Issue 44880 - Solarisgccport: no adjustvisibility patch
Summary: Solarisgccport: no adjustvisibility patch
Status: CLOSED NOT_AN_OOO_ISSUE
Alias: None
Product: porting
Classification: Code
Component: code (show other issues)
Version: 680m85
Hardware: PC Solaris
: P3 Trivial (vote)
Target Milestone: AOO Later
Assignee: taniguchi
QA Contact: issues@porting
URL:
Keywords: needmoreinfo
Depends on:
Blocks:
 
Reported: 2005-03-12 14:55 UTC by taniguchi
Modified: 2007-06-26 05:46 UTC (History)
4 users (show)

See Also:
Issue Type: PATCH
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
pstrules.mk no adjustvisibility patch (4.45 KB, patch)
2005-03-12 14:56 UTC, taniguchi
no flags Details | Diff
rules.mk no adjustvisibility patch (6.87 KB, patch)
2005-03-12 14:56 UTC, taniguchi
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this issue.
Description taniguchi 2005-03-12 14:55:19 UTC
It is not necessary to use adjustvisibility because
  1. gcc doesn't have "fix and continue" function and
  2. debug symbol format for gcc-3.4 or later is dwarf2, not stabs.

These patches disable adjustvisibility for Solaris gcc.
Comment 1 taniguchi 2005-03-12 14:56:11 UTC
Created attachment 23730 [details]
pstrules.mk no adjustvisibility patch
Comment 2 taniguchi 2005-03-12 14:56:53 UTC
Created attachment 23731 [details]
rules.mk no adjustvisibility patch
Comment 3 pavel 2005-03-12 20:18:53 UTC
accept.

ause: BTW, if you look at the patches, don't you think that this whole concept
of ADJUSTVISIBILITY is worth some magic/unification. It looks like these two
patches are long only because it is not well-thought in the source. In ideal
world, both patches would be reduced to two lines in one file.
Comment 4 taniguchi 2005-03-13 05:10:09 UTC
Hi, Pavel. Thank you for your comment but I have two questions.

What is ause?
 
> In ideal world, both patches would be reduced to two lines in one file. 
Is it better for me to rewrite *.mk drastically to centralize ADJUSTVISIBILITY?
Comment 5 pavel 2005-03-13 08:55:12 UTC
taniguchi: your patch is OK. But the sources are not ;-) If you want to fix the
probem in the source itself, you're welcome :-)

Ause (Hans' IRC nick) is build master so he can bring some light into this and
maybe propose better solution (restructuring processing of this) and thus also
taking account about != GCC part with it.

Let's wait for his opinion.
Comment 6 hjs 2005-03-14 10:43:08 UTC
my suggestion would be to have something like

.IF "$(OS)"=="SOLARIS" && "$(product)"=="full" && "$(debug)"=="" ....
ADJUSTVISIBILITY*=$(WRAPCMD) adjustvisibility
.ENDIF

in "unitools.mk" and change the rules to

-.IF "$(OS)"=="SOLARIS" && "$(product)"=="full" && "$(debug)"==""    
+.IF "$(ADJUSTVISIBILITY)"!=""

this is a big patch once, but the desired "two line" solution afterwards.
Comment 7 pavel 2005-03-15 19:38:50 UTC
taniguchi: could you please do that in one go?

hjs: we do not need $(WRAPCMD), because this is really Solaris only. Right?
Comment 8 pavel 2005-04-08 07:34:35 UTC
taniguchi: can you please do that so we can include it into 2.0 or should I
retarget to 2.0.1?
Comment 9 pavel 2005-05-17 09:46:19 UTC
retarget to 2.0.1.
Comment 10 Martin Hollmichel 2005-11-04 09:35:43 UTC
set target to Later, please feel free to reset target to earlier one, if patch
is complete.
Comment 11 stx123 2006-03-29 16:16:04 UTC
Hi "taniguchi", are you going to work on the proposed improvement of the patch?
Comment 12 stx123 2006-06-12 19:21:55 UTC
It seems the patch is not (yet) in a state to be integrated.
Feel free to reopen the issue if you are going to work on it again.
Comment 13 Martin Hollmichel 2007-06-26 05:46:38 UTC
close issue.