrique Rodriguez a écrit :

> On 5/24/07, Emmanuel Lecharny <elecharny@apache.org> wrote:
>
>> ...
>> JDBM is empty, but we might use it (juts if we decide to extend JDBM
>> features, and it's quite likely to happen)
>
>
> It is trivial to add back and for now it is cruft.

yeah, very true... Let's doom it for clarity then :)

>
>> src can be deleted I think. Let me check
>
>
> Thanks.

shoot !

>
>> Btw, did you read my mail about server.xml pb with your last SASL
>> commit? It would be very appreciated if you can fix the pb I got this
>> morning.
>
>
> Yes, and I committed a fix soon after I got a chance to review email.

Oh. I didn't noticed the commit. As a rule of thumb, I think it's better 
to reply to mails than committing a fix with no reply, because I don't 
read all the commits, when I read all the mails :)

>
>> Although, we can't anymore build the server with IBM JVM, which is
>> painfull as this is the JVM I'm using (it's really the fastest !). Can
>> you find a way to fix it ? (I mean, a workaround would be good right
>> now, a perfect solution can wait a few days :)
>
>
> I am aware of the issue and will get around to it when I can.

The profile solution seems to be the best for a short term solution. 
However, a lot of users are compiling the server with Sun JVM, so it's 
not really a pb.
It was just not very cool this morning to have had to fight against my 
.bashrc after having had to download the last version of Sun JVM... 
Might have been better to have wait a few more hours to be sure that we 
have a working branch for SUN, IBM and JRockit JVM on the branch, 
instead of committing it straight to trunk. I'm sure it was not 
intended, of course, and I don't blame you for that. Just trying to make 
sure that the server is a clean place for new commers ;)

>
>> Thos little pbs (it took me 30 mins to have a building and running
>> server this morning) may generate a lot of mails from users, and if we
>> can avoid such mails, this would be good for us !
>
>
> You must know you are at least 8 hrs ahead of me.  Since I thought we
> all follow "collective code ownership" you could have simply consulted
> with other members of the Directory project and disabled the
> server.xml protocols and the failing tests with the IBM JRE.

This is exactly what I did, but when I wrote my mail, the issue with IBM 
was known, but I was the first to experiment pb with server.xml. You 
can't expect people in europ to consult other member of the community 
ate 8am, Paris time, when those other members are just sleeping in 
USA... It's not about me, I was able to handle the situation in 30 
minutes (error messages weren't crystal clear ...), I'm thinking about 
all the users who are not reading the dev ML and who don't know about 
"collective code ownership". Btw, following this "collective code 
ownership" is a best effort constraint. So is the "don't break the 
build" or "don't make it difficult for user to run the server". Again, I 
don't blame you, I just try to get the project stable, and regarding the 
code size, and all the dependencies, this is not an easy task ... So 
far, we are quite successfull ;)

> Certainly Alex will disable failing tests, as he did with the AES-256
> tests, and leave it for me to deal with when I get around to it.

It's not Alex decision, nor mine, I think you are the best person to 
take this decision. If you don't have time in, say, the next few days, 
then we might disable tests. But I think we can wait two weeks, as there 
are a few folks out there using IBM to build the server (I'm one of 
those :). Just tell us what you think is your timeframe to fix this 
little bump on the road, and we will discuss what to do.

Is that fine with you ?

Thanks a lot for your feedback !

Emmanuel


Mime
View raw message