Skip to contentSkip to conversation

package/Makefile.in: Do not add --enable-debug flag.

PATCH New Johan Oudinet 2 participants 3 messages Mar 11, 2015 – Mar 12, 2015

lore
Share & tools
lore

Johan Oudinet

author
#0 lore raw

PATCH 1/1 package/Makefile.in: Do not add --enable-debug flag.

Adding this flag when BR2_ENABLE_DEBUG is activated make several packages to produce binaries that do not work as expected (e.g., dhcp, lame, nano). Moreover, the help message of BR2_ENABLE_DEBUG does not say it is adding this flag. It is supposed to build packages with debugging symbols enabled. So, let it do that only.

Signed-off-by Johan Oudinet <[email protected]>
---
 package/Makefile.in | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
package/Makefile.in +1 -3
index 803b162..2995222 100644
@@ -387,9 +387,7 @@ ifneq ($(BR2_INSTALL_LIBSTDCPP),y)
 TARGET_CONFIGURE_OPTS += CXX=false
 endif
 
-ifeq ($(BR2_ENABLE_DEBUG),y)
-ENABLE_DEBUG := --enable-debug
-else
+ifneq ($(BR2_ENABLE_DEBUG),y)
 ENABLE_DEBUG := --disable-debug
 endif
Signature · 1 line
2.1.0
#1 lore raw

PATCH 1/1 package/Makefile.in: Do not add --enable-debug flag.

Dear Johan Oudinet,

Quoted 7 lines · Johan Oudinet

Adding this flag when BR2_ENABLE_DEBUG is activated make several packages to produce binaries that do not work as expected (e.g., dhcp, lame, nano). Moreover, the help message of BR2_ENABLE_DEBUG does not say it is adding this flag. It is supposed to build packages with debugging symbols enabled. So, let it do that only.

Signed-off-by Johan Oudinet <[email protected]>

I am personally in favor of this change, so thanks for bringing it up.

package/Makefile.in ifneq ($(BR2_INSTALL_LIBSTDCPP),y) · line 392

Quoted 6 lines
package/Makefile.in +1 -3
-ifeq ($(BR2_ENABLE_DEBUG),y)
-ENABLE_DEBUG := --enable-debug
-else
+ifneq ($(BR2_ENABLE_DEBUG),y)
 ENABLE_DEBUG := --disable-debug
 endif

So if we have BR2_ENABLE_DEBUG enabled, then we don't pass --disable-debug. And when BR2_ENABLE_DEBUG is disabled, we're passing it. I'm not sure to understand the logic here.

Shouldn't we simply unconditionally pass --disable-debug, or not pass anything at all?

Also, a number of packages had workarounds in their specific .mk file to avoid --enable-debug. It would be good to get rid of such workarounds as well.

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
author
#2 lore raw

PATCH 1/1 package/Makefile.in: Do not add --enable-debug flag.

Thomas, All,

Quoted 11 lines · Thomas Petazzoni

package/Makefile.in ifneq ($(BR2_INSTALL_LIBSTDCPP),y) · line 392

Quoted 6 lines
package/Makefile.in +1 -3
-ifeq ($(BR2_ENABLE_DEBUG),y)
-ENABLE_DEBUG := --enable-debug
-else
+ifneq ($(BR2_ENABLE_DEBUG),y)
 ENABLE_DEBUG := --disable-debug
 endif

So if we have BR2_ENABLE_DEBUG enabled, then we don't pass --disable-debug. And when BR2_ENABLE_DEBUG is disabled, we're passing it. I'm not sure to understand the logic here.

Quoted 2 lines

Shouldn't we simply unconditionally pass --disable-debug, or not pass anything at all?

I felt bad in passing --disable-debug when BR2_ENABLE_DEBUG is set, and I didn't want to remove it otherwise (in case a package expects it). Still, the result is not logic. Actually, there is no good reason for a package to depends on a --disable-debug flag to produce a version without debugging information (see for example
http://stackoverflow.com/a/4680578/1448926)
I'm going to remove --disable-debug too.

Quoted 4 lines

Also, a number of packages had workarounds in their specific .mk file to avoid --enable-debug. It would be good to get rid of such workarounds as well.

Indeed, I'm going to get rid of such workarounds and send another version for this patch.
Thanks for the quick review.

Signature · 1 line
Johan Oudinet