|
Boost-Build : |
Subject: Re: [Boost-build] Breakage in Doxygen rule
From: Vladimir Prus (vladimir_at_[hidden])
Date: 2008-12-02 13:32:39
John Maddock wrote:
> Vladimir Prus wrote:
>>>> But there is no such directory, and there never will be (as it's
>>>> created by any of the rules), so the del command fails with:
>>>>
>>>> "The system cannot find the file specified."
>>>
>>> I presume this code is supposed to remove stale XML files -- does
>>> the error happens if you touch any of the headers and run bjam
>>> again? I would expect that in that case, stale XML files be removed.
>
> The error is still present because the directory path *never* exists.
Can you explain? For me, on Linux, we try to remove
../../../bin.v2/libs/program_options/doc/gcc-4.2.4/debug/autodoc-xml/*.xml
and there's a bunch of *.xml files there. The directory is created by doxygen, due
to a couple of lines in its (generated) config file:
XML_OUTPUT = autodoc-xml
OUTPUT_DIRECTORY = "../../../bin.v2/libs/program_options/doc/gcc-4.2.4/debug"
Is this not happening for you? Where are the xml files created?
>>>> Now, this doesn't actually terminate the build (which surprises me
>>>> actually), but the error message is identical to some genuine
>>>> errors emitted by Doxygen if it can't find the external tools it
>>>> needs (like latex etc), which makes it darn hard to figure out
>>>> whether the build was OK or not :-(
>>>
>>> We can redirect that 'del' into 'nul', I think.
>
> On Win32 as well?
into 'nul'? Well, *only* on windows :-) "> nul" will redirect stdout; I don't remember
how to redirect stderr offhand.
- Volodya
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