Re: [boost] regression test report time

From: Fernando Cacciola (fernando_cacciola_at_[hidden])
Date: 2005-06-27 13:32:42

"Rene Rivera" <grafik.list_at_[hidden]> escribió en el mensaje 
news:42C0367B.2080907_at_redshift-software.com...
> Fernando Cacciola wrote:
>> That or I am still somnehow commiting Mac sources!!
>>
>> I had used boost inspect before commiting (with the -crlf option) and the
>> only file reported is the one in inspect used precisely to show this, so 
>> I'm
>> sure I had no such files _before_ commiting.
>
> Good :-) But it usually shows up after one commits, and another updates,
> not before the commit.
>
>> I use TortoiseCVS, and it has options named "Check not Unix sandbox" and
>> "Use Unix line endings", which are Unchecked and Checked respectively. 
>> But
>> these options are not supposed to cause such problems, so if I am 
>> commiting
>> Mac files I can't see how.
>
> Did you really mean:
>
> "Check not Unix sandbox" *off*
> "Use Unix line endings" *on*
>
Hmmm, yes...

> Because if you did, that second one is precisely what's causing the
> problem. If your files have DOS EOLs you can't tell *CVS that you have
> Unix EOLs.
>
Oh well, from its "tool-tip" (its only help), which refers to the "checkout" 
command, I really interpreted that as _just_ meaning that files with Unix 
EOLs are/aren't converted to DOS into my local HD (just like my editor(s) 
would/wouldn't do); but I never thought it would mess up with the commit.

> By the way I also use TCVS (and WinCVS) and I don't have those options.
> I have a single "Sandbox DOS/UNIX:" which I set to "Autodetect (default
> to DOS)". Works without problems.
>
Ha, you might have a new version.. I'll check it out.

BTW, I most likely introduced the problem again.
So I re-commited (with the second check turned off).
I hope it fixes the problem now (for good).

Best

Fernando Cacciola
SciSoft