Talk:Gcc issues in Intrepid and later: Difference between revisions

From Openembedded.org
Jump to navigation Jump to search
No edit summary
No edit summary
 
(One intermediate revision by the same user not shown)
Line 15: Line 15:
: Surely if it doesn't build with 4.4, the fix should not be "install and use 4.1", it should be "the fix has been committed".
: Surely if it doesn't build with 4.4, the fix should not be "install and use 4.1", it should be "the fix has been committed".


:: I agree that the proper solution would be "the fix has been committed".  The question is who makes, tests and provides us with this patch?  Most people are not being paid for what they do with OE, so I guess the only sensible answer to that would be "those affected".  Not all devs are using gcc4.4.  So, basically you have three options a) downgrade gcc as suggested, b) prepare the patch yourself (or pay someone to do it), c) wait until somebody else fixes the particular package you want to compile.  I would venture to guess that for the ordinary user the order of preference would be a>b>c --[[User:Laibsch|Laibsch]] 16:36, 18 January 2010 (UTC)
:: I agree that the proper solution would be "the fix has been committed".  The question is who makes, tests and provides us with this patch?  Most people are not being paid for what they do with OE, so I guess the only sensible answer to that question would be "those affected".  Not all devs are using gcc4.4.  So, basically you have three options a) downgrade gcc as suggested, b) prepare the patch yourself (or pay someone to do it), c) wait until somebody else fixes the particular package you want to compile.  I would venture to guess that for the ordinary user the order of preference would be a>b>c --[[User:Laibsch|Laibsch]] 16:36, 18 January 2010 (UTC)
 
:: Oh, since you asked to be corrected (although it has no effect on the above reasoning), the problem is not in bitbake (which also does not use CVS, but git, you may need to update).  The problems are in the upstream projects, although we can fix them in the OE metadata.  But bitbake has nothing to do with this at all. --[[User:Laibsch|Laibsch]] 16:38, 18 January 2010 (UTC)

Latest revision as of 16:38, 18 January 2010

Is this still an issue with OE?

I'm using 9.04 and was at least able to compile one basic image. I didn't go much further though? --Klausfpga 17:49, 9 September 2009 (UTC)

This is definitely still an issue.
I've tried compiling on a fresh install of latest ubuntu and got all these problems and more.
Correct me if i'm wrong but the whole point of the combination of CVS and bitbake is that it can be kept updated so users don't spend days working on workarounds and fixes, most of which could be fixed with some updated packages or .patch's.
I've spent the last 3 days fixing compiler errors and researching the multiple issues that have plagued the build with gcc 4.4 (the default in new ubuntu).
Surely if it doesn't build with 4.4, the fix should not be "install and use 4.1", it should be "the fix has been committed".
I agree that the proper solution would be "the fix has been committed". The question is who makes, tests and provides us with this patch? Most people are not being paid for what they do with OE, so I guess the only sensible answer to that question would be "those affected". Not all devs are using gcc4.4. So, basically you have three options a) downgrade gcc as suggested, b) prepare the patch yourself (or pay someone to do it), c) wait until somebody else fixes the particular package you want to compile. I would venture to guess that for the ordinary user the order of preference would be a>b>c --Laibsch 16:36, 18 January 2010 (UTC)
Oh, since you asked to be corrected (although it has no effect on the above reasoning), the problem is not in bitbake (which also does not use CVS, but git, you may need to update). The problems are in the upstream projects, although we can fix them in the OE metadata. But bitbake has nothing to do with this at all. --Laibsch 16:38, 18 January 2010 (UTC)