Re: Building PCC for "tools" is broken (missing symbol __USE)- PCC bug or NetBSD source tree error?

[Iain Hibbert <plunky%ogmig.net@localhost>]

On 31 July 2014 06:13:00 BST, "William D. Jones" 
<thor0505%comcast.net@localhost> wrote:
>Sorry for the delay in getting back. Iain sent me an email suggesting a
>
>checkout like you suggested, or using prepare_import.sh in 
>$NETBSD_ROOT/external/bsd/pcc/. Is prepare_import.sh up to date?

it should be fine for current PCC sources

 Should
>I 
>use that script instead, or are there subtle issues to using that
>script 
>right now as opposed to applying the diff directly (I've read that
>PCC's 
>files are kept in multiple parts of the source tree)?

I don't think that is the case, feel free to apply the diff directly

It seems as if 
>"prepare_import.sh" implements the equivalent of git subtree (track 
>repository from within repositoty).

I use that script to prepare the PCC sources before import to netbsd tree.. you 
can use it locally and get the same result as if I had imported a new version

if you just apply one patch at a time there is always the chance you miss 
something

>
>Additionally, the bottom of prepare_import.sh is confusing me:
>echo ""
>echo "after testing, use the following command to import from the dist 
>directory,"
>echo ""
>echo "    cvs import src/external/bsd/pcc/dist ragge pcc-${datestamp}"
>echo ""
>echo "providing a ChangeLog in the commit message."
>
>Is that supposed to be your job, Anders, since my username is not
>ragge, and 
>I don't have write access :P.

you can ignore that stuff.. it is a note for somebody who is importing a new 
version, so that over time we are consistent

Iain
(sorry if poor formatting, on my phone)

>
>-----Original Message----- 
>From: Anders Magnusson
>Sent: Saturday, July 26, 2014 3:43 AM
>To: William D. Jones
>Cc: current-users%NetBSD.org@localhost
>Subject: Re: Building PCC for "tools" is broken (missing symbol __USE)-
>PCC 
>bug or NetBSD source tree error?
>
>The commit messages are saved at marc.info, see
>http://marc.info/?l=pcc-commit-list&m=140628065609751&w=2
>
>To get a diff you can use cvs -d
>:pserver:anonymous%pcc.ludd.ltu.se@localhost:/cvsroot rdiff -u -r1.376 -r1.377
>pcc/cc/ccom/cgram.y
>
>Should be directly applyable.
>
>-- Ragge
>
>William D. Jones skrev 2014-07-26 02:05:
>> Do you have a patch that I can apply to my private NetBSD source tree
>of 
>> pcc for the time being? My pcc is compiled into the NetBSD tools 
>> directory, at which point it is used to compile the rest of the
>NetBSD 
>> source tree.
>>
>> -----Original Message----- From: Anders Magnusson
>> Sent: Friday, July 25, 2014 5:41 AM
>> To: William D. Jones ; Iain Hibbert
>> Cc: current-users%NetBSD.org@localhost
>> Subject: Re: Building PCC for "tools" is broken (missing symbol
>__USE)- 
>> PCC bug or NetBSD source tree error?
>>
>> William D. Jones skrev 2014-07-25 07:58:
>>>> Yes, I think you can set UNSUPPORTED_COMPILER.pcc in the Makefile
>and it
>>>> will proceed using a different one
>>>
>>> For the time being, I'll make a note of all the tools that pcc
>cannot
>>> currently build, and set Makefiles accordingly. I'm not sure if
>these
>>> tools under gpl track upstream or not, so at least for tonight I'm
>not
>>> sure if feel comfortable changing the actual source code of binutils
>>> to make pcc compile it successfully. Although perhaps submitting a
>>> patch to binutils that undefines __restrict in arrays if pcc is
>being
>>> used is a possible solution for the time being.
>>>
>> pcc accepted __restrict but not restrict in array declaration.  I
>just
>> added a yacc rule to pcc to accept both (to master tree).
>>
>> -- Ragge
>>
>> -- 
>> William D. Jones
>
>--
>William D. Jones
>Rowan University | ECE | 2012
>Member IEEE
>Member Tau Beta Pi
>thor0505%comcast.net@localhost
>Message sent using 'Windows Live Mail' client. 

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.