| View previous topic :: View next topic |
| Author |
Message |
Christopher Benson-Manica
Guest
|
 Posted: Tue May 17, 2005 4:16 pm Post subject: Problems with incremental link |
 |
|
We are experiencing periodic problems that we believe are related to
the incremental linker. What happens is that projects compiled and
linked via "make" (i.e., only recompile dependent objects whose
sources have changed) will occasionally demonstrate bizarre
nonsensical behavior that is fixed by a full recompile of all objects.
My boss, at least, believes that this happens when data stored by the
linker gets "out of synch" (somehow). Is behavior similar to what
I've described a known issue with version 4.00 of ilink32? Any
suggestions? Is there a way to force a full link of all objects? (It
looks like the -C option might do this, but I figure I'll ask anyway...)
--
Christopher Benson-Manica | I *should* know what I'm talking about - if I
ataru(at)cyberspace.org | don't, I need to know. Flames welcome.
|
|
| Back to top |
|
 |
Christopher Benson-Manica
Guest
|
| Posted: Tue May 17, 2005 4:31 pm Post subject: Re: Problems with incremental link |
|
|
<428a18cf (AT) newsgroups (DOT) borland.com>
Christopher Benson-Manica <ataru (AT) nospam (DOT) cyberspace.org> wrote:
| Quote: | looks like the -C option might do this, but I figure I'll ask anyway...)
|
Well, now I see that -Gn supposedly does this, but we already use that
flag, so I suppose the incremental nature of the link is not to
blame... Now I *really* could use suggestions :-(
--
Christopher Benson-Manica
ataru(at)cyberspace.org
|
|
| Back to top |
|
|
Bob Gonder
Guest
|
| Posted: Tue May 17, 2005 5:43 pm Post subject: Re: Problems with incremental link |
|
|
Christopher Benson-Manica wrote:
| Quote: | We are experiencing periodic problems that we believe are related to
the incremental linker. What happens is that projects compiled and
linked via "make" (i.e., only recompile dependent objects whose
sources have changed) will occasionally demonstrate bizarre
nonsensical behavior that is fixed by a full recompile of all objects.
My boss, at least, believes that this happens when data stored by the
linker gets "out of synch" (somehow). Is behavior similar to what
I've described a known issue with version 4.00 of ilink32? Any
suggestions? Is there a way to force a full link of all objects? (It
looks like the -C option might do this, but I figure I'll ask anyway...)
|
As you already surmised:
/C/Gn
You didn't say what " bizarre nonsensical behavior" means or is.
Might be related to the compile phase.
You might try turning off precompiled headers and see if it is less
bizarre.
|
|
| Back to top |
|
|
Christopher Benson-Manica
Guest
|
| Posted: Tue May 17, 2005 5:56 pm Post subject: Re: Problems with incremental link |
|
|
Bob Gonder <notbg (AT) notmindspring (DOT) invalid> wrote:
| Quote: | You didn't say what " bizarre nonsensical behavior" means or is.
|
Basically, code that is known to be correct misbehaves unpredictably.
It's known to be correct because it works fine following a recompile
of the relevant object (without changing the code).
| Quote: | You might try turning off precompiled headers and see if it is less
bizarre.
|
Hm, that might be something to look at. Thanks.
--
Christopher Benson-Manica | I *should* know what I'm talking about - if I
ataru(at)cyberspace.org | don't, I need to know. Flames welcome.
|
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
