Subject: Re: [Boost-build] Early bjam abort
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2012-05-01 18:04:04
On 05/01/2012 10:54 AM, Nogradi, Chris wrote:
>>> -> 7 Name: <object(c-scanner)@25349#lib/adl>iop_c_pkt.stc
>>> Loc: bin\lib\adl\msvc-8.0\dbg\lnk-sttc\prcs-IOP-SIM\prjct-GTC\thrd-mlt\iop_c_pkt.stc
>>> : Outdated, updating it
>>> : NOCARE
>>> : Depends on <pbin\lib\adl\msvc-8.0\dbg\lnk-sttc\prcs-IOP-SIM\prjct-GTC\thrd-mlt>iop_c_pkt.stc (stable)
>>> : Depends on <pbin\lib\adl\msvc-8.0\dbg\lnk-sttc\prcs-IOP-SIM\prjct-GTC\thrd-mlt>iop_c_pkt.stc (internal node) (stable) (max time)
>>> Why are these duplicate? And why are they updating?
>> The internal node represents the results of header scanning.
> But both the internal node dependency and the file dependency above are stable, so why is this scanner object outdated? The .stc file is a generated file of type H and it is not regenerated since it is stable nor does it need to be regenerated. But for some reason the scanner determines the file to be outdated which appears to cause the rebuilds that I am seeing.
This is a bug in bjam. There should be
no dependency on the internal node in
this case. It's fairly easy to fix.
I'll get to it tonight.
Anyway, the reason for the update is that
the timestamp on the internal node is
newer than the timestamp on the file.
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk