Parsing in distcc pump mode (path problems - "No such file or directory")

--===============2052855284397446441==
Content-Type: multipart/alternative; boundary=001a11353046135e240549a9a46a

--001a11353046135e240549a9a46a
Content-Type: text/plain; charset=UTF-8

I am having issues with strange file paths when parsing c file in distcc's
pump-mode (distcc's plain mode works just fine).

I'm running on Cygwin64, in a system with also uses both CMake and Ninja.
The compiler is not GCC.

I have tried tracing the issue in the source but the indexing of paths is
rather complicated and I hoped someone more familiar with it can offer some
hints/tips as to what the problem may be.  As a user I cannot spend too
much time patching code to suit me.

The error messages I receive are in the format:

WARNING include server: Preprocessing locally. Include server not covering:
Parse file: '/cygdrive/c/*<project base>*/*<path to build.ninja directory>*
/C:\*<project base>*\*<path to source>*\example.c': [Errno 2] No such file
or directory: '/cygdrive/c/*<project base>*/*<path to build.ninja
directory>*/C:\\*<project base>*\\*<path to source>*\\example.c' for
translation unit 'C:\*<project base>*\*<path to source>*\example.c'

(*<path to build.ninja directory>* is also the base for the output files)


I should also mention that I made 1 modification to include_server:

In "compiler_defaults.py", function "_SystemSearchdirsGCC":

Removed the function body as it only works for GCC, and replaced with a
single path to the compilers default include path.

Any suggestions are much appreciated.

Thanks in advance!

--001a11353046135e240549a9a46a
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>I am having issues with strange file paths when parsi=
ng c file in distcc&#39;s pump-mode (distcc&#39;s plain mode works just fin=
e).=C2=A0</div><div><br></div><div>I&#39;m running on Cygwin64, in a system=
 with also uses both CMake and Ninja. =C2=A0</div><div>The compiler is not =
GCC.</div><div><br></div><div>I have tried tracing the issue in the source =
but the indexing of paths is rather complicated and I hoped someone more fa=
miliar with it can offer some hints/tips as to what the problem may be.=C2=
=A0 As a user I cannot spend too much time patching code to suit me.</div><=
div><br></div><div>The error messages I receive are in the format:</div><di=
v><br></div><div>WARNING include server: Preprocessing locally. Include ser=
ver not covering: Parse file: &#39;/cygdrive/c/<i>&lt;project base&gt;</i>/=
<i>&lt;path to build.ninja directory&gt;</i>/C:\<i>&lt;project base&gt;</i>=
\<i>&lt;path to source&gt;</i>\example.c&#39;: [Errno 2] No such file or di=
rectory: &#39;/cygdrive/c/<i>&lt;project base&gt;</i>/<i>&lt;path to build.=
ninja directory&gt;</i>/C:\\<i>&lt;project base&gt;</i>\\<i>&lt;path to sou=
rce&gt;</i>\\example.c&#39; for translation unit &#39;C:\<i>&lt;project bas=
e&gt;</i>\<i>&lt;path to source&gt;</i>\example.c&#39;<br></div><div><br></=
div><div>(<i>&lt;path to build.ninja directory&gt;</i>=C2=A0is also the bas=
e for the output files)</div><div><br></div><div><br></div><div>I should al=
so mention that I made 1 modification to include_server:</div><div><br></di=
v><div>In &quot;compiler_defaults.py&quot;, function &quot;_SystemSearchdir=
sGCC&quot;:</div><div><br></div><div>Removed the function body as it only w=
orks for GCC, and replaced with a single path to the compilers default incl=
ude path.</div><div><br></div><div>Any suggestions are much appreciated.</d=
iv><div><br></div><div>Thanks in advance!</div></div>

--001a11353046135e240549a9a46a--


--===============2052855284397446441==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

__
distcc mailing list            http://distcc.samba.org/
To unsubscribe or change options:
https://lists.samba.org/mailman/listinfo/distcc
--===============2052855284397446441==--