Re: [Monotone-devel] continued problems

[Howard Spindel]

At 03:27 PM 9/1/2005, Matthew Gregan wrote:
> At 2005-08-27T22:13:12-0700, Howard Spindel wrote:
>
> > When I try the db load command, I get:
>
> > monotone: error: sqlite error [1]: near
> > "'H4sIAAAAAAAAAOxdCXgcxZXukSxbtuX7AGxsyyCDAWvGJ2DL2NZh2SNLtpHkI7Y4WjMtqa2e
> > monotone: error:
> > 6WG6RxpxJCbhvkM4kkC4A0kgIYGwSTgScyUBcpFgCElISAwE2EAgHAmE4H1V1T3Tx2tpquTd
> > monotone: error:
> > zbc77c+j6emq/6969erVq1fV3S/s37//7Tmfv3nKdyY/HpLaBhJy0lRilVH4TCcVs7Jd1zWj
>
> > followed by hundreds more lines that look similar except the garbagy
> > characters change.  It looks to me like lines that are supposed to be
> > data are getting interpreted as commands.
>
> Thanks for reporting this.  It turned out to be a Win32 (really, MinGW)
> specific problem where std::cin was returning EOF early after the
> standard I/O streams had been unsynchronized with a call to
> sync_with_stdio(false) when feed input that contained DOS newlines.
>
> The correct solution for this needs a fix in MinGW.  For now, I've
> committed a change to mainline (rev. fc87df) to work around this
> problem.

Thank you for looking into that, Matthew.  One question though...

The problem also occurs when running "db load" from a regular WinXP 
command prompt - no MinGW involved.  So, is your fix effective when 
MinGW is not running?

I guess another way to ask this is: When you say it's a MinGW 
problem, did you mean a problem with the MinGW development 
environment or runtime environment?   If the latter, perhaps the 
problem is not fully understood?

Regards,
Howard