Bug 394959 - CDT GCC build output parser does not understand -isystem
Summary: CDT GCC build output parser does not understand -isystem
Status: NEW
Alias: None
Product: CDT
Classification: Tools
Component: cdt-build (show other bugs)
Version: 8.1.0   Edit
Hardware: PC Linux
: P3 normal with 1 vote (vote)
Target Milestone: ---   Edit
Assignee: cdt-build-inbox@eclipse.org CLA
QA Contact: Jonah Graham CLA
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-11-23 04:58 EST by Serge Pavlovsky CLA
Modified: 2020-09-04 15:21 EDT (History)
6 users (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Serge Pavlovsky CLA 2012-11-23 04:58:38 EST 
CDT GCC build output parser does not understand -imacros and -isystem compiler flags
Comment 1 Andrew Gvozdev CLA 2013-07-13 09:33:11 EDT 
I separated -imacros  issue into bug 412897 and fixed it. That was a simple typing error.
Comment 2 Marc-André Laperle CLA 2013-07-13 17:49:41 EDT 
It this a duplicate of bug 293216?
Comment 3 Andrew Gvozdev CLA 2013-07-15 10:37:53 EDT 
(In reply to comment #2)
> It this a duplicate of bug 293216?
This bug is about a new LSP build output parser. Bug 293216 started about older scanner discovery parser, and there is even a patch for it. Perhaps bug 293216 could be closed when this one is fixed but for now I would prefer to keep them separate.
Comment 4 Fabrizio Iannetti CLA 2013-08-14 04:27:05 EDT 
In order to workaround this (-isystem is used a lot in android source makefiles) I have written a small plugin that contributes a Parser copied from GCC with the addition of -isystem.

Of course this is not really right, as these paths are treated the same way as -I includes. A correct handling will require some deeper changes I take.

Would it make sense that I spend time in trying to implement a fix for the GCC parser? Or is a solution already in the making?
Comment 5 Andrew Gvozdev CLA 2013-08-14 09:37:32 EDT 
As far as I know nobody is working on this currently. I see there is a fair number of people asking for this. If you want to implement it I can review your patches. You can start with a patch that treats -isystem as -I, it would be useful already. Please, submit patches to Gerrit.
Comment 6 Johannes Wienke CLA 2013-08-15 05:29:12 EDT 
Can someone please review the patch in bug 293216 I provided under the light of the new discovery infrastructure. I can't completely remember what I did there and also I do not know much about the implementation of CDT, but maybe I already fixed the new discovery?
Comment 7 Andrew Gvozdev CLA 2013-08-15 09:07:41 EDT 
(In reply to comment #6)
> Can someone please review the patch in bug 293216 I provided under the light of
> the new discovery infrastructure. I can't completely remember what I did there
> and also I do not know much about the implementation of CDT, but maybe I already
> fixed the new discovery?
That patch can not possibly work with new scanner discovery, that's what I meant in comment 3. In addition, as mentioned in comment 4, -isystem is more involved than -I.
Comment 8 Fabrizio Iannetti CLA 2013-08-19 02:56:22 EDT 
Hi,

I have uploaded my change, which -isystem as -I, to gerrit:

https://git.eclipse.org/r/#/c/15584/

cheers
Fabrizio
Comment 9 Andrew Gvozdev CLA 2013-08-19 15:15:25 EDT 
I merged your patch to master, thank you. Again, as you mentioned in comment 4 a correct handling requires some deeper changes - to reorder include paths.
Comment 10 Jonah Graham CLA 2019-12-30 17:05:16 EST 
This bug was assigned and targeted at a now released milestone. As that milestone has now passed, the milestone field has been cleared. If this bug has been fixed, please set the milestone to the version it was fixed in and marked the bug as resolved.